logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x760cc53d82b0>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x760cc53d8610>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x760cc53bec70>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x760cc53562e0>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x760cc536ac10>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x760cc5372bb0>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x760cc53313a0>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x760cc5331400>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['001_cube', '001_cube', '001_cube', '001_cube', '001_cube', '001_cube', '001_cube', '001_cube', '001_cube', '001_cube', '001_cube', '001_cube', '001_cube', '001_cube'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x760cc52ca790>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x760cc52df640>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x760cc52df580>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_001_cube/pretrained'), 'resume_wandb_run': False, 'wandb_id': 'b42i2pxv', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_001_cube', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '001_cube',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 001_cube
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
001_cube already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 001_cube
adding 98 observations
Extended graph 001_cube with new points. New model:
{'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
001_cube not in memory ([])
Adding a new graph to memory.
init object model with id 001_cube
building graph from 15 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x760cc528aa90>
Added new graph with id 001_cube to memory.
Model for 001_cube:
   Contains 15 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '001_cube',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 001_cube
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
001_cube already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 001_cube
adding 99 observations
Extended graph 001_cube with new points. New model:
{'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
001_cube already in memory (['001_cube'])
Updating existing graph for 001_cube
adding 15 observations
Extended graph 001_cube with new points. New model:
{'patch_1': Model for 001_cube:
   Contains 30 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '001_cube',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 001_cube
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
001_cube already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 001_cube
adding 98 observations
Extended graph 001_cube with new points. New model:
{'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
001_cube already in memory (['001_cube'])
Updating existing graph for 001_cube
adding 15 observations
Extended graph 001_cube with new points. New model:
{'patch_1': Model for 001_cube:
   Contains 45 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '001_cube',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 001_cube
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
001_cube already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 001_cube
adding 98 observations
Extended graph 001_cube with new points. New model:
{'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
001_cube already in memory (['001_cube'])
Updating existing graph for 001_cube
adding 15 observations
Extended graph 001_cube with new points. New model:
{'patch_1': Model for 001_cube:
   Contains 60 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '001_cube',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 001_cube
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
001_cube already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 001_cube
adding 98 observations
Extended graph 001_cube with new points. New model:
{'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
001_cube already in memory (['001_cube'])
Updating existing graph for 001_cube
adding 15 observations
Extended graph 001_cube with new points. New model:
{'patch_1': Model for 001_cube:
   Contains 75 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '001_cube',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 001_cube
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
001_cube already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 001_cube
adding 98 observations
Extended graph 001_cube with new points. New model:
{'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
001_cube already in memory (['001_cube'])
Updating existing graph for 001_cube
adding 15 observations
Extended graph 001_cube with new points. New model:
{'patch_1': Model for 001_cube:
   Contains 90 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '001_cube',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 001_cube
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
001_cube already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 001_cube
adding 115 observations
Extended graph 001_cube with new points. New model:
{'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
001_cube already in memory (['001_cube'])
Updating existing graph for 001_cube
adding 41 observations
Extended graph 001_cube with new points. New model:
{'patch_1': Model for 001_cube:
   Contains 131 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '001_cube',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 001_cube
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
001_cube already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 001_cube
adding 115 observations
Extended graph 001_cube with new points. New model:
{'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
001_cube already in memory (['001_cube'])
Updating existing graph for 001_cube
adding 42 observations
Extended graph 001_cube with new points. New model:
{'patch_1': Model for 001_cube:
   Contains 173 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '001_cube',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 001_cube
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
001_cube already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 001_cube
adding 114 observations
Extended graph 001_cube with new points. New model:
{'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
001_cube already in memory (['001_cube'])
Updating existing graph for 001_cube
adding 47 observations
Extended graph 001_cube with new points. New model:
{'patch_1': Model for 001_cube:
   Contains 219 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '001_cube',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 001_cube
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
001_cube already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 001_cube
adding 114 observations
Extended graph 001_cube with new points. New model:
{'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
001_cube already in memory (['001_cube'])
Updating existing graph for 001_cube
adding 44 observations
Extended graph 001_cube with new points. New model:
{'patch_1': Model for 001_cube:
   Contains 261 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '001_cube',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 001_cube
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
001_cube already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 001_cube
adding 114 observations
Extended graph 001_cube with new points. New model:
{'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
001_cube already in memory (['001_cube'])
Updating existing graph for 001_cube
adding 44 observations
Extended graph 001_cube with new points. New model:
{'patch_1': Model for 001_cube:
   Contains 303 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '001_cube',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 001_cube
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
001_cube already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 001_cube
adding 114 observations
Extended graph 001_cube with new points. New model:
{'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
001_cube already in memory (['001_cube'])
Updating existing graph for 001_cube
adding 47 observations
Extended graph 001_cube with new points. New model:
{'patch_1': Model for 001_cube:
   Contains 349 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '001_cube',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 001_cube
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
001_cube already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 001_cube
adding 114 observations
Extended graph 001_cube with new points. New model:
{'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
001_cube already in memory (['001_cube'])
Updating existing graph for 001_cube
adding 42 observations
Extended graph 001_cube with new points. New model:
{'patch_1': Model for 001_cube:
   Contains 391 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '001_cube',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 001_cube
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
001_cube already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 001_cube
adding 114 observations
Extended graph 001_cube with new points. New model:
{'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
001_cube already in memory (['001_cube'])
Updating existing graph for 001_cube
adding 40 observations
Extended graph 001_cube with new points. New model:
{'patch_1': Model for 001_cube:
   Contains 431 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '001_cube',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_001_cube/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x7b252509d370>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x7b252509d6d0>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x7b252500c0d0>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7b25250143a0>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7b2525026cd0>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x7b2525030c70>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x7b2524fee460>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x7b2524fee4c0>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['006_disk', '006_disk', '006_disk', '006_disk', '006_disk', '006_disk', '006_disk', '006_disk', '006_disk', '006_disk', '006_disk', '006_disk', '006_disk', '006_disk'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x7b2524f87850>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7b2524f9c700>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7b2524f9c640>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_006_disk/pretrained'), 'resume_wandb_run': False, 'wandb_id': '24t7urhf', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_006_disk', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '006_disk',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 006_disk
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
006_disk already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 006_disk
adding 99 observations
Extended graph 006_disk with new points. New model:
{'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
006_disk not in memory ([])
Adding a new graph to memory.
init object model with id 006_disk
building graph from 16 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x7b25245d0d60>
Added new graph with id 006_disk to memory.
Model for 006_disk:
   Contains 16 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '006_disk',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 006_disk
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
006_disk already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 006_disk
adding 15 observations
Extended graph 006_disk with new points. New model:
{'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
006_disk already in memory (['006_disk'])
Updating existing graph for 006_disk
adding 7 observations
Extended graph 006_disk with new points. New model:
{'patch_1': Model for 006_disk:
   Contains 23 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '006_disk',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 006_disk
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
006_disk already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 006_disk
adding 99 observations
Extended graph 006_disk with new points. New model:
{'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
006_disk already in memory (['006_disk'])
Updating existing graph for 006_disk
adding 16 observations
Extended graph 006_disk with new points. New model:
{'patch_1': Model for 006_disk:
   Contains 39 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '006_disk',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 006_disk
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
006_disk already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 006_disk
adding 15 observations
Extended graph 006_disk with new points. New model:
{'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
006_disk already in memory (['006_disk'])
Updating existing graph for 006_disk
adding 13 observations
Extended graph 006_disk with new points. New model:
{'patch_1': Model for 006_disk:
   Contains 52 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '006_disk',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 006_disk
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
006_disk already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 006_disk
adding 12 observations
Extended graph 006_disk with new points. New model:
{'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
006_disk already in memory (['006_disk'])
Updating existing graph for 006_disk
adding 1 observations
Extended graph 006_disk with new points. New model:
{'patch_1': Model for 006_disk:
   Contains 53 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '006_disk',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 006_disk
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
006_disk already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 006_disk
adding 11 observations
Extended graph 006_disk with new points. New model:
{'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
006_disk already in memory (['006_disk'])
Updating existing graph for 006_disk
adding 1 observations
Extended graph 006_disk with new points. New model:
{'patch_1': Model for 006_disk:
   Contains 54 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '006_disk',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 006_disk
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
006_disk already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 006_disk
adding 94 observations
Extended graph 006_disk with new points. New model:
{'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
006_disk already in memory (['006_disk'])
Updating existing graph for 006_disk
adding 16 observations
Extended graph 006_disk with new points. New model:
{'patch_1': Model for 006_disk:
   Contains 70 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '006_disk',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 006_disk
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
006_disk already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 006_disk
adding 90 observations
Extended graph 006_disk with new points. New model:
{'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
006_disk already in memory (['006_disk'])
Updating existing graph for 006_disk
adding 16 observations
Extended graph 006_disk with new points. New model:
{'patch_1': Model for 006_disk:
   Contains 86 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '006_disk',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 006_disk
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
006_disk already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 006_disk
adding 93 observations
Extended graph 006_disk with new points. New model:
{'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
006_disk already in memory (['006_disk'])
Updating existing graph for 006_disk
adding 17 observations
Extended graph 006_disk with new points. New model:
{'patch_1': Model for 006_disk:
   Contains 103 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '006_disk',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 006_disk
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
006_disk already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 006_disk
adding 91 observations
Extended graph 006_disk with new points. New model:
{'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
006_disk already in memory (['006_disk'])
Updating existing graph for 006_disk
adding 16 observations
Extended graph 006_disk with new points. New model:
{'patch_1': Model for 006_disk:
   Contains 119 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '006_disk',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 006_disk
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
006_disk already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 006_disk
adding 91 observations
Extended graph 006_disk with new points. New model:
{'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
006_disk already in memory (['006_disk'])
Updating existing graph for 006_disk
adding 16 observations
Extended graph 006_disk with new points. New model:
{'patch_1': Model for 006_disk:
   Contains 135 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '006_disk',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 006_disk
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
006_disk already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 006_disk
adding 93 observations
Extended graph 006_disk with new points. New model:
{'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
006_disk already in memory (['006_disk'])
Updating existing graph for 006_disk
adding 17 observations
Extended graph 006_disk with new points. New model:
{'patch_1': Model for 006_disk:
   Contains 152 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '006_disk',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 006_disk
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
006_disk already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 006_disk
adding 90 observations
Extended graph 006_disk with new points. New model:
{'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
006_disk already in memory (['006_disk'])
Updating existing graph for 006_disk
adding 16 observations
Extended graph 006_disk with new points. New model:
{'patch_1': Model for 006_disk:
   Contains 168 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '006_disk',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 006_disk
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
006_disk already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 006_disk
adding 94 observations
Extended graph 006_disk with new points. New model:
{'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
006_disk already in memory (['006_disk'])
Updating existing graph for 006_disk
adding 16 observations
Extended graph 006_disk with new points. New model:
{'patch_1': Model for 006_disk:
   Contains 184 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '006_disk',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_006_disk/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x7997bae962b0>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x7997bae96610>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x7997bae7ec70>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7997bae152e0>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7997bae29c10>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x7997bae31bb0>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x7997badef3a0>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x7997badef400>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['002_cube_tbp_horz', '002_cube_tbp_horz', '002_cube_tbp_horz', '002_cube_tbp_horz', '002_cube_tbp_horz', '002_cube_tbp_horz', '002_cube_tbp_horz', '002_cube_tbp_horz', '002_cube_tbp_horz', '002_cube_tbp_horz', '002_cube_tbp_horz', '002_cube_tbp_horz', '002_cube_tbp_horz', '002_cube_tbp_horz'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x7997bad88790>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7997bad9d640>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7997bad9d580>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_002_cube_tbp_horz/pretrained'), 'resume_wandb_run': False, 'wandb_id': 'b12if4z0', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_002_cube_tbp_horz', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '002_cube_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 002_cube_tbp_horz
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
002_cube_tbp_horz not in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Adding a new graph to memory.
init object model with id 002_cube_tbp_horz
building graph from 112 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x7997b8810a90>
Added new graph with id 002_cube_tbp_horz to memory.
Model for 002_cube_tbp_horz:
   Contains 112 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,


---Updating memory of learning_module_1---
002_cube_tbp_horz not in memory ([])
Adding a new graph to memory.
init object model with id 002_cube_tbp_horz
building graph from 50 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x7997b8810bb0>
Added new graph with id 002_cube_tbp_horz to memory.
Model for 002_cube_tbp_horz:
   Contains 50 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '002_cube_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 002_cube_tbp_horz
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
002_cube_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 99 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_0': Model for 002_cube_tbp_horz:
   Contains 204 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
002_cube_tbp_horz already in memory (['002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 15 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_1': Model for 002_cube_tbp_horz:
   Contains 65 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '002_cube_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 002_cube_tbp_horz
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
002_cube_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 98 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_0': Model for 002_cube_tbp_horz:
   Contains 295 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
002_cube_tbp_horz already in memory (['002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 15 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_1': Model for 002_cube_tbp_horz:
   Contains 80 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '002_cube_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 002_cube_tbp_horz
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
002_cube_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 98 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_0': Model for 002_cube_tbp_horz:
   Contains 379 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
002_cube_tbp_horz already in memory (['002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 15 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_1': Model for 002_cube_tbp_horz:
   Contains 95 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '002_cube_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 002_cube_tbp_horz
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
002_cube_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 98 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_0': Model for 002_cube_tbp_horz:
   Contains 465 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
002_cube_tbp_horz already in memory (['002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 15 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_1': Model for 002_cube_tbp_horz:
   Contains 110 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '002_cube_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 002_cube_tbp_horz
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
002_cube_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 98 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_0': Model for 002_cube_tbp_horz:
   Contains 551 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
002_cube_tbp_horz already in memory (['002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 15 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_1': Model for 002_cube_tbp_horz:
   Contains 125 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '002_cube_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 002_cube_tbp_horz
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
002_cube_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 115 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_0': Model for 002_cube_tbp_horz:
   Contains 664 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
002_cube_tbp_horz already in memory (['002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 45 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_1': Model for 002_cube_tbp_horz:
   Contains 170 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '002_cube_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 002_cube_tbp_horz
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
002_cube_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 115 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_0': Model for 002_cube_tbp_horz:
   Contains 775 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
002_cube_tbp_horz already in memory (['002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 42 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_1': Model for 002_cube_tbp_horz:
   Contains 212 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '002_cube_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 002_cube_tbp_horz
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
002_cube_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 114 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_0': Model for 002_cube_tbp_horz:
   Contains 884 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
002_cube_tbp_horz already in memory (['002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 51 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_1': Model for 002_cube_tbp_horz:
   Contains 262 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '002_cube_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 002_cube_tbp_horz
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
002_cube_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 114 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_0': Model for 002_cube_tbp_horz:
   Contains 992 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
002_cube_tbp_horz already in memory (['002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 45 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_1': Model for 002_cube_tbp_horz:
   Contains 305 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '002_cube_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 002_cube_tbp_horz
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
002_cube_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 114 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_0': Model for 002_cube_tbp_horz:
   Contains 1102 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
002_cube_tbp_horz already in memory (['002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 44 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_1': Model for 002_cube_tbp_horz:
   Contains 347 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '002_cube_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 002_cube_tbp_horz
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
002_cube_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 114 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_0': Model for 002_cube_tbp_horz:
   Contains 1207 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
002_cube_tbp_horz already in memory (['002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 47 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_1': Model for 002_cube_tbp_horz:
   Contains 393 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '002_cube_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 002_cube_tbp_horz
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
002_cube_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 114 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_0': Model for 002_cube_tbp_horz:
   Contains 1318 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
002_cube_tbp_horz already in memory (['002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 42 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_1': Model for 002_cube_tbp_horz:
   Contains 435 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '002_cube_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 002_cube_tbp_horz
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
002_cube_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 114 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_0': Model for 002_cube_tbp_horz:
   Contains 1425 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
002_cube_tbp_horz already in memory (['002_cube_tbp_horz'])
Updating existing graph for 002_cube_tbp_horz
adding 40 observations
Extended graph 002_cube_tbp_horz with new points. New model:
{'patch_1': Model for 002_cube_tbp_horz:
   Contains 475 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '002_cube_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_002_cube_tbp_horz/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x711f35e562b0>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x711f35e56610>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x711f35e3ec70>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x711f35dd52e0>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x711f35deac10>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x711f35df1bb0>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x711f35db03a0>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x711f35db0400>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['004_cube_numenta_horz', '004_cube_numenta_horz', '004_cube_numenta_horz', '004_cube_numenta_horz', '004_cube_numenta_horz', '004_cube_numenta_horz', '004_cube_numenta_horz', '004_cube_numenta_horz', '004_cube_numenta_horz', '004_cube_numenta_horz', '004_cube_numenta_horz', '004_cube_numenta_horz', '004_cube_numenta_horz', '004_cube_numenta_horz'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x711f353c8790>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x711f353de640>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x711f353de580>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_004_cube_numenta_horz/pretrained'), 'resume_wandb_run': False, 'wandb_id': '3n37ovcy', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_004_cube_numenta_horz', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '004_cube_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 004_cube_numenta_horz
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
004_cube_numenta_horz not in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Adding a new graph to memory.
init object model with id 004_cube_numenta_horz
building graph from 112 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x711f32f90a30>
Added new graph with id 004_cube_numenta_horz to memory.
Model for 004_cube_numenta_horz:
   Contains 112 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,


