Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,19 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+data_files/core_datasets/coffee/demo_src_coffee_task_D0/demo.hdf5 filter=lfs diff=lfs merge=lfs -text
+data_files/core_datasets/coffee/demo_src_coffee_task_D0/demo_failed.hdf5 filter=lfs diff=lfs merge=lfs -text
+data_files/core_datasets/coffee/demo_src_coffee_task_D0/playback_demo_src_coffee_task_D0.mp4 filter=lfs diff=lfs merge=lfs -text
+data_files/core_datasets/coffee/demo_src_coffee_task_D0/playback_demo_src_coffee_task_D0_failed.mp4 filter=lfs diff=lfs merge=lfs -text
+data_files/core_datasets/square/demo_src_square_task_D0/demo.hdf5 filter=lfs diff=lfs merge=lfs -text
+data_files/core_datasets/square/demo_src_square_task_D0/demo_failed.hdf5 filter=lfs diff=lfs merge=lfs -text
+data_files/core_datasets/square/demo_src_square_task_D0/playback_demo_src_square_task_D0.mp4 filter=lfs diff=lfs merge=lfs -text
+data_files/core_datasets/square/demo_src_square_task_D0/playback_demo_src_square_task_D0_failed.mp4 filter=lfs diff=lfs merge=lfs -text
+data_files/core_datasets/stack_three/demo_src_stack_three_task_D0/demo.hdf5 filter=lfs diff=lfs merge=lfs -text
+data_files/core_datasets/stack_three/demo_src_stack_three_task_D0/demo_failed.hdf5 filter=lfs diff=lfs merge=lfs -text
+data_files/core_datasets/stack_three/demo_src_stack_three_task_D0/playback_demo_src_stack_three_task_D0.mp4 filter=lfs diff=lfs merge=lfs -text
+data_files/core_datasets/stack_three/demo_src_stack_three_task_D0/playback_demo_src_stack_three_task_D0_failed.mp4 filter=lfs diff=lfs merge=lfs -text
+data_files/core_datasets/three_piece_assembly/demo_src_three_piece_assembly_task_D0/demo.hdf5 filter=lfs diff=lfs merge=lfs -text
+data_files/core_datasets/three_piece_assembly/demo_src_three_piece_assembly_task_D0/demo_failed.hdf5 filter=lfs diff=lfs merge=lfs -text
+data_files/core_datasets/three_piece_assembly/demo_src_three_piece_assembly_task_D0/playback_demo_src_three_piece_assembly_task_D0.mp4 filter=lfs diff=lfs merge=lfs -text
+data_files/core_datasets/three_piece_assembly/demo_src_three_piece_assembly_task_D0/playback_demo_src_three_piece_assembly_task_D0_failed.mp4 filter=lfs diff=lfs merge=lfs -text

data_files/core_datasets/coffee/demo_src_coffee_task_D0/demo.hdf5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2fc7604d6e07650ffa0d2bf577a0e93303e5405ced78a41b2d6b21109a214cc1
+size 4412320005

data_files/core_datasets/coffee/demo_src_coffee_task_D0/demo_failed.hdf5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:230a57603c97a6e168088aa78a5e4eba68a066f9bf0b31580bc90e2583877878
+size 377295350

data_files/core_datasets/coffee/demo_src_coffee_task_D0/playback_demo_src_coffee_task_D0.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:311ca5545391872b441708622320cfca4b21c4e9625e2613487678e839c2386b
+size 864751

data_files/core_datasets/coffee/demo_src_coffee_task_D0/playback_demo_src_coffee_task_D0_failed.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:57f651a24a9152cfda0884431b644f8ee2c13cc30de2b1b3a38dbb880b9508f4
+size 2112836

data_files/core_datasets/square/demo_src_square_task_D0/demo.hdf5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1ec0ac48c1905bd5f6286c3e51b87d103b5b9c53dd4687f7941190ef104b1623
+size 4217820654