---Updating memory of learning_module_1---
004_cube_numenta_horz not in memory ([])
Adding a new graph to memory.
init object model with id 004_cube_numenta_horz
building graph from 48 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x711f32f90b50>
Added new graph with id 004_cube_numenta_horz to memory.
Model for 004_cube_numenta_horz:
   Contains 48 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '004_cube_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 004_cube_numenta_horz
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
004_cube_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 99 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_0': Model for 004_cube_numenta_horz:
   Contains 204 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
004_cube_numenta_horz already in memory (['004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 15 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_1': Model for 004_cube_numenta_horz:
   Contains 63 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '004_cube_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 004_cube_numenta_horz
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
004_cube_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 98 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_0': Model for 004_cube_numenta_horz:
   Contains 295 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
004_cube_numenta_horz already in memory (['004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 15 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_1': Model for 004_cube_numenta_horz:
   Contains 78 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '004_cube_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 004_cube_numenta_horz
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
004_cube_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 98 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_0': Model for 004_cube_numenta_horz:
   Contains 379 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
004_cube_numenta_horz already in memory (['004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 15 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_1': Model for 004_cube_numenta_horz:
   Contains 93 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '004_cube_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 004_cube_numenta_horz
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
004_cube_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 98 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_0': Model for 004_cube_numenta_horz:
   Contains 465 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
004_cube_numenta_horz already in memory (['004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 15 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_1': Model for 004_cube_numenta_horz:
   Contains 108 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '004_cube_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 004_cube_numenta_horz
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
004_cube_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 98 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_0': Model for 004_cube_numenta_horz:
   Contains 551 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
004_cube_numenta_horz already in memory (['004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 15 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_1': Model for 004_cube_numenta_horz:
   Contains 123 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '004_cube_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 004_cube_numenta_horz
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
004_cube_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 115 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_0': Model for 004_cube_numenta_horz:
   Contains 664 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
004_cube_numenta_horz already in memory (['004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 41 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_1': Model for 004_cube_numenta_horz:
   Contains 164 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '004_cube_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 004_cube_numenta_horz
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
004_cube_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 115 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_0': Model for 004_cube_numenta_horz:
   Contains 775 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
004_cube_numenta_horz already in memory (['004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 43 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_1': Model for 004_cube_numenta_horz:
   Contains 207 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '004_cube_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 004_cube_numenta_horz
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
004_cube_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 114 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_0': Model for 004_cube_numenta_horz:
   Contains 884 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
004_cube_numenta_horz already in memory (['004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 48 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_1': Model for 004_cube_numenta_horz:
   Contains 254 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '004_cube_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 004_cube_numenta_horz
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
004_cube_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 114 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_0': Model for 004_cube_numenta_horz:
   Contains 992 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
004_cube_numenta_horz already in memory (['004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 45 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_1': Model for 004_cube_numenta_horz:
   Contains 297 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '004_cube_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 004_cube_numenta_horz
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
004_cube_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 114 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_0': Model for 004_cube_numenta_horz:
   Contains 1102 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
004_cube_numenta_horz already in memory (['004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 44 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_1': Model for 004_cube_numenta_horz:
   Contains 339 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '004_cube_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 004_cube_numenta_horz
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
004_cube_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 114 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_0': Model for 004_cube_numenta_horz:
   Contains 1207 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
004_cube_numenta_horz already in memory (['004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 47 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_1': Model for 004_cube_numenta_horz:
   Contains 385 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '004_cube_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 004_cube_numenta_horz
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
004_cube_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 114 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_0': Model for 004_cube_numenta_horz:
   Contains 1318 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
004_cube_numenta_horz already in memory (['004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 42 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_1': Model for 004_cube_numenta_horz:
   Contains 427 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '004_cube_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 004_cube_numenta_horz
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
004_cube_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 114 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_0': Model for 004_cube_numenta_horz:
   Contains 1425 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
004_cube_numenta_horz already in memory (['004_cube_numenta_horz'])
Updating existing graph for 004_cube_numenta_horz
adding 40 observations
Extended graph 004_cube_numenta_horz with new points. New model:
{'patch_1': Model for 004_cube_numenta_horz:
   Contains 467 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '004_cube_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_004_cube_numenta_horz/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x7a99f7bdd2e0>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x7a99f7bdd640>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x7a99f7bbcc70>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7a99f7b53310>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7a99f7b66c40>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x7a99f7b6fbe0>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x7a99f7b2d3d0>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x7a99f7b2d430>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['007_disk_tbp_horz', '007_disk_tbp_horz', '007_disk_tbp_horz', '007_disk_tbp_horz', '007_disk_tbp_horz', '007_disk_tbp_horz', '007_disk_tbp_horz', '007_disk_tbp_horz', '007_disk_tbp_horz', '007_disk_tbp_horz', '007_disk_tbp_horz', '007_disk_tbp_horz', '007_disk_tbp_horz', '007_disk_tbp_horz'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x7a99f7ac67c0>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7a99f7adb670>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7a99f7adb5b0>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_007_disk_tbp_horz/pretrained'), 'resume_wandb_run': False, 'wandb_id': 'bw0ijg5c', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_007_disk_tbp_horz', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '007_disk_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 007_disk_tbp_horz
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
007_disk_tbp_horz not in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Adding a new graph to memory.
init object model with id 007_disk_tbp_horz
building graph from 115 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x7a99f5553b80>
Added new graph with id 007_disk_tbp_horz to memory.
Model for 007_disk_tbp_horz:
   Contains 115 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,


---Updating memory of learning_module_1---
007_disk_tbp_horz not in memory ([])
Adding a new graph to memory.
init object model with id 007_disk_tbp_horz
building graph from 51 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x7a99f5553ca0>
Added new graph with id 007_disk_tbp_horz to memory.
Model for 007_disk_tbp_horz:
   Contains 51 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '007_disk_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 007_disk_tbp_horz
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
007_disk_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 15 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_0': Model for 007_disk_tbp_horz:
   Contains 130 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
007_disk_tbp_horz already in memory (['007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 7 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_1': Model for 007_disk_tbp_horz:
   Contains 58 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '007_disk_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 007_disk_tbp_horz
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
007_disk_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 99 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_0': Model for 007_disk_tbp_horz:
   Contains 229 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
007_disk_tbp_horz already in memory (['007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 16 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_1': Model for 007_disk_tbp_horz:
   Contains 74 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '007_disk_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 007_disk_tbp_horz
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
007_disk_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 15 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_0': Model for 007_disk_tbp_horz:
   Contains 244 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
007_disk_tbp_horz already in memory (['007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 13 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_1': Model for 007_disk_tbp_horz:
   Contains 87 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '007_disk_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 007_disk_tbp_horz
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
007_disk_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 12 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_0': Model for 007_disk_tbp_horz:
   Contains 256 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
007_disk_tbp_horz already in memory (['007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 1 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_1': Model for 007_disk_tbp_horz:
   Contains 88 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '007_disk_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 007_disk_tbp_horz
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
007_disk_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 11 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_0': Model for 007_disk_tbp_horz:
   Contains 267 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
007_disk_tbp_horz already in memory (['007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 1 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_1': Model for 007_disk_tbp_horz:
   Contains 89 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '007_disk_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 007_disk_tbp_horz
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
007_disk_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 102 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_0': Model for 007_disk_tbp_horz:
   Contains 368 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
007_disk_tbp_horz already in memory (['007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 34 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_1': Model for 007_disk_tbp_horz:
   Contains 122 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '007_disk_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 007_disk_tbp_horz
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
007_disk_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 96 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_0': Model for 007_disk_tbp_horz:
   Contains 461 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
007_disk_tbp_horz already in memory (['007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 29 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_1': Model for 007_disk_tbp_horz:
   Contains 151 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '007_disk_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 007_disk_tbp_horz
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
007_disk_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 102 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_0': Model for 007_disk_tbp_horz:
   Contains 550 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
007_disk_tbp_horz already in memory (['007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 34 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_1': Model for 007_disk_tbp_horz:
   Contains 182 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '007_disk_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 007_disk_tbp_horz
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
007_disk_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 96 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_0': Model for 007_disk_tbp_horz:
   Contains 635 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
007_disk_tbp_horz already in memory (['007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 33 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_1': Model for 007_disk_tbp_horz:
   Contains 213 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '007_disk_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 007_disk_tbp_horz
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
007_disk_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 91 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_0': Model for 007_disk_tbp_horz:
   Contains 725 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
007_disk_tbp_horz already in memory (['007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 16 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_1': Model for 007_disk_tbp_horz:
   Contains 229 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '007_disk_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 007_disk_tbp_horz
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
007_disk_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 93 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_0': Model for 007_disk_tbp_horz:
   Contains 815 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
007_disk_tbp_horz already in memory (['007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 17 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_1': Model for 007_disk_tbp_horz:
   Contains 246 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '007_disk_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 007_disk_tbp_horz
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
007_disk_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 90 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_0': Model for 007_disk_tbp_horz:
   Contains 897 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
007_disk_tbp_horz already in memory (['007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 16 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_1': Model for 007_disk_tbp_horz:
   Contains 262 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '007_disk_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 007_disk_tbp_horz
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
007_disk_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 94 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_0': Model for 007_disk_tbp_horz:
   Contains 984 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
007_disk_tbp_horz already in memory (['007_disk_tbp_horz'])
Updating existing graph for 007_disk_tbp_horz
adding 16 observations
Extended graph 007_disk_tbp_horz with new points. New model:
{'patch_1': Model for 007_disk_tbp_horz:
   Contains 278 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '007_disk_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_007_disk_tbp_horz/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x76156eb4d310>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x76156eb4d670>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x76156d5fccd0>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x76156d594340>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x76156d5a6c70>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x76156d5b0c10>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x76156d56e400>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x76156d56e460>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['009_disk_numenta_horz', '009_disk_numenta_horz', '009_disk_numenta_horz', '009_disk_numenta_horz', '009_disk_numenta_horz', '009_disk_numenta_horz', '009_disk_numenta_horz', '009_disk_numenta_horz', '009_disk_numenta_horz', '009_disk_numenta_horz', '009_disk_numenta_horz', '009_disk_numenta_horz', '009_disk_numenta_horz', '009_disk_numenta_horz'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x76156d5077f0>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x76156d51c6a0>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x76156d51c5e0>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_009_disk_numenta_horz/pretrained'), 'resume_wandb_run': False, 'wandb_id': 't0l4d0c4', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_009_disk_numenta_horz', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '009_disk_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 009_disk_numenta_horz
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
009_disk_numenta_horz not in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Adding a new graph to memory.
init object model with id 009_disk_numenta_horz
building graph from 115 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x76156af91af0>
Added new graph with id 009_disk_numenta_horz to memory.
Model for 009_disk_numenta_horz:
   Contains 115 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,


---Updating memory of learning_module_1---
009_disk_numenta_horz not in memory ([])
Adding a new graph to memory.
init object model with id 009_disk_numenta_horz
building graph from 47 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x76156af91c10>
Added new graph with id 009_disk_numenta_horz to memory.
Model for 009_disk_numenta_horz:
   Contains 47 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '009_disk_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 009_disk_numenta_horz
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
009_disk_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 15 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_0': Model for 009_disk_numenta_horz:
   Contains 130 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
009_disk_numenta_horz already in memory (['009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 7 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_1': Model for 009_disk_numenta_horz:
   Contains 54 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '009_disk_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 009_disk_numenta_horz
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
009_disk_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 99 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_0': Model for 009_disk_numenta_horz:
   Contains 229 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
009_disk_numenta_horz already in memory (['009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 16 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_1': Model for 009_disk_numenta_horz:
   Contains 70 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '009_disk_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 009_disk_numenta_horz
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
009_disk_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 15 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_0': Model for 009_disk_numenta_horz:
   Contains 244 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
009_disk_numenta_horz already in memory (['009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 13 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_1': Model for 009_disk_numenta_horz:
   Contains 83 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '009_disk_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 009_disk_numenta_horz
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
009_disk_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 12 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_0': Model for 009_disk_numenta_horz:
   Contains 256 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
009_disk_numenta_horz already in memory (['009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 1 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_1': Model for 009_disk_numenta_horz:
   Contains 84 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '009_disk_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 009_disk_numenta_horz
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
009_disk_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 11 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_0': Model for 009_disk_numenta_horz:
   Contains 267 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
009_disk_numenta_horz already in memory (['009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 1 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_1': Model for 009_disk_numenta_horz:
   Contains 85 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '009_disk_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 009_disk_numenta_horz
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
009_disk_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 97 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_0': Model for 009_disk_numenta_horz:
   Contains 363 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
009_disk_numenta_horz already in memory (['009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 25 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_1': Model for 009_disk_numenta_horz:
   Contains 109 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '009_disk_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 009_disk_numenta_horz
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
009_disk_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 96 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_0': Model for 009_disk_numenta_horz:
   Contains 456 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
009_disk_numenta_horz already in memory (['009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 26 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_1': Model for 009_disk_numenta_horz:
   Contains 132 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '009_disk_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 009_disk_numenta_horz
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
009_disk_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 98 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_0': Model for 009_disk_numenta_horz:
   Contains 542 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
009_disk_numenta_horz already in memory (['009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 25 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_1': Model for 009_disk_numenta_horz:
   Contains 155 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '009_disk_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 009_disk_numenta_horz
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
009_disk_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 97 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_0': Model for 009_disk_numenta_horz:
   Contains 628 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
009_disk_numenta_horz already in memory (['009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 31 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_1': Model for 009_disk_numenta_horz:
   Contains 186 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '009_disk_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 009_disk_numenta_horz
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
009_disk_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 91 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_0': Model for 009_disk_numenta_horz:
   Contains 718 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
009_disk_numenta_horz already in memory (['009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 16 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_1': Model for 009_disk_numenta_horz:
   Contains 202 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '009_disk_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 009_disk_numenta_horz
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
009_disk_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 93 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_0': Model for 009_disk_numenta_horz:
   Contains 808 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
009_disk_numenta_horz already in memory (['009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 17 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_1': Model for 009_disk_numenta_horz:
   Contains 219 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '009_disk_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 009_disk_numenta_horz
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
009_disk_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 90 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_0': Model for 009_disk_numenta_horz:
   Contains 890 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
009_disk_numenta_horz already in memory (['009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 16 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_1': Model for 009_disk_numenta_horz:
   Contains 235 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '009_disk_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 009_disk_numenta_horz
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
009_disk_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 94 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_0': Model for 009_disk_numenta_horz:
   Contains 977 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
009_disk_numenta_horz already in memory (['009_disk_numenta_horz'])
Updating existing graph for 009_disk_numenta_horz
adding 16 observations
Extended graph 009_disk_numenta_horz with new points. New model:
{'patch_1': Model for 009_disk_numenta_horz:
   Contains 251 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '009_disk_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_009_disk_numenta_horz/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x70134334e2b0>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x70134334e610>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x70134333ec70>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7013432d52e0>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7013432eac10>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x7013432f1bb0>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x7013432af3a0>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x7013432af400>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['011_cylinder', '011_cylinder', '011_cylinder', '011_cylinder', '011_cylinder', '011_cylinder', '011_cylinder', '011_cylinder', '011_cylinder', '011_cylinder', '011_cylinder', '011_cylinder', '011_cylinder', '011_cylinder'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x701343248790>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x70134325d640>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x70134325d580>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_011_cylinder/pretrained'), 'resume_wandb_run': False, 'wandb_id': 'y6yb1d0p', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_011_cylinder', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '011_cylinder',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 011_cylinder
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
011_cylinder already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 011_cylinder
adding 80 observations
Extended graph 011_cylinder with new points. New model:
{'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
011_cylinder not in memory ([])
Adding a new graph to memory.
init object model with id 011_cylinder
building graph from 16 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x701343213160>
Added new graph with id 011_cylinder to memory.
Model for 011_cylinder:
   Contains 16 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '011_cylinder',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 011_cylinder
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
011_cylinder already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 011_cylinder
adding 80 observations
Extended graph 011_cylinder with new points. New model:
{'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
011_cylinder already in memory (['011_cylinder'])
Updating existing graph for 011_cylinder
adding 15 observations
Extended graph 011_cylinder with new points. New model:
{'patch_1': Model for 011_cylinder:
   Contains 31 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '011_cylinder',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 011_cylinder
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
011_cylinder already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 011_cylinder
adding 80 observations
Extended graph 011_cylinder with new points. New model:
{'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
011_cylinder already in memory (['011_cylinder'])
Updating existing graph for 011_cylinder
adding 16 observations
Extended graph 011_cylinder with new points. New model:
{'patch_1': Model for 011_cylinder:
   Contains 47 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '011_cylinder',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 011_cylinder
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
011_cylinder already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 011_cylinder
adding 80 observations
Extended graph 011_cylinder with new points. New model:
{'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
011_cylinder already in memory (['011_cylinder'])
Updating existing graph for 011_cylinder
adding 15 observations
Extended graph 011_cylinder with new points. New model:
{'patch_1': Model for 011_cylinder:
   Contains 62 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '011_cylinder',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 011_cylinder
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
011_cylinder already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 011_cylinder
adding 96 observations
Extended graph 011_cylinder with new points. New model:
{'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
011_cylinder already in memory (['011_cylinder'])
Updating existing graph for 011_cylinder
adding 12 observations
Extended graph 011_cylinder with new points. New model:
{'patch_1': Model for 011_cylinder:
   Contains 74 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '011_cylinder',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 011_cylinder
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
011_cylinder already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 011_cylinder
adding 96 observations
Extended graph 011_cylinder with new points. New model:
{'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
011_cylinder already in memory (['011_cylinder'])
Updating existing graph for 011_cylinder
adding 12 observations
Extended graph 011_cylinder with new points. New model:
{'patch_1': Model for 011_cylinder:
   Contains 86 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '011_cylinder',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 011_cylinder
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
011_cylinder already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 011_cylinder
adding 81 observations
Extended graph 011_cylinder with new points. New model:
{'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
011_cylinder already in memory (['011_cylinder'])
Updating existing graph for 011_cylinder
adding 14 observations
Extended graph 011_cylinder with new points. New model:
{'patch_1': Model for 011_cylinder:
   Contains 100 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '011_cylinder',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 011_cylinder
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
011_cylinder already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 011_cylinder
adding 85 observations
Extended graph 011_cylinder with new points. New model:
{'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
011_cylinder already in memory (['011_cylinder'])
Updating existing graph for 011_cylinder
adding 18 observations
Extended graph 011_cylinder with new points. New model:
{'patch_1': Model for 011_cylinder:
   Contains 118 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '011_cylinder',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 011_cylinder
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
011_cylinder already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 011_cylinder
adding 77 observations
Extended graph 011_cylinder with new points. New model:
{'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
011_cylinder already in memory (['011_cylinder'])
Updating existing graph for 011_cylinder
adding 18 observations
Extended graph 011_cylinder with new points. New model:
{'patch_1': Model for 011_cylinder:
   Contains 136 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '011_cylinder',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 011_cylinder
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
011_cylinder already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 011_cylinder
adding 78 observations
Extended graph 011_cylinder with new points. New model:
{'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
011_cylinder already in memory (['011_cylinder'])
Updating existing graph for 011_cylinder
adding 15 observations
Extended graph 011_cylinder with new points. New model:
{'patch_1': Model for 011_cylinder:
   Contains 150 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '011_cylinder',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 011_cylinder
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
011_cylinder already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 011_cylinder
adding 76 observations
Extended graph 011_cylinder with new points. New model:
{'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
011_cylinder already in memory (['011_cylinder'])
Updating existing graph for 011_cylinder
adding 15 observations
Extended graph 011_cylinder with new points. New model:
{'patch_1': Model for 011_cylinder:
   Contains 165 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '011_cylinder',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 011_cylinder
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
011_cylinder already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 011_cylinder
adding 78 observations
Extended graph 011_cylinder with new points. New model:
{'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
011_cylinder already in memory (['011_cylinder'])
Updating existing graph for 011_cylinder
adding 18 observations
Extended graph 011_cylinder with new points. New model:
{'patch_1': Model for 011_cylinder:
   Contains 183 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '011_cylinder',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 011_cylinder
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
011_cylinder already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 011_cylinder
adding 87 observations
Extended graph 011_cylinder with new points. New model:
{'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
011_cylinder already in memory (['011_cylinder'])
Updating existing graph for 011_cylinder
adding 18 observations
Extended graph 011_cylinder with new points. New model:
{'patch_1': Model for 011_cylinder:
   Contains 201 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '011_cylinder',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 011_cylinder
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
011_cylinder already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 011_cylinder
adding 79 observations
Extended graph 011_cylinder with new points. New model:
{'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
011_cylinder already in memory (['011_cylinder'])
Updating existing graph for 011_cylinder
adding 14 observations
Extended graph 011_cylinder with new points. New model:
{'patch_1': Model for 011_cylinder:
   Contains 215 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '011_cylinder',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_011_cylinder/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x77642325d190>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x77642325d4f0>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x77642323dfa0>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7764231d31c0>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7764231e7ee0>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x7764231f0a90>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x7764231ad280>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x7764231ad2e0>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['016_sphere', '016_sphere', '016_sphere', '016_sphere', '016_sphere', '016_sphere', '016_sphere', '016_sphere', '016_sphere', '016_sphere', '016_sphere', '016_sphere', '016_sphere', '016_sphere'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x7764227c1fa0>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7764227db520>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7764227db460>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_016_sphere/pretrained'), 'resume_wandb_run': False, 'wandb_id': 'txog80yr', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_016_sphere', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '016_sphere',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 016_sphere
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
016_sphere already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 016_sphere
adding 68 observations
Extended graph 016_sphere with new points. New model:
{'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
016_sphere not in memory ([])
Adding a new graph to memory.
init object model with id 016_sphere
building graph from 22 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x776422793490>
Added new graph with id 016_sphere to memory.
Model for 016_sphere:
   Contains 22 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '016_sphere',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 016_sphere
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
016_sphere already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 016_sphere
adding 68 observations
Extended graph 016_sphere with new points. New model:
{'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
016_sphere already in memory (['016_sphere'])
Updating existing graph for 016_sphere
adding 23 observations
Extended graph 016_sphere with new points. New model:
{'patch_1': Model for 016_sphere:
   Contains 45 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '016_sphere',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 016_sphere
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
016_sphere already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 016_sphere
adding 79 observations
Extended graph 016_sphere with new points. New model:
{'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
016_sphere already in memory (['016_sphere'])
Updating existing graph for 016_sphere
adding 25 observations
Extended graph 016_sphere with new points. New model:
{'patch_1': Model for 016_sphere:
   Contains 70 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '016_sphere',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 016_sphere
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
016_sphere already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 016_sphere
adding 71 observations
Extended graph 016_sphere with new points. New model:
{'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
016_sphere already in memory (['016_sphere'])
Updating existing graph for 016_sphere
adding 20 observations
Extended graph 016_sphere with new points. New model:
{'patch_1': Model for 016_sphere:
   Contains 90 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '016_sphere',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 016_sphere
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
016_sphere already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 016_sphere
adding 70 observations
Extended graph 016_sphere with new points. New model:
{'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
016_sphere already in memory (['016_sphere'])
Updating existing graph for 016_sphere
adding 22 observations
Extended graph 016_sphere with new points. New model:
{'patch_1': Model for 016_sphere:
   Contains 112 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '016_sphere',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 016_sphere
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
016_sphere already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 016_sphere
adding 74 observations
Extended graph 016_sphere with new points. New model:
{'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
016_sphere already in memory (['016_sphere'])
Updating existing graph for 016_sphere
adding 21 observations
Extended graph 016_sphere with new points. New model:
{'patch_1': Model for 016_sphere:
   Contains 133 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '016_sphere',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 016_sphere
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
016_sphere already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 016_sphere
adding 70 observations
Extended graph 016_sphere with new points. New model:
{'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
016_sphere already in memory (['016_sphere'])
Updating existing graph for 016_sphere
adding 24 observations
Extended graph 016_sphere with new points. New model:
{'patch_1': Model for 016_sphere:
   Contains 156 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '016_sphere',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 016_sphere
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
016_sphere already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 016_sphere
adding 71 observations
Extended graph 016_sphere with new points. New model:
{'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
016_sphere already in memory (['016_sphere'])
Updating existing graph for 016_sphere
adding 24 observations
Extended graph 016_sphere with new points. New model:
{'patch_1': Model for 016_sphere:
   Contains 180 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '016_sphere',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 016_sphere
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
016_sphere already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 016_sphere
adding 70 observations
Extended graph 016_sphere with new points. New model:
{'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
016_sphere already in memory (['016_sphere'])
Updating existing graph for 016_sphere
adding 22 observations
Extended graph 016_sphere with new points. New model:
{'patch_1': Model for 016_sphere:
   Contains 202 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '016_sphere',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 016_sphere
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
016_sphere already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 016_sphere
adding 70 observations
Extended graph 016_sphere with new points. New model:
{'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
016_sphere already in memory (['016_sphere'])
Updating existing graph for 016_sphere
adding 21 observations
Extended graph 016_sphere with new points. New model:
{'patch_1': Model for 016_sphere:
   Contains 223 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '016_sphere',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 016_sphere
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
016_sphere already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 016_sphere
adding 67 observations
Extended graph 016_sphere with new points. New model:
{'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
016_sphere already in memory (['016_sphere'])
Updating existing graph for 016_sphere
adding 21 observations
Extended graph 016_sphere with new points. New model:
{'patch_1': Model for 016_sphere:
   Contains 243 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '016_sphere',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 016_sphere
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
016_sphere already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 016_sphere
adding 67 observations
Extended graph 016_sphere with new points. New model:
{'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
016_sphere already in memory (['016_sphere'])
Updating existing graph for 016_sphere
adding 20 observations
Extended graph 016_sphere with new points. New model:
{'patch_1': Model for 016_sphere:
   Contains 263 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '016_sphere',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 016_sphere
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
016_sphere already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 016_sphere
adding 68 observations
Extended graph 016_sphere with new points. New model:
{'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
016_sphere already in memory (['016_sphere'])
Updating existing graph for 016_sphere
adding 23 observations
Extended graph 016_sphere with new points. New model:
{'patch_1': Model for 016_sphere:
   Contains 286 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '016_sphere',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 016_sphere
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
016_sphere already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 016_sphere
adding 69 observations
Extended graph 016_sphere with new points. New model:
{'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
016_sphere already in memory (['016_sphere'])
Updating existing graph for 016_sphere
adding 22 observations
Extended graph 016_sphere with new points. New model:
{'patch_1': Model for 016_sphere:
   Contains 306 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '016_sphere',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_016_sphere/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x74c5f28972b0>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x74c5f2897610>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x74c5f287ec70>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x74c5f28152e0>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x74c5f2828c10>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x74c5f2831bb0>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x74c5f27f03a0>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x74c5f27f0400>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['023_mug', '023_mug', '023_mug', '023_mug', '023_mug', '023_mug', '023_mug', '023_mug', '023_mug', '023_mug', '023_mug', '023_mug', '023_mug', '023_mug'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x74c5f2788790>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x74c5f279e640>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x74c5f279e580>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_023_mug/pretrained'), 'resume_wandb_run': False, 'wandb_id': 'vpyrkk53', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_023_mug', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '023_mug',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 023_mug
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
023_mug already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 023_mug
adding 97 observations
Extended graph 023_mug with new points. New model:
{'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
023_mug not in memory ([])
Adding a new graph to memory.
init object model with id 023_mug
building graph from 35 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x74c5f1dd4a60>
Added new graph with id 023_mug to memory.
Model for 023_mug:
   Contains 35 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '023_mug',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 023_mug
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
023_mug already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 023_mug
adding 80 observations
Extended graph 023_mug with new points. New model:
{'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
023_mug already in memory (['023_mug'])
Updating existing graph for 023_mug
adding 24 observations
Extended graph 023_mug with new points. New model:
{'patch_1': Model for 023_mug:
   Contains 59 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '023_mug',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 023_mug
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
023_mug already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 023_mug
adding 92 observations
Extended graph 023_mug with new points. New model:
{'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
023_mug already in memory (['023_mug'])
Updating existing graph for 023_mug
adding 30 observations
Extended graph 023_mug with new points. New model:
{'patch_1': Model for 023_mug:
   Contains 88 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '023_mug',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 023_mug
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
023_mug already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 023_mug
adding 91 observations
Extended graph 023_mug with new points. New model:
{'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
023_mug already in memory (['023_mug'])
Updating existing graph for 023_mug
adding 45 observations
Extended graph 023_mug with new points. New model:
{'patch_1': Model for 023_mug:
   Contains 132 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '023_mug',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 023_mug
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
023_mug already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 023_mug
adding 156 observations
Extended graph 023_mug with new points. New model:
{'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
023_mug already in memory (['023_mug'])
Updating existing graph for 023_mug
adding 63 observations
Extended graph 023_mug with new points. New model:
{'patch_1': Model for 023_mug:
   Contains 194 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '023_mug',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 023_mug
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
023_mug already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 023_mug
adding 75 observations
Extended graph 023_mug with new points. New model:
{'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
023_mug already in memory (['023_mug'])
Updating existing graph for 023_mug
adding 13 observations
Extended graph 023_mug with new points. New model:
{'patch_1': Model for 023_mug:
   Contains 206 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '023_mug',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 023_mug
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
023_mug already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 023_mug
adding 90 observations
Extended graph 023_mug with new points. New model:
{'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
023_mug already in memory (['023_mug'])
Updating existing graph for 023_mug
adding 25 observations
Extended graph 023_mug with new points. New model:
{'patch_1': Model for 023_mug:
   Contains 229 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '023_mug',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 023_mug
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
023_mug already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 023_mug
adding 89 observations
Extended graph 023_mug with new points. New model:
{'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
023_mug already in memory (['023_mug'])
Updating existing graph for 023_mug
adding 24 observations
Extended graph 023_mug with new points. New model:
{'patch_1': Model for 023_mug:
   Contains 253 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '023_mug',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 023_mug
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
023_mug already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 023_mug
adding 103 observations
Extended graph 023_mug with new points. New model:
{'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
023_mug already in memory (['023_mug'])
Updating existing graph for 023_mug
adding 45 observations
Extended graph 023_mug with new points. New model:
{'patch_1': Model for 023_mug:
   Contains 297 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '023_mug',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 023_mug
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
023_mug already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 023_mug
adding 102 observations
Extended graph 023_mug with new points. New model:
{'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
023_mug already in memory (['023_mug'])
Updating existing graph for 023_mug
adding 44 observations
Extended graph 023_mug with new points. New model:
{'patch_1': Model for 023_mug:
   Contains 339 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '023_mug',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 023_mug
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
023_mug already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 023_mug
adding 86 observations
Extended graph 023_mug with new points. New model:
{'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
023_mug already in memory (['023_mug'])
Updating existing graph for 023_mug
adding 24 observations
Extended graph 023_mug with new points. New model:
{'patch_1': Model for 023_mug:
   Contains 362 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '023_mug',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 023_mug
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
023_mug already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 023_mug
adding 92 observations
Extended graph 023_mug with new points. New model:
{'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
023_mug already in memory (['023_mug'])
Updating existing graph for 023_mug
adding 29 observations
Extended graph 023_mug with new points. New model:
{'patch_1': Model for 023_mug:
   Contains 391 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '023_mug',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 023_mug
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
023_mug already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 023_mug
adding 108 observations
Extended graph 023_mug with new points. New model:
{'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
023_mug already in memory (['023_mug'])
Updating existing graph for 023_mug
adding 51 observations
Extended graph 023_mug with new points. New model:
{'patch_1': Model for 023_mug:
   Contains 441 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '023_mug',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 023_mug
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
023_mug already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Updating existing graph for 023_mug
adding 102 observations
Extended graph 023_mug with new points. New model:
{'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
023_mug already in memory (['023_mug'])
Updating existing graph for 023_mug
adding 40 observations
Extended graph 023_mug with new points. New model:
{'patch_1': Model for 023_mug:
   Contains 479 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '023_mug',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_023_mug/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x7e0f210d82e0>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x7e0f210d8640>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x7e0f21040c70>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7e0f21057310>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7e0f21069c40>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x7e0f21073be0>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x7e0f210313d0>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x7e0f21031430>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['012_cylinder_tbp_horz', '012_cylinder_tbp_horz', '012_cylinder_tbp_horz', '012_cylinder_tbp_horz', '012_cylinder_tbp_horz', '012_cylinder_tbp_horz', '012_cylinder_tbp_horz', '012_cylinder_tbp_horz', '012_cylinder_tbp_horz', '012_cylinder_tbp_horz', '012_cylinder_tbp_horz', '012_cylinder_tbp_horz', '012_cylinder_tbp_horz', '012_cylinder_tbp_horz'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x7e0f20fca7c0>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7e0f20fdf670>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7e0f20fdf5b0>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_012_cylinder_tbp_horz/pretrained'), 'resume_wandb_run': False, 'wandb_id': 'sl92729m', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_012_cylinder_tbp_horz', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '012_cylinder_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 012_cylinder_tbp_horz
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
012_cylinder_tbp_horz not in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Adding a new graph to memory.
init object model with id 012_cylinder_tbp_horz
building graph from 97 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x7e0f1ea90820>
Added new graph with id 012_cylinder_tbp_horz to memory.
Model for 012_cylinder_tbp_horz:
   Contains 97 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,


---Updating memory of learning_module_1---
012_cylinder_tbp_horz not in memory ([])
Adding a new graph to memory.
init object model with id 012_cylinder_tbp_horz
building graph from 51 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x7e0f1ea90940>
Added new graph with id 012_cylinder_tbp_horz to memory.
Model for 012_cylinder_tbp_horz:
   Contains 51 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '012_cylinder_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 012_cylinder_tbp_horz
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
012_cylinder_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 80 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_0': Model for 012_cylinder_tbp_horz:
   Contains 177 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
012_cylinder_tbp_horz already in memory (['012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 15 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_1': Model for 012_cylinder_tbp_horz:
   Contains 66 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '012_cylinder_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 012_cylinder_tbp_horz
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
012_cylinder_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 80 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_0': Model for 012_cylinder_tbp_horz:
   Contains 257 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
012_cylinder_tbp_horz already in memory (['012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 16 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_1': Model for 012_cylinder_tbp_horz:
   Contains 82 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '012_cylinder_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 012_cylinder_tbp_horz
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
012_cylinder_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 82 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_0': Model for 012_cylinder_tbp_horz:
   Contains 339 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
012_cylinder_tbp_horz already in memory (['012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 17 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_1': Model for 012_cylinder_tbp_horz:
   Contains 99 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '012_cylinder_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 012_cylinder_tbp_horz
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
012_cylinder_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 96 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_0': Model for 012_cylinder_tbp_horz:
   Contains 435 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
012_cylinder_tbp_horz already in memory (['012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 12 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_1': Model for 012_cylinder_tbp_horz:
   Contains 111 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '012_cylinder_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 012_cylinder_tbp_horz
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
012_cylinder_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 96 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_0': Model for 012_cylinder_tbp_horz:
   Contains 531 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
012_cylinder_tbp_horz already in memory (['012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 12 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_1': Model for 012_cylinder_tbp_horz:
   Contains 123 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '012_cylinder_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 012_cylinder_tbp_horz
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
012_cylinder_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 88 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_0': Model for 012_cylinder_tbp_horz:
   Contains 617 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
012_cylinder_tbp_horz already in memory (['012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 26 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_1': Model for 012_cylinder_tbp_horz:
   Contains 148 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '012_cylinder_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 012_cylinder_tbp_horz
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
012_cylinder_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 88 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_0': Model for 012_cylinder_tbp_horz:
   Contains 702 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
012_cylinder_tbp_horz already in memory (['012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 25 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_1': Model for 012_cylinder_tbp_horz:
   Contains 171 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '012_cylinder_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 012_cylinder_tbp_horz
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
012_cylinder_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 82 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_0': Model for 012_cylinder_tbp_horz:
   Contains 783 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
012_cylinder_tbp_horz already in memory (['012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 23 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_1': Model for 012_cylinder_tbp_horz:
   Contains 193 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '012_cylinder_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 012_cylinder_tbp_horz
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
012_cylinder_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 82 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_0': Model for 012_cylinder_tbp_horz:
   Contains 861 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
012_cylinder_tbp_horz already in memory (['012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 21 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_1': Model for 012_cylinder_tbp_horz:
   Contains 212 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '012_cylinder_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 012_cylinder_tbp_horz
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
012_cylinder_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 76 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_0': Model for 012_cylinder_tbp_horz:
   Contains 936 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
012_cylinder_tbp_horz already in memory (['012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 15 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_1': Model for 012_cylinder_tbp_horz:
   Contains 227 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '012_cylinder_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 012_cylinder_tbp_horz
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
012_cylinder_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 78 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_0': Model for 012_cylinder_tbp_horz:
   Contains 1009 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
012_cylinder_tbp_horz already in memory (['012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 18 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_1': Model for 012_cylinder_tbp_horz:
   Contains 245 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '012_cylinder_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 012_cylinder_tbp_horz
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
012_cylinder_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 87 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_0': Model for 012_cylinder_tbp_horz:
   Contains 1096 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
012_cylinder_tbp_horz already in memory (['012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 18 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_1': Model for 012_cylinder_tbp_horz:
   Contains 263 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '012_cylinder_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 012_cylinder_tbp_horz
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
012_cylinder_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 79 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_0': Model for 012_cylinder_tbp_horz:
   Contains 1173 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
012_cylinder_tbp_horz already in memory (['012_cylinder_tbp_horz'])
Updating existing graph for 012_cylinder_tbp_horz
adding 14 observations
Extended graph 012_cylinder_tbp_horz with new points. New model:
{'patch_1': Model for 012_cylinder_tbp_horz:
   Contains 277 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '012_cylinder_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_012_cylinder_tbp_horz/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x7975424d72b0>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x7975424d7610>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x7975424bfc70>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7975424562e0>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x797542469c10>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x797542471bb0>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x7975424303a0>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x797542430400>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['014_cylinder_numenta_horz', '014_cylinder_numenta_horz', '014_cylinder_numenta_horz', '014_cylinder_numenta_horz', '014_cylinder_numenta_horz', '014_cylinder_numenta_horz', '014_cylinder_numenta_horz', '014_cylinder_numenta_horz', '014_cylinder_numenta_horz', '014_cylinder_numenta_horz', '014_cylinder_numenta_horz', '014_cylinder_numenta_horz', '014_cylinder_numenta_horz', '014_cylinder_numenta_horz'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x7975423c9790>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7975423de640>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7975423de580>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_014_cylinder_numenta_horz/pretrained'), 'resume_wandb_run': False, 'wandb_id': '3sf30dfk', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_014_cylinder_numenta_horz', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '014_cylinder_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 014_cylinder_numenta_horz
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
014_cylinder_numenta_horz not in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Adding a new graph to memory.
init object model with id 014_cylinder_numenta_horz
building graph from 93 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x79753fece3d0>
Added new graph with id 014_cylinder_numenta_horz to memory.
Model for 014_cylinder_numenta_horz:
   Contains 93 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,