data_files/core_datasets/square/demo_src_square_task_D0/demo_failed.hdf5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1ce39c97fa06226a3d8cfdd9bc0e04838445c3674902d7c3302fe1fe02a9401a
+size 357640726

data_files/core_datasets/square/demo_src_square_task_D0/playback_demo_src_square_task_D0.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:183c160b9a13d6efcd8caf405639e40b44068679df23adfdda6d4a027c760f88
+size 794761

data_files/core_datasets/square/demo_src_square_task_D0/playback_demo_src_square_task_D0_failed.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5c8ac43ff90c1130233140f4f8b74e35b9d01b6f22533a9444ef1699ffcff90c
+size 1816861

data_files/core_datasets/stack_three/demo_src_stack_three_task_D0/demo.hdf5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e42e47342fc7f4b551e3144feaec2a5b93693be1413c6451bd7ec1b02171a6c
+size 6727027862

data_files/core_datasets/stack_three/demo_src_stack_three_task_D0/demo_failed.hdf5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a613e84bbef67c2988c9a8026a76ed5243ce23eb317903bbccf5bae3332b5241
+size 566136387

data_files/core_datasets/stack_three/demo_src_stack_three_task_D0/playback_demo_src_stack_three_task_D0.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6db09d4f8d4baa5127c24584a896c1067a188b306197a275ab5af243e1217abc
+size 1330658

data_files/core_datasets/stack_three/demo_src_stack_three_task_D0/playback_demo_src_stack_three_task_D0_failed.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8a50e9867ea7e53777ca9b31876d88dc96bf6383dd1e341fa5dc4dfaa4c3b5b7
+size 3388300

data_files/core_datasets/three_piece_assembly/demo_src_three_piece_assembly_task_D0/demo.hdf5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:54eb5cdde12f3b81964af2c5ad8959c5b195e484d6918a9d66d6b2d2603f3506
+size 8757077583

data_files/core_datasets/three_piece_assembly/demo_src_three_piece_assembly_task_D0/demo_failed.hdf5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7308392cef2b161be5158a3c1649a3a14d18d36c290269ab16a543cfddf1b080
+size 728144620

data_files/core_datasets/three_piece_assembly/demo_src_three_piece_assembly_task_D0/playback_demo_src_three_piece_assembly_task_D0.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d2fb3cea52682d99da5eddb708eef65295f970e621bd5d176384e5d708cd5729
+size 1341633

data_files/core_datasets/three_piece_assembly/demo_src_three_piece_assembly_task_D0/playback_demo_src_three_piece_assembly_task_D0_failed.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e0a9ab24c961528b1c610519194025273c8b2e2b1f602ea7c949b4968321e80f
+size 3650934