---Updating memory of learning_module_1---
014_cylinder_numenta_horz not in memory ([])
Adding a new graph to memory.
init object model with id 014_cylinder_numenta_horz
building graph from 33 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x79753fece4f0>
Added new graph with id 014_cylinder_numenta_horz to memory.
Model for 014_cylinder_numenta_horz:
   Contains 33 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '014_cylinder_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 014_cylinder_numenta_horz
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
014_cylinder_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 80 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_0': Model for 014_cylinder_numenta_horz:
   Contains 173 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
014_cylinder_numenta_horz already in memory (['014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 15 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_1': Model for 014_cylinder_numenta_horz:
   Contains 48 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '014_cylinder_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 014_cylinder_numenta_horz
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
014_cylinder_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 80 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_0': Model for 014_cylinder_numenta_horz:
   Contains 253 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
014_cylinder_numenta_horz already in memory (['014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 16 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_1': Model for 014_cylinder_numenta_horz:
   Contains 64 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '014_cylinder_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 014_cylinder_numenta_horz
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
014_cylinder_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 81 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_0': Model for 014_cylinder_numenta_horz:
   Contains 334 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
014_cylinder_numenta_horz already in memory (['014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 15 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_1': Model for 014_cylinder_numenta_horz:
   Contains 79 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '014_cylinder_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 014_cylinder_numenta_horz
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
014_cylinder_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 96 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_0': Model for 014_cylinder_numenta_horz:
   Contains 430 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
014_cylinder_numenta_horz already in memory (['014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 12 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_1': Model for 014_cylinder_numenta_horz:
   Contains 91 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '014_cylinder_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 014_cylinder_numenta_horz
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
014_cylinder_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 96 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_0': Model for 014_cylinder_numenta_horz:
   Contains 526 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
014_cylinder_numenta_horz already in memory (['014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 12 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_1': Model for 014_cylinder_numenta_horz:
   Contains 103 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '014_cylinder_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 014_cylinder_numenta_horz
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
014_cylinder_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 85 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_0': Model for 014_cylinder_numenta_horz:
   Contains 609 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
014_cylinder_numenta_horz already in memory (['014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 21 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_1': Model for 014_cylinder_numenta_horz:
   Contains 124 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '014_cylinder_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 014_cylinder_numenta_horz
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
014_cylinder_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 91 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_0': Model for 014_cylinder_numenta_horz:
   Contains 698 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
014_cylinder_numenta_horz already in memory (['014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 27 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_1': Model for 014_cylinder_numenta_horz:
   Contains 151 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '014_cylinder_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 014_cylinder_numenta_horz
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
014_cylinder_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 81 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_0': Model for 014_cylinder_numenta_horz:
   Contains 778 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
014_cylinder_numenta_horz already in memory (['014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 23 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_1': Model for 014_cylinder_numenta_horz:
   Contains 173 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '014_cylinder_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 014_cylinder_numenta_horz
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
014_cylinder_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 83 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_0': Model for 014_cylinder_numenta_horz:
   Contains 857 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
014_cylinder_numenta_horz already in memory (['014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 21 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_1': Model for 014_cylinder_numenta_horz:
   Contains 193 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '014_cylinder_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 014_cylinder_numenta_horz
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
014_cylinder_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 76 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_0': Model for 014_cylinder_numenta_horz:
   Contains 932 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
014_cylinder_numenta_horz already in memory (['014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 15 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_1': Model for 014_cylinder_numenta_horz:
   Contains 208 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '014_cylinder_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 014_cylinder_numenta_horz
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
014_cylinder_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 78 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_0': Model for 014_cylinder_numenta_horz:
   Contains 1005 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
014_cylinder_numenta_horz already in memory (['014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 18 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_1': Model for 014_cylinder_numenta_horz:
   Contains 226 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '014_cylinder_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 014_cylinder_numenta_horz
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
014_cylinder_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 87 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_0': Model for 014_cylinder_numenta_horz:
   Contains 1092 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
014_cylinder_numenta_horz already in memory (['014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 18 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_1': Model for 014_cylinder_numenta_horz:
   Contains 244 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '014_cylinder_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 014_cylinder_numenta_horz
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
014_cylinder_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 79 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_0': Model for 014_cylinder_numenta_horz:
   Contains 1169 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
014_cylinder_numenta_horz already in memory (['014_cylinder_numenta_horz'])
Updating existing graph for 014_cylinder_numenta_horz
adding 14 observations
Extended graph 014_cylinder_numenta_horz with new points. New model:
{'patch_1': Model for 014_cylinder_numenta_horz:
   Contains 258 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '014_cylinder_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_014_cylinder_numenta_horz/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x7bcd74ed71f0>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x7bcd74ed7550>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x7bcd74e40b80>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7bcd74e56220>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7bcd74e6af40>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x7bcd74e73af0>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x7bcd74e302e0>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x7bcd74e30340>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['017_sphere_tbp_horz', '017_sphere_tbp_horz', '017_sphere_tbp_horz', '017_sphere_tbp_horz', '017_sphere_tbp_horz', '017_sphere_tbp_horz', '017_sphere_tbp_horz', '017_sphere_tbp_horz', '017_sphere_tbp_horz', '017_sphere_tbp_horz', '017_sphere_tbp_horz', '017_sphere_tbp_horz', '017_sphere_tbp_horz', '017_sphere_tbp_horz'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x7bcd74dc96d0>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7bcd74dde580>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7bcd74dde4c0>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_017_sphere_tbp_horz/pretrained'), 'resume_wandb_run': False, 'wandb_id': 'j0jany7p', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_017_sphere_tbp_horz', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '017_sphere_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 017_sphere_tbp_horz
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
017_sphere_tbp_horz not in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Adding a new graph to memory.
init object model with id 017_sphere_tbp_horz
building graph from 98 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x7bcd72890ac0>
Added new graph with id 017_sphere_tbp_horz to memory.
Model for 017_sphere_tbp_horz:
   Contains 98 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,


---Updating memory of learning_module_1---
017_sphere_tbp_horz not in memory ([])
Adding a new graph to memory.
init object model with id 017_sphere_tbp_horz
building graph from 69 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x7bcd72890be0>
Added new graph with id 017_sphere_tbp_horz to memory.
Model for 017_sphere_tbp_horz:
   Contains 69 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '017_sphere_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 017_sphere_tbp_horz
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
017_sphere_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 68 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_0': Model for 017_sphere_tbp_horz:
   Contains 166 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
017_sphere_tbp_horz already in memory (['017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 23 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_1': Model for 017_sphere_tbp_horz:
   Contains 92 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '017_sphere_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 017_sphere_tbp_horz
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
017_sphere_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 79 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_0': Model for 017_sphere_tbp_horz:
   Contains 245 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
017_sphere_tbp_horz already in memory (['017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 25 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_1': Model for 017_sphere_tbp_horz:
   Contains 117 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '017_sphere_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 017_sphere_tbp_horz
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
017_sphere_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 71 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_0': Model for 017_sphere_tbp_horz:
   Contains 316 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
017_sphere_tbp_horz already in memory (['017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 20 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_1': Model for 017_sphere_tbp_horz:
   Contains 137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '017_sphere_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 017_sphere_tbp_horz
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
017_sphere_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 70 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_0': Model for 017_sphere_tbp_horz:
   Contains 386 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
017_sphere_tbp_horz already in memory (['017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 22 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_1': Model for 017_sphere_tbp_horz:
   Contains 159 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '017_sphere_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 017_sphere_tbp_horz
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
017_sphere_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 74 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_0': Model for 017_sphere_tbp_horz:
   Contains 460 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
017_sphere_tbp_horz already in memory (['017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 21 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_1': Model for 017_sphere_tbp_horz:
   Contains 180 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '017_sphere_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 017_sphere_tbp_horz
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
017_sphere_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 79 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_0': Model for 017_sphere_tbp_horz:
   Contains 539 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
017_sphere_tbp_horz already in memory (['017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 35 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_1': Model for 017_sphere_tbp_horz:
   Contains 214 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '017_sphere_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 017_sphere_tbp_horz
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
017_sphere_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 78 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_0': Model for 017_sphere_tbp_horz:
   Contains 612 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
017_sphere_tbp_horz already in memory (['017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 33 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_1': Model for 017_sphere_tbp_horz:
   Contains 246 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '017_sphere_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 017_sphere_tbp_horz
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
017_sphere_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 77 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_0': Model for 017_sphere_tbp_horz:
   Contains 688 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
017_sphere_tbp_horz already in memory (['017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 24 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_1': Model for 017_sphere_tbp_horz:
   Contains 269 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '017_sphere_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 017_sphere_tbp_horz
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
017_sphere_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 76 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_0': Model for 017_sphere_tbp_horz:
   Contains 763 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
017_sphere_tbp_horz already in memory (['017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 31 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_1': Model for 017_sphere_tbp_horz:
   Contains 296 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '017_sphere_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 017_sphere_tbp_horz
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
017_sphere_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 67 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_0': Model for 017_sphere_tbp_horz:
   Contains 830 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
017_sphere_tbp_horz already in memory (['017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 21 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_1': Model for 017_sphere_tbp_horz:
   Contains 316 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '017_sphere_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 017_sphere_tbp_horz
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
017_sphere_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 67 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_0': Model for 017_sphere_tbp_horz:
   Contains 894 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
017_sphere_tbp_horz already in memory (['017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 20 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_1': Model for 017_sphere_tbp_horz:
   Contains 336 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '017_sphere_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 017_sphere_tbp_horz
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
017_sphere_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 68 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_0': Model for 017_sphere_tbp_horz:
   Contains 960 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
017_sphere_tbp_horz already in memory (['017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 23 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_1': Model for 017_sphere_tbp_horz:
   Contains 359 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '017_sphere_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 017_sphere_tbp_horz
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
017_sphere_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 69 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_0': Model for 017_sphere_tbp_horz:
   Contains 1028 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
017_sphere_tbp_horz already in memory (['017_sphere_tbp_horz'])
Updating existing graph for 017_sphere_tbp_horz
adding 22 observations
Extended graph 017_sphere_tbp_horz with new points. New model:
{'patch_1': Model for 017_sphere_tbp_horz:
   Contains 379 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '017_sphere_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_017_sphere_tbp_horz/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x7d91d0495280>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x7d91d04955e0>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x7d91d047efa0>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7d91d04142b0>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7d91d0428be0>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x7d91d0430b80>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x7d91d03ee370>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x7d91d03ee3d0>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['019_sphere_numenta_horz', '019_sphere_numenta_horz', '019_sphere_numenta_horz', '019_sphere_numenta_horz', '019_sphere_numenta_horz', '019_sphere_numenta_horz', '019_sphere_numenta_horz', '019_sphere_numenta_horz', '019_sphere_numenta_horz', '019_sphere_numenta_horz', '019_sphere_numenta_horz', '019_sphere_numenta_horz', '019_sphere_numenta_horz', '019_sphere_numenta_horz'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x7d91d0387760>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7d91d039d610>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7d91d039d550>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_019_sphere_numenta_horz/pretrained'), 'resume_wandb_run': False, 'wandb_id': '356nu4xo', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_019_sphere_numenta_horz', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '019_sphere_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 019_sphere_numenta_horz
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
019_sphere_numenta_horz not in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Adding a new graph to memory.
init object model with id 019_sphere_numenta_horz
building graph from 92 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x7d91cd5ce610>
Added new graph with id 019_sphere_numenta_horz to memory.
Model for 019_sphere_numenta_horz:
   Contains 92 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,


---Updating memory of learning_module_1---
019_sphere_numenta_horz not in memory ([])
Adding a new graph to memory.
init object model with id 019_sphere_numenta_horz
building graph from 47 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x7d91cd5ce730>
Added new graph with id 019_sphere_numenta_horz to memory.
Model for 019_sphere_numenta_horz:
   Contains 47 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '019_sphere_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 019_sphere_numenta_horz
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
019_sphere_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 68 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_0': Model for 019_sphere_numenta_horz:
   Contains 160 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
019_sphere_numenta_horz already in memory (['019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 23 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_1': Model for 019_sphere_numenta_horz:
   Contains 70 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '019_sphere_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 019_sphere_numenta_horz
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
019_sphere_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 79 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_0': Model for 019_sphere_numenta_horz:
   Contains 239 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
019_sphere_numenta_horz already in memory (['019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 25 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_1': Model for 019_sphere_numenta_horz:
   Contains 95 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '019_sphere_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 019_sphere_numenta_horz
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
019_sphere_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 71 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_0': Model for 019_sphere_numenta_horz:
   Contains 310 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
019_sphere_numenta_horz already in memory (['019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 20 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_1': Model for 019_sphere_numenta_horz:
   Contains 115 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '019_sphere_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 019_sphere_numenta_horz
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
019_sphere_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 70 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_0': Model for 019_sphere_numenta_horz:
   Contains 380 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
019_sphere_numenta_horz already in memory (['019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 22 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_1': Model for 019_sphere_numenta_horz:
   Contains 137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '019_sphere_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 019_sphere_numenta_horz
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
019_sphere_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 74 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_0': Model for 019_sphere_numenta_horz:
   Contains 454 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
019_sphere_numenta_horz already in memory (['019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 21 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_1': Model for 019_sphere_numenta_horz:
   Contains 158 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '019_sphere_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 019_sphere_numenta_horz
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
019_sphere_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 75 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_0': Model for 019_sphere_numenta_horz:
   Contains 528 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
019_sphere_numenta_horz already in memory (['019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 27 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_1': Model for 019_sphere_numenta_horz:
   Contains 184 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '019_sphere_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 019_sphere_numenta_horz
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
019_sphere_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 75 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_0': Model for 019_sphere_numenta_horz:
   Contains 599 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
019_sphere_numenta_horz already in memory (['019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 27 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_1': Model for 019_sphere_numenta_horz:
   Contains 210 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '019_sphere_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 019_sphere_numenta_horz
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
019_sphere_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 76 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_0': Model for 019_sphere_numenta_horz:
   Contains 674 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
019_sphere_numenta_horz already in memory (['019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 26 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_1': Model for 019_sphere_numenta_horz:
   Contains 236 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '019_sphere_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 019_sphere_numenta_horz
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
019_sphere_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 75 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_0': Model for 019_sphere_numenta_horz:
   Contains 748 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
019_sphere_numenta_horz already in memory (['019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 28 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_1': Model for 019_sphere_numenta_horz:
   Contains 262 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '019_sphere_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 019_sphere_numenta_horz
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
019_sphere_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 67 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_0': Model for 019_sphere_numenta_horz:
   Contains 815 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
019_sphere_numenta_horz already in memory (['019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 21 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_1': Model for 019_sphere_numenta_horz:
   Contains 282 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '019_sphere_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 019_sphere_numenta_horz
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
019_sphere_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 67 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_0': Model for 019_sphere_numenta_horz:
   Contains 879 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
019_sphere_numenta_horz already in memory (['019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 20 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_1': Model for 019_sphere_numenta_horz:
   Contains 302 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '019_sphere_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 019_sphere_numenta_horz
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
019_sphere_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 68 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_0': Model for 019_sphere_numenta_horz:
   Contains 945 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
019_sphere_numenta_horz already in memory (['019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 23 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_1': Model for 019_sphere_numenta_horz:
   Contains 325 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '019_sphere_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 019_sphere_numenta_horz
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
019_sphere_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 69 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_0': Model for 019_sphere_numenta_horz:
   Contains 1013 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
019_sphere_numenta_horz already in memory (['019_sphere_numenta_horz'])
Updating existing graph for 019_sphere_numenta_horz
adding 22 observations
Extended graph 019_sphere_numenta_horz with new points. New model:
{'patch_1': Model for 019_sphere_numenta_horz:
   Contains 345 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '019_sphere_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_019_sphere_numenta_horz/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x79e3a6bd6310>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x79e3a6bd6670>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x79e3a6bb6cd0>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x79e3a6b55340>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x79e3a6b67c70>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x79e3a6b71c10>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x79e3a6b2f400>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x79e3a6b2f460>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['024_mug_tbp_horz', '024_mug_tbp_horz', '024_mug_tbp_horz', '024_mug_tbp_horz', '024_mug_tbp_horz', '024_mug_tbp_horz', '024_mug_tbp_horz', '024_mug_tbp_horz', '024_mug_tbp_horz', '024_mug_tbp_horz', '024_mug_tbp_horz', '024_mug_tbp_horz', '024_mug_tbp_horz', '024_mug_tbp_horz'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x79e3a6ac87f0>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x79e3a6add6a0>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x79e3a6add5e0>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_024_mug_tbp_horz/pretrained'), 'resume_wandb_run': False, 'wandb_id': 'y90qvxdt', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_024_mug_tbp_horz', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '024_mug_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 024_mug_tbp_horz
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
024_mug_tbp_horz not in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Adding a new graph to memory.
init object model with id 024_mug_tbp_horz
building graph from 119 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x79e3a4558100>
Added new graph with id 024_mug_tbp_horz to memory.
Model for 024_mug_tbp_horz:
   Contains 119 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,


---Updating memory of learning_module_1---
024_mug_tbp_horz not in memory ([])
Adding a new graph to memory.
init object model with id 024_mug_tbp_horz
building graph from 73 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x79e3a45581f0>
Added new graph with id 024_mug_tbp_horz to memory.
Model for 024_mug_tbp_horz:
   Contains 73 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '024_mug_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 024_mug_tbp_horz
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
024_mug_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 80 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_0': Model for 024_mug_tbp_horz:
   Contains 199 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
024_mug_tbp_horz already in memory (['024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 24 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_1': Model for 024_mug_tbp_horz:
   Contains 97 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '024_mug_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 024_mug_tbp_horz
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
024_mug_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 92 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_0': Model for 024_mug_tbp_horz:
   Contains 290 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
024_mug_tbp_horz already in memory (['024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 30 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_1': Model for 024_mug_tbp_horz:
   Contains 126 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '024_mug_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 024_mug_tbp_horz
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
024_mug_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 91 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_0': Model for 024_mug_tbp_horz:
   Contains 380 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
024_mug_tbp_horz already in memory (['024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 48 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_1': Model for 024_mug_tbp_horz:
   Contains 173 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '024_mug_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 024_mug_tbp_horz
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
024_mug_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 156 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_0': Model for 024_mug_tbp_horz:
   Contains 535 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
024_mug_tbp_horz already in memory (['024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 63 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_1': Model for 024_mug_tbp_horz:
   Contains 235 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '024_mug_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 024_mug_tbp_horz
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
024_mug_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 75 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_0': Model for 024_mug_tbp_horz:
   Contains 609 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
024_mug_tbp_horz already in memory (['024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 13 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_1': Model for 024_mug_tbp_horz:
   Contains 247 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '024_mug_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 024_mug_tbp_horz
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
024_mug_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 100 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_0': Model for 024_mug_tbp_horz:
   Contains 706 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
024_mug_tbp_horz already in memory (['024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 44 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_1': Model for 024_mug_tbp_horz:
   Contains 287 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '024_mug_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 024_mug_tbp_horz
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
024_mug_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 97 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_0': Model for 024_mug_tbp_horz:
   Contains 798 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
024_mug_tbp_horz already in memory (['024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 32 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_1': Model for 024_mug_tbp_horz:
   Contains 317 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '024_mug_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 024_mug_tbp_horz
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
024_mug_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 110 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_0': Model for 024_mug_tbp_horz:
   Contains 905 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
024_mug_tbp_horz already in memory (['024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 57 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_1': Model for 024_mug_tbp_horz:
   Contains 372 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '024_mug_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 024_mug_tbp_horz
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
024_mug_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 104 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_0': Model for 024_mug_tbp_horz:
   Contains 1003 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
024_mug_tbp_horz already in memory (['024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 53 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_1': Model for 024_mug_tbp_horz:
   Contains 422 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '024_mug_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 024_mug_tbp_horz
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
024_mug_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 86 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_0': Model for 024_mug_tbp_horz:
   Contains 1085 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
024_mug_tbp_horz already in memory (['024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 24 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_1': Model for 024_mug_tbp_horz:
   Contains 445 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '024_mug_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 024_mug_tbp_horz
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
024_mug_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 92 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_0': Model for 024_mug_tbp_horz:
   Contains 1171 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
024_mug_tbp_horz already in memory (['024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 29 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_1': Model for 024_mug_tbp_horz:
   Contains 474 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '024_mug_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 024_mug_tbp_horz
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
024_mug_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 108 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_0': Model for 024_mug_tbp_horz:
   Contains 1268 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
024_mug_tbp_horz already in memory (['024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 51 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_1': Model for 024_mug_tbp_horz:
   Contains 524 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '024_mug_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 024_mug_tbp_horz
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
024_mug_tbp_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 102 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_0': Model for 024_mug_tbp_horz:
   Contains 1361 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
024_mug_tbp_horz already in memory (['024_mug_tbp_horz'])
Updating existing graph for 024_mug_tbp_horz
adding 40 observations
Extended graph 024_mug_tbp_horz with new points. New model:
{'patch_1': Model for 024_mug_tbp_horz:
   Contains 562 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '024_mug_tbp_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_024_mug_tbp_horz/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x76b7bc8cf220>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x76b7bc8cf580>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x76b7bc8bfe20>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x76b7bc855250>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x76b7bc869b80>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x76b7bc872b20>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x76b7bc830310>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x76b7bc830370>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['026_mug_numenta_horz', '026_mug_numenta_horz', '026_mug_numenta_horz', '026_mug_numenta_horz', '026_mug_numenta_horz', '026_mug_numenta_horz', '026_mug_numenta_horz', '026_mug_numenta_horz', '026_mug_numenta_horz', '026_mug_numenta_horz', '026_mug_numenta_horz', '026_mug_numenta_horz', '026_mug_numenta_horz', '026_mug_numenta_horz'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x76b7bc7c8700>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x76b7bc7de5b0>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x76b7bc7de4f0>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_026_mug_numenta_horz/pretrained'), 'resume_wandb_run': False, 'wandb_id': 'izt6pxf0', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_026_mug_numenta_horz', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '026_mug_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 026_mug_numenta_horz
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
026_mug_numenta_horz not in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Adding a new graph to memory.
init object model with id 026_mug_numenta_horz
building graph from 111 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x76b7ba2d2bb0>
Added new graph with id 026_mug_numenta_horz to memory.
Model for 026_mug_numenta_horz:
   Contains 111 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,


---Updating memory of learning_module_1---
026_mug_numenta_horz not in memory ([])
Adding a new graph to memory.
init object model with id 026_mug_numenta_horz
building graph from 60 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x76b7ba2d2cd0>
Added new graph with id 026_mug_numenta_horz to memory.
Model for 026_mug_numenta_horz:
   Contains 60 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '026_mug_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 026_mug_numenta_horz
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
026_mug_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 80 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_0': Model for 026_mug_numenta_horz:
   Contains 191 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
026_mug_numenta_horz already in memory (['026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 24 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_1': Model for 026_mug_numenta_horz:
   Contains 84 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '026_mug_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 026_mug_numenta_horz
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
026_mug_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 92 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_0': Model for 026_mug_numenta_horz:
   Contains 282 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
026_mug_numenta_horz already in memory (['026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 30 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_1': Model for 026_mug_numenta_horz:
   Contains 113 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '026_mug_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 026_mug_numenta_horz
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
026_mug_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 92 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_0': Model for 026_mug_numenta_horz:
   Contains 373 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
026_mug_numenta_horz already in memory (['026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 47 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_1': Model for 026_mug_numenta_horz:
   Contains 159 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '026_mug_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 026_mug_numenta_horz
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
026_mug_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 156 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_0': Model for 026_mug_numenta_horz:
   Contains 528 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
026_mug_numenta_horz already in memory (['026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 63 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_1': Model for 026_mug_numenta_horz:
   Contains 221 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '026_mug_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 026_mug_numenta_horz
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
026_mug_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 75 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_0': Model for 026_mug_numenta_horz:
   Contains 602 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
026_mug_numenta_horz already in memory (['026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 13 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_1': Model for 026_mug_numenta_horz:
   Contains 233 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '026_mug_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 026_mug_numenta_horz
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
026_mug_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 97 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_0': Model for 026_mug_numenta_horz:
   Contains 696 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
026_mug_numenta_horz already in memory (['026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 36 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_1': Model for 026_mug_numenta_horz:
   Contains 266 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '026_mug_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 026_mug_numenta_horz
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
026_mug_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 98 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_0': Model for 026_mug_numenta_horz:
   Contains 789 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
026_mug_numenta_horz already in memory (['026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 35 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_1': Model for 026_mug_numenta_horz:
   Contains 298 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '026_mug_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 026_mug_numenta_horz
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
026_mug_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 107 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_0': Model for 026_mug_numenta_horz:
   Contains 894 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
026_mug_numenta_horz already in memory (['026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 52 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_1': Model for 026_mug_numenta_horz:
   Contains 348 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '026_mug_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 026_mug_numenta_horz
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
026_mug_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 107 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_0': Model for 026_mug_numenta_horz:
   Contains 994 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
026_mug_numenta_horz already in memory (['026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 59 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_1': Model for 026_mug_numenta_horz:
   Contains 403 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '026_mug_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 026_mug_numenta_horz
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
026_mug_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 86 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_0': Model for 026_mug_numenta_horz:
   Contains 1076 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
026_mug_numenta_horz already in memory (['026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 24 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_1': Model for 026_mug_numenta_horz:
   Contains 426 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '026_mug_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 026_mug_numenta_horz
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
026_mug_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 92 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_0': Model for 026_mug_numenta_horz:
   Contains 1162 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
026_mug_numenta_horz already in memory (['026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 29 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_1': Model for 026_mug_numenta_horz:
   Contains 455 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '026_mug_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 026_mug_numenta_horz
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
026_mug_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 108 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_0': Model for 026_mug_numenta_horz:
   Contains 1259 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
026_mug_numenta_horz already in memory (['026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 51 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_1': Model for 026_mug_numenta_horz:
   Contains 505 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '026_mug_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 026_mug_numenta_horz
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
026_mug_numenta_horz already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 102 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_0': Model for 026_mug_numenta_horz:
   Contains 1352 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
026_mug_numenta_horz already in memory (['026_mug_numenta_horz'])
Updating existing graph for 026_mug_numenta_horz
adding 40 observations
Extended graph 026_mug_numenta_horz with new points. New model:
{'patch_1': Model for 026_mug_numenta_horz:
   Contains 543 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '026_mug_numenta_horz',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_026_mug_numenta_horz/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x712fb0ed6310>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x712fb0ed6670>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x712fb0eb6cd0>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x712fb0e54340>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x712fb0e67c70>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x712fb0e70c10>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x712fb0e2f400>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x712fb0e2f460>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['003_cube_tbp_vert', '003_cube_tbp_vert', '003_cube_tbp_vert', '003_cube_tbp_vert', '003_cube_tbp_vert', '003_cube_tbp_vert', '003_cube_tbp_vert', '003_cube_tbp_vert', '003_cube_tbp_vert', '003_cube_tbp_vert', '003_cube_tbp_vert', '003_cube_tbp_vert', '003_cube_tbp_vert', '003_cube_tbp_vert'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x712fb0dc77f0>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x712fb0ddd6a0>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x712fb0ddd5e0>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_003_cube_tbp_vert/pretrained'), 'resume_wandb_run': False, 'wandb_id': 'ju8orrvo', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_003_cube_tbp_vert', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '003_cube_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 003_cube_tbp_vert
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
003_cube_tbp_vert not in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Adding a new graph to memory.
init object model with id 003_cube_tbp_vert
building graph from 112 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x712fae890ac0>
Added new graph with id 003_cube_tbp_vert to memory.
Model for 003_cube_tbp_vert:
   Contains 112 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,


---Updating memory of learning_module_1---
003_cube_tbp_vert not in memory ([])
Adding a new graph to memory.
init object model with id 003_cube_tbp_vert
building graph from 49 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x712fae890be0>
Added new graph with id 003_cube_tbp_vert to memory.
Model for 003_cube_tbp_vert:
   Contains 49 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '003_cube_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 003_cube_tbp_vert
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
003_cube_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 99 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_0': Model for 003_cube_tbp_vert:
   Contains 204 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
003_cube_tbp_vert already in memory (['003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 15 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_1': Model for 003_cube_tbp_vert:
   Contains 64 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '003_cube_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 003_cube_tbp_vert
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
003_cube_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 98 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_0': Model for 003_cube_tbp_vert:
   Contains 295 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
003_cube_tbp_vert already in memory (['003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 15 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_1': Model for 003_cube_tbp_vert:
   Contains 79 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '003_cube_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 003_cube_tbp_vert
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
003_cube_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 98 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_0': Model for 003_cube_tbp_vert:
   Contains 379 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
003_cube_tbp_vert already in memory (['003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 15 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_1': Model for 003_cube_tbp_vert:
   Contains 94 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '003_cube_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 003_cube_tbp_vert
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
003_cube_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 98 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_0': Model for 003_cube_tbp_vert:
   Contains 465 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
003_cube_tbp_vert already in memory (['003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 15 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_1': Model for 003_cube_tbp_vert:
   Contains 109 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '003_cube_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 003_cube_tbp_vert
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
003_cube_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 98 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_0': Model for 003_cube_tbp_vert:
   Contains 551 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
003_cube_tbp_vert already in memory (['003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 15 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_1': Model for 003_cube_tbp_vert:
   Contains 124 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '003_cube_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 003_cube_tbp_vert
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
003_cube_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 115 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_0': Model for 003_cube_tbp_vert:
   Contains 663 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
003_cube_tbp_vert already in memory (['003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 44 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_1': Model for 003_cube_tbp_vert:
   Contains 167 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '003_cube_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 003_cube_tbp_vert
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
003_cube_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 115 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_0': Model for 003_cube_tbp_vert:
   Contains 774 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
003_cube_tbp_vert already in memory (['003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 44 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_1': Model for 003_cube_tbp_vert:
   Contains 211 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '003_cube_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 003_cube_tbp_vert
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
003_cube_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 114 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_0': Model for 003_cube_tbp_vert:
   Contains 882 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
003_cube_tbp_vert already in memory (['003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 49 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_1': Model for 003_cube_tbp_vert:
   Contains 257 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '003_cube_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 003_cube_tbp_vert
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
003_cube_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 114 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_0': Model for 003_cube_tbp_vert:
   Contains 990 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
003_cube_tbp_vert already in memory (['003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 45 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_1': Model for 003_cube_tbp_vert:
   Contains 298 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '003_cube_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 003_cube_tbp_vert
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
003_cube_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 114 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_0': Model for 003_cube_tbp_vert:
   Contains 1100 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
003_cube_tbp_vert already in memory (['003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 44 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_1': Model for 003_cube_tbp_vert:
   Contains 340 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '003_cube_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 003_cube_tbp_vert
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
003_cube_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 114 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_0': Model for 003_cube_tbp_vert:
   Contains 1205 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
003_cube_tbp_vert already in memory (['003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 47 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_1': Model for 003_cube_tbp_vert:
   Contains 386 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '003_cube_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 003_cube_tbp_vert
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
003_cube_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 114 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_0': Model for 003_cube_tbp_vert:
   Contains 1316 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
003_cube_tbp_vert already in memory (['003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 42 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_1': Model for 003_cube_tbp_vert:
   Contains 428 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '003_cube_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 003_cube_tbp_vert
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
003_cube_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 114 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_0': Model for 003_cube_tbp_vert:
   Contains 1423 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
003_cube_tbp_vert already in memory (['003_cube_tbp_vert'])
Updating existing graph for 003_cube_tbp_vert
adding 40 observations
Extended graph 003_cube_tbp_vert with new points. New model:
{'patch_1': Model for 003_cube_tbp_vert:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '003_cube_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_003_cube_tbp_vert/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x7b347969e280>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x7b347969e5e0>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x7b347967efa0>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7b34796142b0>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7b3479628be0>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x7b3479630b80>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x7b34795ef370>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x7b34795ef3d0>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['005_cube_numenta_vert', '005_cube_numenta_vert', '005_cube_numenta_vert', '005_cube_numenta_vert', '005_cube_numenta_vert', '005_cube_numenta_vert', '005_cube_numenta_vert', '005_cube_numenta_vert', '005_cube_numenta_vert', '005_cube_numenta_vert', '005_cube_numenta_vert', '005_cube_numenta_vert', '005_cube_numenta_vert', '005_cube_numenta_vert'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x7b3479587760>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7b347959d610>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7b347959d550>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_005_cube_numenta_vert/pretrained'), 'resume_wandb_run': False, 'wandb_id': 'ws52dl18', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_005_cube_numenta_vert', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '005_cube_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 005_cube_numenta_vert
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
005_cube_numenta_vert not in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Adding a new graph to memory.
init object model with id 005_cube_numenta_vert
building graph from 114 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x7b34770509d0>
Added new graph with id 005_cube_numenta_vert to memory.
Model for 005_cube_numenta_vert:
   Contains 114 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,


---Updating memory of learning_module_1---
005_cube_numenta_vert not in memory ([])
Adding a new graph to memory.
init object model with id 005_cube_numenta_vert
building graph from 45 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x7b3477050af0>
Added new graph with id 005_cube_numenta_vert to memory.
Model for 005_cube_numenta_vert:
   Contains 45 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '005_cube_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 005_cube_numenta_vert
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
005_cube_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 99 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_0': Model for 005_cube_numenta_vert:
   Contains 206 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
005_cube_numenta_vert already in memory (['005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 15 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_1': Model for 005_cube_numenta_vert:
   Contains 60 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '005_cube_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 005_cube_numenta_vert
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
005_cube_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 98 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_0': Model for 005_cube_numenta_vert:
   Contains 297 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
005_cube_numenta_vert already in memory (['005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 15 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_1': Model for 005_cube_numenta_vert:
   Contains 75 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '005_cube_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 005_cube_numenta_vert
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
005_cube_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 98 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_0': Model for 005_cube_numenta_vert:
   Contains 381 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
005_cube_numenta_vert already in memory (['005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 15 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_1': Model for 005_cube_numenta_vert:
   Contains 90 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '005_cube_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 005_cube_numenta_vert
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
005_cube_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 98 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_0': Model for 005_cube_numenta_vert:
   Contains 467 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
005_cube_numenta_vert already in memory (['005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 15 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_1': Model for 005_cube_numenta_vert:
   Contains 105 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '005_cube_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 005_cube_numenta_vert
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
005_cube_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 98 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_0': Model for 005_cube_numenta_vert:
   Contains 553 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
005_cube_numenta_vert already in memory (['005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 15 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_1': Model for 005_cube_numenta_vert:
   Contains 120 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '005_cube_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 005_cube_numenta_vert
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
005_cube_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 115 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_0': Model for 005_cube_numenta_vert:
   Contains 665 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
005_cube_numenta_vert already in memory (['005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 41 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_1': Model for 005_cube_numenta_vert:
   Contains 161 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '005_cube_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 005_cube_numenta_vert
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
005_cube_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 115 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_0': Model for 005_cube_numenta_vert:
   Contains 776 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
005_cube_numenta_vert already in memory (['005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 42 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_1': Model for 005_cube_numenta_vert:
   Contains 203 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '005_cube_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 005_cube_numenta_vert
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
005_cube_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 114 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_0': Model for 005_cube_numenta_vert:
   Contains 884 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
005_cube_numenta_vert already in memory (['005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 47 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_1': Model for 005_cube_numenta_vert:
   Contains 249 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '005_cube_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 005_cube_numenta_vert
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
005_cube_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 114 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_0': Model for 005_cube_numenta_vert:
   Contains 992 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
005_cube_numenta_vert already in memory (['005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 45 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_1': Model for 005_cube_numenta_vert:
   Contains 292 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '005_cube_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 005_cube_numenta_vert
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
005_cube_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 114 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_0': Model for 005_cube_numenta_vert:
   Contains 1102 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
005_cube_numenta_vert already in memory (['005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 44 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_1': Model for 005_cube_numenta_vert:
   Contains 334 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '005_cube_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 005_cube_numenta_vert
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
005_cube_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 114 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_0': Model for 005_cube_numenta_vert:
   Contains 1207 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
005_cube_numenta_vert already in memory (['005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 47 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_1': Model for 005_cube_numenta_vert:
   Contains 380 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '005_cube_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 005_cube_numenta_vert
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
005_cube_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 114 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_0': Model for 005_cube_numenta_vert:
   Contains 1318 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
005_cube_numenta_vert already in memory (['005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 42 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_1': Model for 005_cube_numenta_vert:
   Contains 422 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '005_cube_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 005_cube_numenta_vert
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
005_cube_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 114 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_0': Model for 005_cube_numenta_vert:
   Contains 1425 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
005_cube_numenta_vert already in memory (['005_cube_numenta_vert'])
Updating existing graph for 005_cube_numenta_vert
adding 40 observations
Extended graph 005_cube_numenta_vert with new points. New model:
{'patch_1': Model for 005_cube_numenta_vert:
   Contains 462 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '005_cube_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_005_cube_numenta_vert/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x799fd5e5d2b0>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x799fd5e5d610>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x799fd5e3dc70>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x799fd5dd42e0>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x799fd5de7c10>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x799fd5df0bb0>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x799fd5dae3a0>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x799fd5dae400>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['008_disk_tbp_vert', '008_disk_tbp_vert', '008_disk_tbp_vert', '008_disk_tbp_vert', '008_disk_tbp_vert', '008_disk_tbp_vert', '008_disk_tbp_vert', '008_disk_tbp_vert', '008_disk_tbp_vert', '008_disk_tbp_vert', '008_disk_tbp_vert', '008_disk_tbp_vert', '008_disk_tbp_vert', '008_disk_tbp_vert'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x799fd53c7790>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x799fd53dc640>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x799fd53dc580>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_008_disk_tbp_vert/pretrained'), 'resume_wandb_run': False, 'wandb_id': 'cj844g23', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_008_disk_tbp_vert', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '008_disk_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 008_disk_tbp_vert
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
008_disk_tbp_vert not in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Adding a new graph to memory.
init object model with id 008_disk_tbp_vert
building graph from 116 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x799fd2fd1af0>
Added new graph with id 008_disk_tbp_vert to memory.
Model for 008_disk_tbp_vert:
   Contains 116 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,