data_files/core_train_configs/bc_rnn_image_ds_coffee_D0_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_coffee_D0_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 400,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "./data_files/core_datasets/coffee/demo_src_coffee_task_D0/demo.hdf5",
+        "output_dir": "./tmp_files/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "coffee_D0",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_coffee_D1_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_coffee_D1_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 400,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/coffee/demo_src_coffee_task_D1/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "coffee_D1",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_coffee_D2_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_coffee_D2_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 400,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/coffee/demo_src_coffee_task_D2/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "coffee_D2",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_coffee_preparation_D0_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_coffee_preparation_D0_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 800,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/coffee_preparation/demo_src_coffee_preparation_task_D0/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "coffee_preparation_D0",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_coffee_preparation_D1_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_coffee_preparation_D1_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 800,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/coffee_preparation/demo_src_coffee_preparation_task_D1/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "coffee_preparation_D1",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_hammer_cleanup_D0_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_hammer_cleanup_D0_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 500,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/hammer_cleanup/demo_src_hammer_cleanup_task_D0/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "hammer_cleanup_D0",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_hammer_cleanup_D1_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_hammer_cleanup_D1_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 500,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/hammer_cleanup/demo_src_hammer_cleanup_task_D1/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "hammer_cleanup_D1",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_kitchen_D0_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_kitchen_D0_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 800,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/kitchen/demo_src_kitchen_task_D0/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "kitchen_D0",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_kitchen_D1_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_kitchen_D1_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 800,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/kitchen/demo_src_kitchen_task_D1/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "kitchen_D1",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_mug_cleanup_D0_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_mug_cleanup_D0_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 500,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/mug_cleanup/demo_src_mug_cleanup_task_D0/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "mug_cleanup_D0",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_mug_cleanup_D1_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_mug_cleanup_D1_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 500,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/mug_cleanup/demo_src_mug_cleanup_task_D1/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "mug_cleanup_D1",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_mug_cleanup_O1_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_mug_cleanup_O1_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 500,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/mug_cleanup/demo_src_mug_cleanup_task_O1/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "mug_cleanup_O1",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_mug_cleanup_O2_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_mug_cleanup_O2_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 500,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/mug_cleanup/demo_src_mug_cleanup_task_O2/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "mug_cleanup_O2",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_nut_assembly_D0_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_nut_assembly_D0_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 500,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/nut_assembly/demo_src_nut_assembly_task_D0/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "nut_assembly_D0",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_pick_place_D0_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_pick_place_D0_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 1000,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/pick_place/demo_src_pick_place_task_D0/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "pick_place_D0",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_square_D0_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_square_D0_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 400,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "./data_files/core_datasets/square/demo_src_square_task_D0/demo.hdf5",
+        "output_dir": "./tmp_files/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "square_D0",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_square_D1_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_square_D1_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 400,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/square/demo_src_square_task_D1/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "square_D1",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_square_D2_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_square_D2_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 400,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/square/demo_src_square_task_D2/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "square_D2",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_stack_D0_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_stack_D0_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 400,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/stack/demo_src_stack_task_D0/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "stack_D0",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_stack_D1_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_stack_D1_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 400,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/stack/demo_src_stack_task_D1/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "stack_D1",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_stack_three_D0_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_stack_three_D0_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 400,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "./data_files/core_datasets/stack_three/demo_src_stack_three_task_D0/demo.hdf5",
+        "output_dir": "./tmp_files/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "stack_three_D0",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_stack_three_D1_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_stack_three_D1_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 400,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/stack_three/demo_src_stack_three_task_D1/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "stack_three_D1",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_threading_D0_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_threading_D0_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 400,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/threading/demo_src_threading_task_D0/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "threading_D0",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_threading_D1_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_threading_D1_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 400,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/threading/demo_src_threading_task_D1/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "threading_D1",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_threading_D2_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_threading_D2_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 400,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/threading/demo_src_threading_task_D2/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "threading_D2",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_three_piece_assembly_D0_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_three_piece_assembly_D0_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 500,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "./data_files/core_datasets/three_piece_assembly/demo_src_three_piece_assembly_task_D0/demo.hdf5",
+        "output_dir": "./tmp_files/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "three_piece_assembly_D0",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_three_piece_assembly_D1_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_three_piece_assembly_D1_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 500,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/three_piece_assembly/demo_src_three_piece_assembly_task_D1/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "three_piece_assembly_D1",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_image_ds_three_piece_assembly_D2_seed_101.json

@@ -0,0 +1,241 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_image_ds_three_piece_assembly_D2_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 20,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 500,
+        "validation_epoch_every_n_steps": 50,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 500,
+            "rate": 20,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/three_piece_assembly/demo_src_three_piece_assembly_task_D2/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 16,
+        "num_epochs": 600,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.0001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 1000,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos"
+                ],
+                "rgb": [
+                    "agentview_image",
+                    "robot0_eye_in_hand_image"
+                ],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {
+                    "feature_dimension": 64,
+                    "flatten": true,
+                    "backbone_class": "ResNet18Conv",
+                    "backbone_kwargs": {
+                        "pretrained": false,
+                        "input_coord_conv": false
+                    },
+                    "pool_class": "SpatialSoftmax",
+                    "pool_kwargs": {
+                        "num_kp": 32,
+                        "learnable_temperature": false,
+                        "temperature": 1.0,
+                        "noise_std": 0.0,
+                        "output_variance": false
+                    }
+                },
+                "obs_randomizer_class": "CropRandomizer",
+                "obs_randomizer_kwargs": {
+                    "crop_height": 76,
+                    "crop_width": 76,
+                    "num_crops": 1,
+                    "pos_enc": false
+                }
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "three_piece_assembly_D2",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_low_dim_ds_coffee_D0_seed_101.json