---Updating memory of learning_module_1---
008_disk_tbp_vert not in memory ([])
Adding a new graph to memory.
init object model with id 008_disk_tbp_vert
building graph from 48 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x799fd2fd1c10>
Added new graph with id 008_disk_tbp_vert to memory.
Model for 008_disk_tbp_vert:
   Contains 48 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '008_disk_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 008_disk_tbp_vert
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
008_disk_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 15 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_0': Model for 008_disk_tbp_vert:
   Contains 131 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
008_disk_tbp_vert already in memory (['008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 7 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_1': Model for 008_disk_tbp_vert:
   Contains 55 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '008_disk_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 008_disk_tbp_vert
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
008_disk_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 99 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_0': Model for 008_disk_tbp_vert:
   Contains 230 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
008_disk_tbp_vert already in memory (['008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 16 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_1': Model for 008_disk_tbp_vert:
   Contains 71 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '008_disk_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 008_disk_tbp_vert
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
008_disk_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 15 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_0': Model for 008_disk_tbp_vert:
   Contains 245 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
008_disk_tbp_vert already in memory (['008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 13 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_1': Model for 008_disk_tbp_vert:
   Contains 84 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '008_disk_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 008_disk_tbp_vert
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
008_disk_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 12 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_0': Model for 008_disk_tbp_vert:
   Contains 257 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
008_disk_tbp_vert already in memory (['008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 1 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_1': Model for 008_disk_tbp_vert:
   Contains 85 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '008_disk_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 008_disk_tbp_vert
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
008_disk_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 11 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_0': Model for 008_disk_tbp_vert:
   Contains 268 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
008_disk_tbp_vert already in memory (['008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 1 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_1': Model for 008_disk_tbp_vert:
   Contains 86 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '008_disk_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 008_disk_tbp_vert
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
008_disk_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 103 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_0': Model for 008_disk_tbp_vert:
   Contains 369 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
008_disk_tbp_vert already in memory (['008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 35 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_1': Model for 008_disk_tbp_vert:
   Contains 121 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '008_disk_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 008_disk_tbp_vert
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
008_disk_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 98 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_0': Model for 008_disk_tbp_vert:
   Contains 462 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
008_disk_tbp_vert already in memory (['008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 28 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_1': Model for 008_disk_tbp_vert:
   Contains 144 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '008_disk_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 008_disk_tbp_vert
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
008_disk_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 104 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_0': Model for 008_disk_tbp_vert:
   Contains 553 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
008_disk_tbp_vert already in memory (['008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 38 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_1': Model for 008_disk_tbp_vert:
   Contains 176 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '008_disk_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 008_disk_tbp_vert
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
008_disk_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 97 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_0': Model for 008_disk_tbp_vert:
   Contains 640 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
008_disk_tbp_vert already in memory (['008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 30 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_1': Model for 008_disk_tbp_vert:
   Contains 201 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '008_disk_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 008_disk_tbp_vert
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
008_disk_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 91 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_0': Model for 008_disk_tbp_vert:
   Contains 730 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
008_disk_tbp_vert already in memory (['008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 16 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_1': Model for 008_disk_tbp_vert:
   Contains 217 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '008_disk_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 008_disk_tbp_vert
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
008_disk_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 93 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_0': Model for 008_disk_tbp_vert:
   Contains 820 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
008_disk_tbp_vert already in memory (['008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 17 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_1': Model for 008_disk_tbp_vert:
   Contains 234 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '008_disk_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 008_disk_tbp_vert
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
008_disk_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 90 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_0': Model for 008_disk_tbp_vert:
   Contains 902 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
008_disk_tbp_vert already in memory (['008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 16 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_1': Model for 008_disk_tbp_vert:
   Contains 250 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '008_disk_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 008_disk_tbp_vert
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
008_disk_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 94 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_0': Model for 008_disk_tbp_vert:
   Contains 989 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
008_disk_tbp_vert already in memory (['008_disk_tbp_vert'])
Updating existing graph for 008_disk_tbp_vert
adding 16 observations
Extended graph 008_disk_tbp_vert with new points. New model:
{'patch_1': Model for 008_disk_tbp_vert:
   Contains 266 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '008_disk_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_008_disk_tbp_vert/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x7b6131f56340>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x7b6131f566a0>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x7b6131ecd0a0>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7b6131ed5370>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7b6131ee8ca0>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x7b6131ef1c40>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x7b6131eaf430>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x7b6131eaf490>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['010_disk_numenta_vert', '010_disk_numenta_vert', '010_disk_numenta_vert', '010_disk_numenta_vert', '010_disk_numenta_vert', '010_disk_numenta_vert', '010_disk_numenta_vert', '010_disk_numenta_vert', '010_disk_numenta_vert', '010_disk_numenta_vert', '010_disk_numenta_vert', '010_disk_numenta_vert', '010_disk_numenta_vert', '010_disk_numenta_vert'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x7b6131e48820>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7b6131e5d6d0>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7b6131e5d610>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_010_disk_numenta_vert/pretrained'), 'resume_wandb_run': False, 'wandb_id': '0opa6cur', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_010_disk_numenta_vert', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '010_disk_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 010_disk_numenta_vert
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
010_disk_numenta_vert not in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Adding a new graph to memory.
init object model with id 010_disk_numenta_vert
building graph from 116 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x7b6130191b80>
Added new graph with id 010_disk_numenta_vert to memory.
Model for 010_disk_numenta_vert:
   Contains 116 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,


---Updating memory of learning_module_1---
010_disk_numenta_vert not in memory ([])
Adding a new graph to memory.
init object model with id 010_disk_numenta_vert
building graph from 48 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x7b6130191ca0>
Added new graph with id 010_disk_numenta_vert to memory.
Model for 010_disk_numenta_vert:
   Contains 48 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '010_disk_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 010_disk_numenta_vert
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
010_disk_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 15 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_0': Model for 010_disk_numenta_vert:
   Contains 131 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
010_disk_numenta_vert already in memory (['010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 7 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_1': Model for 010_disk_numenta_vert:
   Contains 55 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '010_disk_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 010_disk_numenta_vert
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
010_disk_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 99 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_0': Model for 010_disk_numenta_vert:
   Contains 230 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
010_disk_numenta_vert already in memory (['010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 16 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_1': Model for 010_disk_numenta_vert:
   Contains 71 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '010_disk_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 010_disk_numenta_vert
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
010_disk_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 15 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_0': Model for 010_disk_numenta_vert:
   Contains 245 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
010_disk_numenta_vert already in memory (['010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 13 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_1': Model for 010_disk_numenta_vert:
   Contains 84 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '010_disk_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 010_disk_numenta_vert
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
010_disk_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 12 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_0': Model for 010_disk_numenta_vert:
   Contains 257 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
010_disk_numenta_vert already in memory (['010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 1 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_1': Model for 010_disk_numenta_vert:
   Contains 85 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '010_disk_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 010_disk_numenta_vert
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
010_disk_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 11 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_0': Model for 010_disk_numenta_vert:
   Contains 268 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
010_disk_numenta_vert already in memory (['010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 1 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_1': Model for 010_disk_numenta_vert:
   Contains 86 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '010_disk_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 010_disk_numenta_vert
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
010_disk_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 101 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_0': Model for 010_disk_numenta_vert:
   Contains 367 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
010_disk_numenta_vert already in memory (['010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 30 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_1': Model for 010_disk_numenta_vert:
   Contains 114 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '010_disk_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 010_disk_numenta_vert
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
010_disk_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 99 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_0': Model for 010_disk_numenta_vert:
   Contains 461 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
010_disk_numenta_vert already in memory (['010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 27 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_1': Model for 010_disk_numenta_vert:
   Contains 138 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '010_disk_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 010_disk_numenta_vert
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
010_disk_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 103 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_0': Model for 010_disk_numenta_vert:
   Contains 549 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
010_disk_numenta_vert already in memory (['010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 30 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_1': Model for 010_disk_numenta_vert:
   Contains 164 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '010_disk_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 010_disk_numenta_vert
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
010_disk_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 95 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_0': Model for 010_disk_numenta_vert:
   Contains 634 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
010_disk_numenta_vert already in memory (['010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 26 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_1': Model for 010_disk_numenta_vert:
   Contains 185 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '010_disk_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 010_disk_numenta_vert
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
010_disk_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 91 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_0': Model for 010_disk_numenta_vert:
   Contains 724 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
010_disk_numenta_vert already in memory (['010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 16 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_1': Model for 010_disk_numenta_vert:
   Contains 201 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '010_disk_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 010_disk_numenta_vert
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
010_disk_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 93 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_0': Model for 010_disk_numenta_vert:
   Contains 814 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
010_disk_numenta_vert already in memory (['010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 17 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_1': Model for 010_disk_numenta_vert:
   Contains 218 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '010_disk_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 010_disk_numenta_vert
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
010_disk_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 90 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_0': Model for 010_disk_numenta_vert:
   Contains 896 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
010_disk_numenta_vert already in memory (['010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 16 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_1': Model for 010_disk_numenta_vert:
   Contains 234 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '010_disk_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 010_disk_numenta_vert
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
010_disk_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 94 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_0': Model for 010_disk_numenta_vert:
   Contains 983 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
010_disk_numenta_vert already in memory (['010_disk_numenta_vert'])
Updating existing graph for 010_disk_numenta_vert
adding 16 observations
Extended graph 010_disk_numenta_vert with new points. New model:
{'patch_1': Model for 010_disk_numenta_vert:
   Contains 250 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '010_disk_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_010_disk_numenta_vert/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x71ae1ddd62e0>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x71ae1ddd6640>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x71ae1ddbec70>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x71ae1dd55310>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x71ae1dd68c40>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x71ae1dd71be0>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x71ae1dd2f3d0>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x71ae1dd2f430>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['013_cylinder_tbp_vert', '013_cylinder_tbp_vert', '013_cylinder_tbp_vert', '013_cylinder_tbp_vert', '013_cylinder_tbp_vert', '013_cylinder_tbp_vert', '013_cylinder_tbp_vert', '013_cylinder_tbp_vert', '013_cylinder_tbp_vert', '013_cylinder_tbp_vert', '013_cylinder_tbp_vert', '013_cylinder_tbp_vert', '013_cylinder_tbp_vert', '013_cylinder_tbp_vert'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x71ae1dcc87c0>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x71ae1dcde670>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x71ae1dcde5b0>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_013_cylinder_tbp_vert/pretrained'), 'resume_wandb_run': False, 'wandb_id': 'a92lhgwl', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_013_cylinder_tbp_vert', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '013_cylinder_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 013_cylinder_tbp_vert
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
013_cylinder_tbp_vert not in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Adding a new graph to memory.
init object model with id 013_cylinder_tbp_vert
building graph from 106 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x71ae1b791bb0>
Added new graph with id 013_cylinder_tbp_vert to memory.
Model for 013_cylinder_tbp_vert:
   Contains 106 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,


---Updating memory of learning_module_1---
013_cylinder_tbp_vert not in memory ([])
Adding a new graph to memory.
init object model with id 013_cylinder_tbp_vert
building graph from 61 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x71ae1b791cd0>
Added new graph with id 013_cylinder_tbp_vert to memory.
Model for 013_cylinder_tbp_vert:
   Contains 61 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '013_cylinder_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 013_cylinder_tbp_vert
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
013_cylinder_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 80 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_0': Model for 013_cylinder_tbp_vert:
   Contains 186 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
013_cylinder_tbp_vert already in memory (['013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 15 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_1': Model for 013_cylinder_tbp_vert:
   Contains 76 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '013_cylinder_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 013_cylinder_tbp_vert
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
013_cylinder_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 80 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_0': Model for 013_cylinder_tbp_vert:
   Contains 266 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
013_cylinder_tbp_vert already in memory (['013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 16 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_1': Model for 013_cylinder_tbp_vert:
   Contains 92 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '013_cylinder_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 013_cylinder_tbp_vert
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
013_cylinder_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 80 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_0': Model for 013_cylinder_tbp_vert:
   Contains 346 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
013_cylinder_tbp_vert already in memory (['013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 15 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_1': Model for 013_cylinder_tbp_vert:
   Contains 107 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '013_cylinder_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 013_cylinder_tbp_vert
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
013_cylinder_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 96 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_0': Model for 013_cylinder_tbp_vert:
   Contains 442 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
013_cylinder_tbp_vert already in memory (['013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 12 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_1': Model for 013_cylinder_tbp_vert:
   Contains 119 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '013_cylinder_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 013_cylinder_tbp_vert
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
013_cylinder_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 96 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_0': Model for 013_cylinder_tbp_vert:
   Contains 538 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
013_cylinder_tbp_vert already in memory (['013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 12 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_1': Model for 013_cylinder_tbp_vert:
   Contains 131 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '013_cylinder_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 013_cylinder_tbp_vert
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
013_cylinder_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 81 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_0': Model for 013_cylinder_tbp_vert:
   Contains 617 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
013_cylinder_tbp_vert already in memory (['013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 15 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_1': Model for 013_cylinder_tbp_vert:
   Contains 146 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '013_cylinder_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 013_cylinder_tbp_vert
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
013_cylinder_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 85 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_0': Model for 013_cylinder_tbp_vert:
   Contains 701 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
013_cylinder_tbp_vert already in memory (['013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 18 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_1': Model for 013_cylinder_tbp_vert:
   Contains 164 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '013_cylinder_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 013_cylinder_tbp_vert
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
013_cylinder_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 77 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_0': Model for 013_cylinder_tbp_vert:
   Contains 777 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
013_cylinder_tbp_vert already in memory (['013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 19 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_1': Model for 013_cylinder_tbp_vert:
   Contains 183 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '013_cylinder_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 013_cylinder_tbp_vert
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
013_cylinder_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 78 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_0': Model for 013_cylinder_tbp_vert:
   Contains 851 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
013_cylinder_tbp_vert already in memory (['013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 15 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_1': Model for 013_cylinder_tbp_vert:
   Contains 197 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '013_cylinder_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 013_cylinder_tbp_vert
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
013_cylinder_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 76 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_0': Model for 013_cylinder_tbp_vert:
   Contains 926 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
013_cylinder_tbp_vert already in memory (['013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 15 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_1': Model for 013_cylinder_tbp_vert:
   Contains 212 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '013_cylinder_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 013_cylinder_tbp_vert
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
013_cylinder_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 78 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_0': Model for 013_cylinder_tbp_vert:
   Contains 999 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
013_cylinder_tbp_vert already in memory (['013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 18 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_1': Model for 013_cylinder_tbp_vert:
   Contains 230 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '013_cylinder_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 013_cylinder_tbp_vert
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
013_cylinder_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 87 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_0': Model for 013_cylinder_tbp_vert:
   Contains 1086 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
013_cylinder_tbp_vert already in memory (['013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 18 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_1': Model for 013_cylinder_tbp_vert:
   Contains 248 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '013_cylinder_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 013_cylinder_tbp_vert
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
013_cylinder_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 79 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_0': Model for 013_cylinder_tbp_vert:
   Contains 1163 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
013_cylinder_tbp_vert already in memory (['013_cylinder_tbp_vert'])
Updating existing graph for 013_cylinder_tbp_vert
adding 14 observations
Extended graph 013_cylinder_tbp_vert with new points. New model:
{'patch_1': Model for 013_cylinder_tbp_vert:
   Contains 262 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '013_cylinder_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_013_cylinder_tbp_vert/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x71d78cace370>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x71d78cace6d0>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x71d78ca4d0d0>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x71d78ca553a0>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x71d78ca67cd0>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x71d78ca71c70>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x71d78ca2f460>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x71d78ca2f4c0>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['015_cylinder_numenta_vert', '015_cylinder_numenta_vert', '015_cylinder_numenta_vert', '015_cylinder_numenta_vert', '015_cylinder_numenta_vert', '015_cylinder_numenta_vert', '015_cylinder_numenta_vert', '015_cylinder_numenta_vert', '015_cylinder_numenta_vert', '015_cylinder_numenta_vert', '015_cylinder_numenta_vert', '015_cylinder_numenta_vert', '015_cylinder_numenta_vert', '015_cylinder_numenta_vert'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x71d78c9c8850>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x71d78c9dd700>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x71d78c9dd640>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_015_cylinder_numenta_vert/pretrained'), 'resume_wandb_run': False, 'wandb_id': 'k1wzcejg', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_015_cylinder_numenta_vert', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '015_cylinder_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 015_cylinder_numenta_vert
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
015_cylinder_numenta_vert not in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Adding a new graph to memory.
init object model with id 015_cylinder_numenta_vert
building graph from 100 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x71d78a490940>
Added new graph with id 015_cylinder_numenta_vert to memory.
Model for 015_cylinder_numenta_vert:
   Contains 100 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,