@@ -0,0 +1,218 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_low_dim_ds_coffee_D0_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 50,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 100,
+        "validation_epoch_every_n_steps": 10,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 400,
+            "rate": 50,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/coffee/demo_src_coffee_task_D0/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 100,
+        "num_epochs": 2000,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 400,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos",
+                    "object"
+                ],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "coffee_D0",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_low_dim_ds_coffee_D1_seed_101.json

@@ -0,0 +1,218 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_low_dim_ds_coffee_D1_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 50,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 100,
+        "validation_epoch_every_n_steps": 10,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 400,
+            "rate": 50,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/coffee/demo_src_coffee_task_D1/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 100,
+        "num_epochs": 2000,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 400,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos",
+                    "object"
+                ],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "coffee_D1",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_low_dim_ds_coffee_D2_seed_101.json

@@ -0,0 +1,218 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_low_dim_ds_coffee_D2_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 50,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 100,
+        "validation_epoch_every_n_steps": 10,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 400,
+            "rate": 50,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/coffee/demo_src_coffee_task_D2/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 100,
+        "num_epochs": 2000,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 400,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos",
+                    "object"
+                ],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "coffee_D2",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_low_dim_ds_coffee_preparation_D0_seed_101.json

@@ -0,0 +1,218 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_low_dim_ds_coffee_preparation_D0_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 50,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 100,
+        "validation_epoch_every_n_steps": 10,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 800,
+            "rate": 50,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/coffee_preparation/demo_src_coffee_preparation_task_D0/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 100,
+        "num_epochs": 2000,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 400,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos",
+                    "object"
+                ],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "coffee_preparation_D0",
+            101
+        ]
+    }
+}

data_files/core_train_configs/bc_rnn_low_dim_ds_coffee_preparation_D1_seed_101.json