---Updating memory of learning_module_1---
015_cylinder_numenta_vert not in memory ([])
Adding a new graph to memory.
init object model with id 015_cylinder_numenta_vert
building graph from 51 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x71d78a490a60>
Added new graph with id 015_cylinder_numenta_vert to memory.
Model for 015_cylinder_numenta_vert:
   Contains 51 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '015_cylinder_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 015_cylinder_numenta_vert
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
015_cylinder_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 80 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_0': Model for 015_cylinder_numenta_vert:
   Contains 180 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
015_cylinder_numenta_vert already in memory (['015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 15 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_1': Model for 015_cylinder_numenta_vert:
   Contains 66 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '015_cylinder_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 015_cylinder_numenta_vert
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
015_cylinder_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 80 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_0': Model for 015_cylinder_numenta_vert:
   Contains 260 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
015_cylinder_numenta_vert already in memory (['015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 16 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_1': Model for 015_cylinder_numenta_vert:
   Contains 82 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '015_cylinder_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 015_cylinder_numenta_vert
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
015_cylinder_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 80 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_0': Model for 015_cylinder_numenta_vert:
   Contains 340 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
015_cylinder_numenta_vert already in memory (['015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 15 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_1': Model for 015_cylinder_numenta_vert:
   Contains 97 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '015_cylinder_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 015_cylinder_numenta_vert
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
015_cylinder_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 96 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_0': Model for 015_cylinder_numenta_vert:
   Contains 436 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
015_cylinder_numenta_vert already in memory (['015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 12 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_1': Model for 015_cylinder_numenta_vert:
   Contains 109 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '015_cylinder_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 015_cylinder_numenta_vert
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
015_cylinder_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 96 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_0': Model for 015_cylinder_numenta_vert:
   Contains 532 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
015_cylinder_numenta_vert already in memory (['015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 12 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_1': Model for 015_cylinder_numenta_vert:
   Contains 121 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '015_cylinder_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 015_cylinder_numenta_vert
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
015_cylinder_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 81 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_0': Model for 015_cylinder_numenta_vert:
   Contains 611 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
015_cylinder_numenta_vert already in memory (['015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 14 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_1': Model for 015_cylinder_numenta_vert:
   Contains 135 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '015_cylinder_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 015_cylinder_numenta_vert
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
015_cylinder_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 85 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_0': Model for 015_cylinder_numenta_vert:
   Contains 695 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
015_cylinder_numenta_vert already in memory (['015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 18 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_1': Model for 015_cylinder_numenta_vert:
   Contains 153 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '015_cylinder_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 015_cylinder_numenta_vert
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
015_cylinder_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 77 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_0': Model for 015_cylinder_numenta_vert:
   Contains 771 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
015_cylinder_numenta_vert already in memory (['015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 18 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_1': Model for 015_cylinder_numenta_vert:
   Contains 171 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '015_cylinder_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 015_cylinder_numenta_vert
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
015_cylinder_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 78 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_0': Model for 015_cylinder_numenta_vert:
   Contains 845 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
015_cylinder_numenta_vert already in memory (['015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 16 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_1': Model for 015_cylinder_numenta_vert:
   Contains 186 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '015_cylinder_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 015_cylinder_numenta_vert
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
015_cylinder_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 76 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_0': Model for 015_cylinder_numenta_vert:
   Contains 920 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
015_cylinder_numenta_vert already in memory (['015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 15 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_1': Model for 015_cylinder_numenta_vert:
   Contains 201 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '015_cylinder_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 015_cylinder_numenta_vert
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
015_cylinder_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 78 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_0': Model for 015_cylinder_numenta_vert:
   Contains 993 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
015_cylinder_numenta_vert already in memory (['015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 18 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_1': Model for 015_cylinder_numenta_vert:
   Contains 219 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '015_cylinder_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 015_cylinder_numenta_vert
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
015_cylinder_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 87 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_0': Model for 015_cylinder_numenta_vert:
   Contains 1080 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
015_cylinder_numenta_vert already in memory (['015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 18 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_1': Model for 015_cylinder_numenta_vert:
   Contains 237 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '015_cylinder_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 015_cylinder_numenta_vert
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
015_cylinder_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 79 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_0': Model for 015_cylinder_numenta_vert:
   Contains 1157 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
015_cylinder_numenta_vert already in memory (['015_cylinder_numenta_vert'])
Updating existing graph for 015_cylinder_numenta_vert
adding 14 observations
Extended graph 015_cylinder_numenta_vert with new points. New model:
{'patch_1': Model for 015_cylinder_numenta_vert:
   Contains 251 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '015_cylinder_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_015_cylinder_numenta_vert/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x7ff8b09502b0>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x7ff8b0950610>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x7ff8b08c0c70>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7ff8b08d62e0>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7ff8b08ebc10>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x7ff8b08f2bb0>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x7ff8b08b13a0>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x7ff8b08b1400>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['018_sphere_tbp_vert', '018_sphere_tbp_vert', '018_sphere_tbp_vert', '018_sphere_tbp_vert', '018_sphere_tbp_vert', '018_sphere_tbp_vert', '018_sphere_tbp_vert', '018_sphere_tbp_vert', '018_sphere_tbp_vert', '018_sphere_tbp_vert', '018_sphere_tbp_vert', '018_sphere_tbp_vert', '018_sphere_tbp_vert', '018_sphere_tbp_vert'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x7ff8b084a790>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7ff8b085f640>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7ff8b085f580>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_018_sphere_tbp_vert/pretrained'), 'resume_wandb_run': False, 'wandb_id': '3bk4k7aa', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_018_sphere_tbp_vert', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '018_sphere_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 018_sphere_tbp_vert
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
018_sphere_tbp_vert not in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Adding a new graph to memory.
init object model with id 018_sphere_tbp_vert
building graph from 101 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x7ff8ae30fd60>
Added new graph with id 018_sphere_tbp_vert to memory.
Model for 018_sphere_tbp_vert:
   Contains 101 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,


---Updating memory of learning_module_1---
018_sphere_tbp_vert not in memory ([])
Adding a new graph to memory.
init object model with id 018_sphere_tbp_vert
building graph from 76 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x7ff8ae30fe80>
Added new graph with id 018_sphere_tbp_vert to memory.
Model for 018_sphere_tbp_vert:
   Contains 76 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '018_sphere_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 018_sphere_tbp_vert
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
018_sphere_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 68 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_0': Model for 018_sphere_tbp_vert:
   Contains 169 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
018_sphere_tbp_vert already in memory (['018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 23 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_1': Model for 018_sphere_tbp_vert:
   Contains 99 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '018_sphere_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 018_sphere_tbp_vert
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
018_sphere_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 79 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_0': Model for 018_sphere_tbp_vert:
   Contains 248 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
018_sphere_tbp_vert already in memory (['018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 25 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_1': Model for 018_sphere_tbp_vert:
   Contains 124 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '018_sphere_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 018_sphere_tbp_vert
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
018_sphere_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 71 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_0': Model for 018_sphere_tbp_vert:
   Contains 319 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
018_sphere_tbp_vert already in memory (['018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 20 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_1': Model for 018_sphere_tbp_vert:
   Contains 144 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '018_sphere_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 018_sphere_tbp_vert
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
018_sphere_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 70 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_0': Model for 018_sphere_tbp_vert:
   Contains 389 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
018_sphere_tbp_vert already in memory (['018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 22 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_1': Model for 018_sphere_tbp_vert:
   Contains 166 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '018_sphere_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 018_sphere_tbp_vert
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
018_sphere_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 74 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_0': Model for 018_sphere_tbp_vert:
   Contains 463 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
018_sphere_tbp_vert already in memory (['018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 21 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_1': Model for 018_sphere_tbp_vert:
   Contains 187 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '018_sphere_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 018_sphere_tbp_vert
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
018_sphere_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 72 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_0': Model for 018_sphere_tbp_vert:
   Contains 534 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
018_sphere_tbp_vert already in memory (['018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 26 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_1': Model for 018_sphere_tbp_vert:
   Contains 213 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '018_sphere_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 018_sphere_tbp_vert
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
018_sphere_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 71 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_0': Model for 018_sphere_tbp_vert:
   Contains 601 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
018_sphere_tbp_vert already in memory (['018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 25 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_1': Model for 018_sphere_tbp_vert:
   Contains 238 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '018_sphere_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 018_sphere_tbp_vert
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
018_sphere_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 71 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_0': Model for 018_sphere_tbp_vert:
   Contains 671 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
018_sphere_tbp_vert already in memory (['018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 23 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_1': Model for 018_sphere_tbp_vert:
   Contains 261 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '018_sphere_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 018_sphere_tbp_vert
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
018_sphere_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 72 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_0': Model for 018_sphere_tbp_vert:
   Contains 742 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
018_sphere_tbp_vert already in memory (['018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 23 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_1': Model for 018_sphere_tbp_vert:
   Contains 282 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '018_sphere_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 018_sphere_tbp_vert
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
018_sphere_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 67 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_0': Model for 018_sphere_tbp_vert:
   Contains 809 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
018_sphere_tbp_vert already in memory (['018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 21 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_1': Model for 018_sphere_tbp_vert:
   Contains 302 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '018_sphere_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 018_sphere_tbp_vert
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
018_sphere_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 67 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_0': Model for 018_sphere_tbp_vert:
   Contains 873 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
018_sphere_tbp_vert already in memory (['018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 20 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_1': Model for 018_sphere_tbp_vert:
   Contains 322 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '018_sphere_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 018_sphere_tbp_vert
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
018_sphere_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 68 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_0': Model for 018_sphere_tbp_vert:
   Contains 939 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
018_sphere_tbp_vert already in memory (['018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 23 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_1': Model for 018_sphere_tbp_vert:
   Contains 345 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '018_sphere_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 018_sphere_tbp_vert
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
018_sphere_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 69 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_0': Model for 018_sphere_tbp_vert:
   Contains 1007 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
018_sphere_tbp_vert already in memory (['018_sphere_tbp_vert'])
Updating existing graph for 018_sphere_tbp_vert
adding 22 observations
Extended graph 018_sphere_tbp_vert with new points. New model:
{'patch_1': Model for 018_sphere_tbp_vert:
   Contains 365 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '018_sphere_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_018_sphere_tbp_vert/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x766cf2957250>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x766cf29575b0>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x766cf28c0dc0>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x766cf28d6280>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x766cf28eabb0>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x766cf28f2b50>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x766cf28b0340>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x766cf28b03a0>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['020_sphere_numenta_vert', '020_sphere_numenta_vert', '020_sphere_numenta_vert', '020_sphere_numenta_vert', '020_sphere_numenta_vert', '020_sphere_numenta_vert', '020_sphere_numenta_vert', '020_sphere_numenta_vert', '020_sphere_numenta_vert', '020_sphere_numenta_vert', '020_sphere_numenta_vert', '020_sphere_numenta_vert', '020_sphere_numenta_vert', '020_sphere_numenta_vert'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x766cf2849730>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x766cf285e5e0>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x766cf285e520>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_020_sphere_numenta_vert/pretrained'), 'resume_wandb_run': False, 'wandb_id': '1oqt3pnv', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_020_sphere_numenta_vert', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '020_sphere_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 020_sphere_numenta_vert
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
020_sphere_numenta_vert not in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Adding a new graph to memory.
init object model with id 020_sphere_numenta_vert
building graph from 91 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x766cf030f700>
Added new graph with id 020_sphere_numenta_vert to memory.
Model for 020_sphere_numenta_vert:
   Contains 91 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,


---Updating memory of learning_module_1---
020_sphere_numenta_vert not in memory ([])
Adding a new graph to memory.
init object model with id 020_sphere_numenta_vert
building graph from 54 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x766cf030f820>
Added new graph with id 020_sphere_numenta_vert to memory.
Model for 020_sphere_numenta_vert:
   Contains 54 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '020_sphere_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 020_sphere_numenta_vert
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
020_sphere_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 68 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_0': Model for 020_sphere_numenta_vert:
   Contains 159 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
020_sphere_numenta_vert already in memory (['020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 23 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_1': Model for 020_sphere_numenta_vert:
   Contains 77 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '020_sphere_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 020_sphere_numenta_vert
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
020_sphere_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 79 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_0': Model for 020_sphere_numenta_vert:
   Contains 238 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
020_sphere_numenta_vert already in memory (['020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 25 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_1': Model for 020_sphere_numenta_vert:
   Contains 102 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '020_sphere_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 020_sphere_numenta_vert
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
020_sphere_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 71 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_0': Model for 020_sphere_numenta_vert:
   Contains 309 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
020_sphere_numenta_vert already in memory (['020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 20 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_1': Model for 020_sphere_numenta_vert:
   Contains 122 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '020_sphere_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 020_sphere_numenta_vert
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
020_sphere_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 70 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_0': Model for 020_sphere_numenta_vert:
   Contains 379 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
020_sphere_numenta_vert already in memory (['020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 22 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_1': Model for 020_sphere_numenta_vert:
   Contains 144 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '020_sphere_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 020_sphere_numenta_vert
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
020_sphere_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 74 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_0': Model for 020_sphere_numenta_vert:
   Contains 453 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
020_sphere_numenta_vert already in memory (['020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 21 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_1': Model for 020_sphere_numenta_vert:
   Contains 165 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '020_sphere_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 020_sphere_numenta_vert
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
020_sphere_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 71 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_0': Model for 020_sphere_numenta_vert:
   Contains 523 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
020_sphere_numenta_vert already in memory (['020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 24 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_1': Model for 020_sphere_numenta_vert:
   Contains 188 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '020_sphere_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 020_sphere_numenta_vert
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
020_sphere_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 71 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_0': Model for 020_sphere_numenta_vert:
   Contains 590 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
020_sphere_numenta_vert already in memory (['020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 24 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_1': Model for 020_sphere_numenta_vert:
   Contains 212 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '020_sphere_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 020_sphere_numenta_vert
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
020_sphere_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 71 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_0': Model for 020_sphere_numenta_vert:
   Contains 660 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
020_sphere_numenta_vert already in memory (['020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 22 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_1': Model for 020_sphere_numenta_vert:
   Contains 234 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '020_sphere_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 020_sphere_numenta_vert
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
020_sphere_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 72 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_0': Model for 020_sphere_numenta_vert:
   Contains 731 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
020_sphere_numenta_vert already in memory (['020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 21 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_1': Model for 020_sphere_numenta_vert:
   Contains 254 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '020_sphere_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 020_sphere_numenta_vert
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
020_sphere_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 67 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_0': Model for 020_sphere_numenta_vert:
   Contains 798 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
020_sphere_numenta_vert already in memory (['020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 21 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_1': Model for 020_sphere_numenta_vert:
   Contains 274 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '020_sphere_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 020_sphere_numenta_vert
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
020_sphere_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 67 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_0': Model for 020_sphere_numenta_vert:
   Contains 862 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
020_sphere_numenta_vert already in memory (['020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 20 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_1': Model for 020_sphere_numenta_vert:
   Contains 294 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '020_sphere_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 020_sphere_numenta_vert
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
020_sphere_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 68 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_0': Model for 020_sphere_numenta_vert:
   Contains 928 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
020_sphere_numenta_vert already in memory (['020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 23 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_1': Model for 020_sphere_numenta_vert:
   Contains 317 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '020_sphere_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 020_sphere_numenta_vert
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
020_sphere_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 69 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_0': Model for 020_sphere_numenta_vert:
   Contains 996 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
020_sphere_numenta_vert already in memory (['020_sphere_numenta_vert'])
Updating existing graph for 020_sphere_numenta_vert
adding 22 observations
Extended graph 020_sphere_numenta_vert with new points. New model:
{'patch_1': Model for 020_sphere_numenta_vert:
   Contains 337 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '020_sphere_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_020_sphere_numenta_vert/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x7015720952e0>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x701572095640>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x70157207dc70>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x701572014310>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x701572028c40>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x701572030be0>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x701571fef3d0>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x701571fef430>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['025_mug_tbp_vert', '025_mug_tbp_vert', '025_mug_tbp_vert', '025_mug_tbp_vert', '025_mug_tbp_vert', '025_mug_tbp_vert', '025_mug_tbp_vert', '025_mug_tbp_vert', '025_mug_tbp_vert', '025_mug_tbp_vert', '025_mug_tbp_vert', '025_mug_tbp_vert', '025_mug_tbp_vert', '025_mug_tbp_vert'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x701571f887c0>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x701571f9d670>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x701571f9d5b0>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_025_mug_tbp_vert/pretrained'), 'resume_wandb_run': False, 'wandb_id': 'o3v1zqaa', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_025_mug_tbp_vert', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '025_mug_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 025_mug_tbp_vert
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
025_mug_tbp_vert not in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Adding a new graph to memory.
init object model with id 025_mug_tbp_vert
building graph from 126 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x70157021e2b0>
Added new graph with id 025_mug_tbp_vert to memory.
Model for 025_mug_tbp_vert:
   Contains 126 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,


---Updating memory of learning_module_1---
025_mug_tbp_vert not in memory ([])
Adding a new graph to memory.
init object model with id 025_mug_tbp_vert
building graph from 77 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x70157021e3d0>
Added new graph with id 025_mug_tbp_vert to memory.
Model for 025_mug_tbp_vert:
   Contains 77 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '025_mug_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 025_mug_tbp_vert
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
025_mug_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 80 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_0': Model for 025_mug_tbp_vert:
   Contains 206 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
025_mug_tbp_vert already in memory (['025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 24 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_1': Model for 025_mug_tbp_vert:
   Contains 101 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '025_mug_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 025_mug_tbp_vert
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
025_mug_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 92 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_0': Model for 025_mug_tbp_vert:
   Contains 297 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
025_mug_tbp_vert already in memory (['025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 30 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_1': Model for 025_mug_tbp_vert:
   Contains 130 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '025_mug_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 025_mug_tbp_vert
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
025_mug_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 91 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_0': Model for 025_mug_tbp_vert:
   Contains 387 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
025_mug_tbp_vert already in memory (['025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 45 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_1': Model for 025_mug_tbp_vert:
   Contains 174 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '025_mug_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 025_mug_tbp_vert
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
025_mug_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 156 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_0': Model for 025_mug_tbp_vert:
   Contains 541 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
025_mug_tbp_vert already in memory (['025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 63 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_1': Model for 025_mug_tbp_vert:
   Contains 236 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '025_mug_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 025_mug_tbp_vert
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
025_mug_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 75 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_0': Model for 025_mug_tbp_vert:
   Contains 615 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
025_mug_tbp_vert already in memory (['025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 13 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_1': Model for 025_mug_tbp_vert:
   Contains 248 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '025_mug_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 025_mug_tbp_vert
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
025_mug_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 93 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_0': Model for 025_mug_tbp_vert:
   Contains 705 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
025_mug_tbp_vert already in memory (['025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 32 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_1': Model for 025_mug_tbp_vert:
   Contains 277 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '025_mug_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 025_mug_tbp_vert
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
025_mug_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 93 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_0': Model for 025_mug_tbp_vert:
   Contains 795 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
025_mug_tbp_vert already in memory (['025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 28 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_1': Model for 025_mug_tbp_vert:
   Contains 304 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '025_mug_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 025_mug_tbp_vert
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
025_mug_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 105 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_0': Model for 025_mug_tbp_vert:
   Contains 897 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
025_mug_tbp_vert already in memory (['025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 49 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_1': Model for 025_mug_tbp_vert:
   Contains 351 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '025_mug_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 025_mug_tbp_vert
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
025_mug_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 102 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_0': Model for 025_mug_tbp_vert:
   Contains 995 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
025_mug_tbp_vert already in memory (['025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 45 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_1': Model for 025_mug_tbp_vert:
   Contains 394 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '025_mug_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 025_mug_tbp_vert
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
025_mug_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 86 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_0': Model for 025_mug_tbp_vert:
   Contains 1077 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
025_mug_tbp_vert already in memory (['025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 24 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_1': Model for 025_mug_tbp_vert:
   Contains 417 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '025_mug_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 025_mug_tbp_vert
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
025_mug_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 92 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_0': Model for 025_mug_tbp_vert:
   Contains 1163 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
025_mug_tbp_vert already in memory (['025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 29 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_1': Model for 025_mug_tbp_vert:
   Contains 446 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '025_mug_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 025_mug_tbp_vert
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
025_mug_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 108 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_0': Model for 025_mug_tbp_vert:
   Contains 1260 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
025_mug_tbp_vert already in memory (['025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 51 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_1': Model for 025_mug_tbp_vert:
   Contains 496 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '025_mug_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 025_mug_tbp_vert
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
025_mug_tbp_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 102 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_0': Model for 025_mug_tbp_vert:
   Contains 1353 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
025_mug_tbp_vert already in memory (['025_mug_tbp_vert'])
Updating existing graph for 025_mug_tbp_vert
adding 40 observations
Extended graph 025_mug_tbp_vert with new points. New model:
{'patch_1': Model for 025_mug_tbp_vert:
   Contains 534 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '025_mug_tbp_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_025_mug_tbp_vert/pretrained
logger initialized
{'do_eval': False, 'do_train': True, 'monty_config': {'monty_class': <class 'tbp.monty.frameworks.models.evidence_matching.model.MontyForEvidenceGraphMatching'>, 'monty_args': {'num_exploratory_steps': 1000, 'min_eval_steps': 3, 'min_train_steps': 3, 'max_total_steps': 2500}, 'motor_system_config': <tbp.monty.frameworks.models.motor_system.MotorSystem object at 0x7fd8cb0542b0>, 'learning_modules': {'learning_module_0': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x7fd8cb054610>, 'learning_module_1': <tbp.monty.frameworks.models.evidence_matching.learning_module.EvidenceGraphLM object at 0x7fd8cb03dc70>}, 'sensor_modules': {'sensor_module_0': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7fd8cafd32e0>, 'sensor_module_1': <tbp.monty.frameworks.models.sensor_modules.CameraSM object at 0x7fd8cafe8c10>, 'sensor_module_2': <tbp.monty.frameworks.models.sensor_modules.Probe object at 0x7fd8cafefbb0>}, 'sm_to_agent_dict': {'patch_0': 'agent_id_0', 'patch_1': 'agent_id_0', 'view_finder': 'agent_id_0'}, 'sm_to_lm_matrix': [[0], [1]], 'lm_to_lm_matrix': None, 'lm_to_lm_vote_matrix': None}, 'environment': {'env_init_args': {'objects': [{'name': 'coneSolid', 'position': [0.0, 1.5, -0.1]}], 'scene_id': None, 'seed': 42, 'data_path': '/home/hlee/tbp/data/habitat/objects/compositional_objects', 'agents': {'agent_args': {'agent_id': 'agent_id_0', 'sensor_ids': ['patch_0', 'patch_1', 'view_finder'], 'height': 0.0, 'position': [0.0, 1.5, 0.2], 'resolutions': [[64, 64], [64, 64], [64, 64]], 'positions': [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], 'rotations': [[1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0], [1.0, 0.0, 0.0, 0.0]], 'semantics': [False, False, False], 'zooms': [10.0, 5.0, 1.0]}, 'agent_type': <class 'tbp.monty.simulators.habitat.agents.MultiSensorAgent'>}}, 'env_init_func': <class 'tbp.monty.simulators.habitat.environment.HabitatEnvironment'>, 'transform': [<tbp.monty.frameworks.environment_utils.transforms.MissingToMaxDepth object at 0x7fd8cafae3a0>, <tbp.monty.frameworks.environment_utils.transforms.DepthTo3DLocations object at 0x7fd8cafae400>]}, 'train_env_interface_args': {'parent_to_child_mapping': None, 'object_names': ['027_mug_numenta_vert', '027_mug_numenta_vert', '027_mug_numenta_vert', '027_mug_numenta_vert', '027_mug_numenta_vert', '027_mug_numenta_vert', '027_mug_numenta_vert', '027_mug_numenta_vert', '027_mug_numenta_vert', '027_mug_numenta_vert', '027_mug_numenta_vert', '027_mug_numenta_vert', '027_mug_numenta_vert', '027_mug_numenta_vert'], 'object_init_sampler': <tbp.monty.frameworks.environments.object_init_samplers.Predefined object at 0x7fd8ca5c7790>, 'positioning_procedures': [<tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7fd8ca5dc640>, <tbp.monty.frameworks.environments.positioning_procedures.GetGoodViewFactory object at 0x7fd8ca5dc580>]}, 'train_env_interface_class': <class 'tbp.monty.experiment.environment.OneObjectPerEpisodeInterface'>, 'logging': {'monty_log_level': 'SILENT', 'monty_handlers': [], 'wandb_handlers': [], 'python_log_level': 'INFO', 'python_log_to_file': True, 'python_log_to_stderr': True, 'output_dir': PosixPath('/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_027_mug_numenta_vert/pretrained'), 'resume_wandb_run': False, 'wandb_id': 'r4exmrmb', 'wandb_group': '20260630_compositional_configs', 'run_name': 'supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_027_mug_numenta_vert', 'log_parallel_wandb': False}, 'show_sensor_output': False, 'max_train_steps': 1000, 'max_eval_steps': 500, 'max_total_steps': 6000, 'n_train_epochs': 1, 'n_eval_epochs': 3, 'model_name_or_path': '/home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_2d_children/pretrained/', 'min_lms_match': 1, 'seed': 42, 'supervised_lm_ids': 'all'}
loading models
loading 001_cube with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 001_cube:
   Contains 1411 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 006_disk with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 006_disk:
   Contains 949 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 011_cylinder with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 011_cylinder:
   Contains 1137 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 016_sphere with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 016_sphere:
   Contains 969 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 023_mug with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 023_mug:
   Contains 1315 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 021_logo_tbp with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 021_logo_tbp:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading 022_logo_numenta with features from dict_keys(['patch_0'])
Loaded {'patch_0': Model for 022_logo_numenta:
   Contains 377 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
} for patch_0
loading models
running train epoch 0
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '027_mug_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 027_mug_numenta_vert
running train epoch 0 train episode 0
resetting RNG to seed 1060955053
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
027_mug_numenta_vert not in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta'])
Adding a new graph to memory.
init object model with id 027_mug_numenta_vert
building graph from 112 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x7fd8c81d3c10>
Added new graph with id 027_mug_numenta_vert to memory.
Model for 027_mug_numenta_vert:
   Contains 112 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,


---Updating memory of learning_module_1---
027_mug_numenta_vert not in memory ([])
Adding a new graph to memory.
init object model with id 027_mug_numenta_vert
building graph from 58 observations
built graph <tbp.monty.frameworks.utils.object_model_utils.NumpyGraph object at 0x7fd8c81d3d00>
Added new graph with id 027_mug_numenta_vert to memory.
Model for 027_mug_numenta_vert:
   Contains 58 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,



Going from 0 to 1 of 14
New primary target: {'euler_rotation': [0, 90, 0],
 'object': '027_mug_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.        , 0.70710678, 0.        ]),
 'rotation': (0.7071067811865476, 0.0, 0.7071067811865475, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 027_mug_numenta_vert
running train epoch 0 train episode 1
resetting RNG to seed 1260428609
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
027_mug_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 80 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_0': Model for 027_mug_numenta_vert:
   Contains 192 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
027_mug_numenta_vert already in memory (['027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 24 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_1': Model for 027_mug_numenta_vert:
   Contains 82 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 1 to 2 of 14
New primary target: {'euler_rotation': [0, 180, 0],
 'object': '027_mug_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([6.123234e-17, 0.000000e+00, 1.000000e+00, 0.000000e+00]),
 'rotation': (6.123233995736766e-17, 0.0, 1.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 027_mug_numenta_vert
running train epoch 0 train episode 2
resetting RNG to seed 3997959863
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
027_mug_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 92 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_0': Model for 027_mug_numenta_vert:
   Contains 283 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
027_mug_numenta_vert already in memory (['027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 30 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_1': Model for 027_mug_numenta_vert:
   Contains 111 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 2 to 3 of 14
New primary target: {'euler_rotation': [0, 270, 0],
 'object': '027_mug_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.70710678,  0.        ,  0.70710678,  0.        ]),
 'rotation': (-0.7071067811865475, 0.0, 0.7071067811865476, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 027_mug_numenta_vert
running train epoch 0 train episode 3
resetting RNG to seed 99937259
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
027_mug_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 91 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_0': Model for 027_mug_numenta_vert:
   Contains 373 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
027_mug_numenta_vert already in memory (['027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 45 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_1': Model for 027_mug_numenta_vert:
   Contains 155 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 3 to 4 of 14
New primary target: {'euler_rotation': [90, 0, 0],
 'object': '027_mug_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([0.70710678, 0.70710678, 0.        , 0.        ]),
 'rotation': (0.7071067811865476, 0.7071067811865475, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 027_mug_numenta_vert
running train epoch 0 train episode 4
resetting RNG to seed 560786102
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
027_mug_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 156 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_0': Model for 027_mug_numenta_vert:
   Contains 527 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
027_mug_numenta_vert already in memory (['027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 63 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_1': Model for 027_mug_numenta_vert:
   Contains 217 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 4 to 5 of 14
New primary target: {'euler_rotation': [90, 180, 0],
 'object': '027_mug_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 4.32978028e-17,  4.32978028e-17,  7.07106781e-01, -7.07106781e-01]),
 'rotation': (4.329780281177467e-17,
              4.329780281177466e-17,
              0.7071067811865476,
              -0.7071067811865475),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 027_mug_numenta_vert
running train epoch 0 train episode 5
resetting RNG to seed 1953320932
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
027_mug_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 75 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_0': Model for 027_mug_numenta_vert:
   Contains 601 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
027_mug_numenta_vert already in memory (['027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 13 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_1': Model for 027_mug_numenta_vert:
   Contains 229 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 5 to 6 of 14
New primary target: {'euler_rotation': [35, 45, 0],
 'object': '027_mug_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957,  0.27781593,  0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              0.27781593346944056,
              0.36497167621709875,
              -0.11507512748638375),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 027_mug_numenta_vert
running train epoch 0 train episode 6
resetting RNG to seed 2492039860
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
027_mug_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 92 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_0': Model for 027_mug_numenta_vert:
   Contains 689 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
027_mug_numenta_vert already in memory (['027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 29 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_1': Model for 027_mug_numenta_vert:
   Contains 255 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 6 to 7 of 14
New primary target: {'euler_rotation': [325, 45, 0],
 'object': '027_mug_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957,  0.27781593, -0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              0.2778159334694408,
              -0.3649716762170987,
              -0.11507512748638384),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 027_mug_numenta_vert
running train epoch 0 train episode 7
resetting RNG to seed 2201955992
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
027_mug_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 92 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_0': Model for 027_mug_numenta_vert:
   Contains 778 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
027_mug_numenta_vert already in memory (['027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 25 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_1': Model for 027_mug_numenta_vert:
   Contains 279 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 7 to 8 of 14
New primary target: {'euler_rotation': [35, 315, 0],
 'object': '027_mug_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.88111957, -0.27781593,  0.36497168, -0.11507513]),
 'rotation': (-0.8811195706053617,
              -0.27781593346944056,
              0.36497167621709886,
              -0.11507512748638378),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 027_mug_numenta_vert
running train epoch 0 train episode 8
resetting RNG to seed 306583017
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
027_mug_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 105 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_0': Model for 027_mug_numenta_vert:
   Contains 880 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
027_mug_numenta_vert already in memory (['027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 44 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_1': Model for 027_mug_numenta_vert:
   Contains 322 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 8 to 9 of 14
New primary target: {'euler_rotation': [325, 315, 0],
 'object': '027_mug_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.88111957, -0.27781593, -0.36497168, -0.11507513]),
 'rotation': (0.8811195706053617,
              -0.2778159334694408,
              -0.3649716762170988,
              -0.11507512748638386),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 027_mug_numenta_vert
running train epoch 0 train episode 9
resetting RNG to seed 583356519
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
027_mug_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 102 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_0': Model for 027_mug_numenta_vert:
   Contains 978 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
027_mug_numenta_vert already in memory (['027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 46 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_1': Model for 027_mug_numenta_vert:
   Contains 366 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 9 to 10 of 14
New primary target: {'euler_rotation': [35, 135, 0],
 'object': '027_mug_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168,  0.11507513,  0.88111957, -0.27781593]),
 'rotation': (0.3649716762170988,
              0.11507512748638377,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 027_mug_numenta_vert
running train epoch 0 train episode 10
resetting RNG to seed 1177517334
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
027_mug_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 86 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_0': Model for 027_mug_numenta_vert:
   Contains 1060 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
027_mug_numenta_vert already in memory (['027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 24 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_1': Model for 027_mug_numenta_vert:
   Contains 389 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 10 to 11 of 14
New primary target: {'euler_rotation': [325, 135, 0],
 'object': '027_mug_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168,  0.11507513, -0.88111957, -0.27781593]),
 'rotation': (-0.36497167621709875,
              0.11507512748638385,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 027_mug_numenta_vert
running train epoch 0 train episode 11
resetting RNG to seed 111915506
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
027_mug_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 92 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_0': Model for 027_mug_numenta_vert:
   Contains 1146 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
027_mug_numenta_vert already in memory (['027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 29 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_1': Model for 027_mug_numenta_vert:
   Contains 418 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 11 to 12 of 14
New primary target: {'euler_rotation': [35, 225, 0],
 'object': '027_mug_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([-0.36497168, -0.11507513,  0.88111957, -0.27781593]),
 'rotation': (-0.3649716762170987,
              -0.11507512748638374,
              0.8811195706053617,
              -0.27781593346944056),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 027_mug_numenta_vert
running train epoch 0 train episode 12
resetting RNG to seed 1538901097
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
027_mug_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 108 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_0': Model for 027_mug_numenta_vert:
   Contains 1243 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
027_mug_numenta_vert already in memory (['027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 51 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_1': Model for 027_mug_numenta_vert:
   Contains 468 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 12 to 13 of 14
New primary target: {'euler_rotation': [325, 225, 0],
 'object': '027_mug_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([ 0.36497168, -0.11507513, -0.88111957, -0.27781593]),
 'rotation': (0.3649716762170987,
              -0.11507512748638382,
              -0.8811195706053617,
              -0.2778159334694408),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
Running a simulation to model object: 027_mug_numenta_vert
running train epoch 0 train episode 13
resetting RNG to seed 678633877
Going into exploratory mode after 0 steps

---Updating memory of learning_module_0---
027_mug_numenta_vert already in memory (['001_cube', '006_disk', '011_cylinder', '016_sphere', '023_mug', '021_logo_tbp', '022_logo_numenta', '027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 102 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_0': Model for 027_mug_numenta_vert:
   Contains 1336 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 19])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           min_depth - 12:13,
           mean_depth - 13:14,
           hsv - 14:17,
           principal_curvatures_log - 17:19,
}

---Updating memory of learning_module_1---
027_mug_numenta_vert already in memory (['027_mug_numenta_vert'])
Updating existing graph for 027_mug_numenta_vert
adding 40 observations
Extended graph 027_mug_numenta_vert with new points. New model:
{'patch_1': Model for 027_mug_numenta_vert:
   Contains 506 points in graph.
   Feature grid shape: torch.Size([200, 200, 200, 17])
   Stored features and their indexes:
           pose_vectors - 0:9,
           pose_fully_defined - 9:10,
           on_object - 10:11,
           object_coverage - 11:12,
           hsv - 12:15,
           principal_curvatures_log - 15:17,
}


Going from 13 to 0 of 14
New primary target: {'euler_rotation': [0, 0, 0],
 'object': '027_mug_numenta_vert',
 'position': (0.0, 1.5, 0.0),
 'quat_rotation': array([1., 0., 0., 0.]),
 'rotation': (1.0, 0.0, 0.0, 0.0),
 'scale': (1.0, 1.0, 1.0),
 'semantic_id': 1}
saving model to /home/hlee/tbp/results/compositional_pretraining_20260630/supervised_pre_training_objects_with_stickers_monolithic_models/supervised_pre_training_objects_with_stickers_monolithic_models-parallel_train_episode_027_mug_numenta_vert/pretrained