@@ -0,0 +1,218 @@
+{
+    "algo_name": "bc",
+    "experiment": {
+        "name": "bc_rnn_low_dim_ds_coffee_preparation_D1_seed_101",
+        "validate": false,
+        "logging": {
+            "terminal_output_to_txt": true,
+            "log_tb": true,
+            "log_wandb": false,
+            "wandb_proj_name": "debug"
+        },
+        "save": {
+            "enabled": true,
+            "every_n_seconds": null,
+            "every_n_epochs": 50,
+            "epochs": [],
+            "on_best_validation": false,
+            "on_best_rollout_return": false,
+            "on_best_rollout_success_rate": true
+        },
+        "epoch_every_n_steps": 100,
+        "validation_epoch_every_n_steps": 10,
+        "env": null,
+        "additional_envs": null,
+        "render": false,
+        "render_video": true,
+        "keep_all_videos": false,
+        "video_skip": 5,
+        "rollout": {
+            "enabled": true,
+            "n": 50,
+            "horizon": 800,
+            "rate": 50,
+            "warmstart": 0,
+            "terminate_on_success": true
+        }
+    },
+    "train": {
+        "data": "/tmp/core_datasets/coffee_preparation/demo_src_coffee_preparation_task_D1/demo.hdf5",
+        "output_dir": "/tmp/core_training_results",
+        "num_data_workers": 4,
+        "hdf5_cache_mode": "low_dim",
+        "hdf5_use_swmr": true,
+        "hdf5_load_next_obs": false,
+        "hdf5_normalize_obs": false,
+        "hdf5_filter_key": null,
+        "hdf5_validation_filter_key": null,
+        "seq_length": 10,
+        "pad_seq_length": true,
+        "frame_stack": 1,
+        "pad_frame_stack": true,
+        "dataset_keys": [
+            "actions",
+            "rewards",
+            "dones"
+        ],
+        "goal_mode": null,
+        "cuda": true,
+        "batch_size": 100,
+        "num_epochs": 2000,
+        "seed": 101
+    },
+    "algo": {
+        "optim_params": {
+            "policy": {
+                "optimizer_type": "adam",
+                "learning_rate": {
+                    "initial": 0.001,
+                    "decay_factor": 0.1,
+                    "epoch_schedule": [],
+                    "scheduler_type": "multistep"
+                },
+                "regularization": {
+                    "L2": 0.0
+                }
+            }
+        },
+        "loss": {
+            "l2_weight": 1.0,
+            "l1_weight": 0.0,
+            "cos_weight": 0.0
+        },
+        "actor_layer_dims": [],
+        "gaussian": {
+            "enabled": false,
+            "fixed_std": false,
+            "init_std": 0.1,
+            "min_std": 0.01,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "gmm": {
+            "enabled": true,
+            "num_modes": 5,
+            "min_std": 0.0001,
+            "std_activation": "softplus",
+            "low_noise_eval": true
+        },
+        "vae": {
+            "enabled": false,
+            "latent_dim": 14,
+            "latent_clip": null,
+            "kl_weight": 1.0,
+            "decoder": {
+                "is_conditioned": true,
+                "reconstruction_sum_across_elements": false
+            },
+            "prior": {
+                "learn": false,
+                "is_conditioned": false,
+                "use_gmm": false,
+                "gmm_num_modes": 10,
+                "gmm_learn_weights": false,
+                "use_categorical": false,
+                "categorical_dim": 10,
+                "categorical_gumbel_softmax_hard": false,
+                "categorical_init_temp": 1.0,
+                "categorical_temp_anneal_step": 0.001,
+                "categorical_min_temp": 0.3
+            },
+            "encoder_layer_dims": [
+                300,
+                400
+            ],
+            "decoder_layer_dims": [
+                300,
+                400
+            ],
+            "prior_layer_dims": [
+                300,
+                400
+            ]
+        },
+        "rnn": {
+            "enabled": true,
+            "horizon": 10,
+            "hidden_dim": 400,
+            "rnn_type": "LSTM",
+            "num_layers": 2,
+            "open_loop": false,
+            "kwargs": {
+                "bidirectional": false
+            }
+        },
+        "transformer": {
+            "enabled": false,
+            "context_length": 10,
+            "embed_dim": 512,
+            "num_layers": 6,
+            "num_heads": 8,
+            "emb_dropout": 0.1,
+            "attn_dropout": 0.1,
+            "block_output_dropout": 0.1,
+            "sinusoidal_embedding": false,
+            "activation": "gelu",
+            "supervise_all_steps": false,
+            "nn_parameter_for_timesteps": true
+        }
+    },
+    "observation": {
+        "modalities": {
+            "obs": {
+                "low_dim": [
+                    "robot0_eef_pos",
+                    "robot0_eef_quat",
+                    "robot0_gripper_qpos",
+                    "object"
+                ],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            },
+            "goal": {
+                "low_dim": [],
+                "rgb": [],
+                "depth": [],
+                "scan": []
+            }
+        },
+        "encoder": {
+            "low_dim": {
+                "core_class": null,
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "rgb": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "depth": {
+                "core_class": "VisualCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            },
+            "scan": {
+                "core_class": "ScanCore",
+                "core_kwargs": {},
+                "obs_randomizer_class": null,
+                "obs_randomizer_kwargs": {}
+            }
+        }
+    },
+    "meta": {
+        "hp_base_config_file": "/tmp/core_train_configs/bc.json",
+        "hp_keys": [
+            "ds",
+            "seed"
+        ],
+        "hp_values": [
+            "coffee_preparation_D1",
+            101
+        ]
+    }
+}