maximellerbach HF Staff
Drop the bbox image size step: bbox targets are [0, 1] image fractions
4eb6dcc verified | { | |
| "name": "policy_preprocessor", | |
| "steps": [ | |
| { | |
| "registry_name": "render_runtime_messages_processor", | |
| "config": { | |
| "recipe": { | |
| "messages": [ | |
| { | |
| "role": "user", | |
| "content": "${task}", | |
| "stream": "low_level", | |
| "target": false, | |
| "if_present": null, | |
| "tool_calls_from": null | |
| }, | |
| { | |
| "role": "assistant", | |
| "content": "BBoxJSON: ${bbox}", | |
| "stream": "low_level", | |
| "target": true, | |
| "if_present": "bbox", | |
| "tool_calls_from": null | |
| }, | |
| { | |
| "role": "assistant", | |
| "content": "Subtask: ${subtask}", | |
| "stream": "low_level", | |
| "target": true, | |
| "if_present": "subtask", | |
| "tool_calls_from": null | |
| } | |
| ], | |
| "bindings": { | |
| "bbox": "emitted_at(t, style=vqa, role=assistant, camera=observation.images.exterior)" | |
| }, | |
| "blend": null, | |
| "weight": null, | |
| "route": null | |
| } | |
| } | |
| }, | |
| { | |
| "registry_name": "render_training_messages_processor", | |
| "config": { | |
| "recipe": null | |
| } | |
| }, | |
| { | |
| "registry_name": "rename_observations_processor", | |
| "config": { | |
| "rename_map": {} | |
| } | |
| }, | |
| { | |
| "registry_name": "to_batch_processor", | |
| "config": {} | |
| }, | |
| { | |
| "registry_name": "g05_image_transform", | |
| "config": { | |
| "camera_order": [ | |
| "observation.images.head_rgb", | |
| "observation.images.left_wrist_rgb", | |
| "observation.images.right_wrist_rgb" | |
| ], | |
| "camera_sizes": { | |
| "observation.images.head_rgb": [ | |
| 256, | |
| 256 | |
| ], | |
| "observation.images.left_wrist_rgb": [ | |
| 256, | |
| 256 | |
| ], | |
| "observation.images.right_wrist_rgb": [ | |
| 256, | |
| 256 | |
| ] | |
| }, | |
| "mean": [ | |
| 0.5, | |
| 0.5, | |
| 0.5 | |
| ], | |
| "std": [ | |
| 0.5, | |
| 0.5, | |
| 0.5 | |
| ], | |
| "optional_camera_keys": [] | |
| } | |
| }, | |
| { | |
| "registry_name": "g05_action_operation_mask", | |
| "config": { | |
| "action_parts": [ | |
| [ | |
| "left_arm", | |
| 6 | |
| ], | |
| [ | |
| "left_gripper", | |
| 1 | |
| ], | |
| [ | |
| "right_arm", | |
| 6 | |
| ], | |
| [ | |
| "right_gripper", | |
| 1 | |
| ] | |
| ], | |
| "joint_threshold": null, | |
| "gripper_threshold": null, | |
| "velocity_threshold": 0.1, | |
| "eef_threshold": 0.001, | |
| "dim_thresholds": { | |
| "left_arm": [ | |
| 0.00756, | |
| 0.00662, | |
| 0.0432, | |
| 0.00732, | |
| 0.00493, | |
| 0.00612 | |
| ], | |
| "right_arm": [ | |
| 0.0115, | |
| 0.00717, | |
| 0.0619, | |
| 0.012, | |
| 0.00737, | |
| 0.00918 | |
| ] | |
| } | |
| } | |
| }, | |
| { | |
| "registry_name": "g05_relative_joint_actions", | |
| "config": { | |
| "enabled": true, | |
| "exclude_joints": [ | |
| "gripper" | |
| ], | |
| "action_names": [ | |
| "left_arm.0", | |
| "left_arm.1", | |
| "left_arm.2", | |
| "left_arm.3", | |
| "left_arm.4", | |
| "left_arm.5", | |
| "left_gripper.0", | |
| "right_arm.0", | |
| "right_arm.1", | |
| "right_arm.2", | |
| "right_arm.3", | |
| "right_arm.4", | |
| "right_arm.5", | |
| "right_gripper.0" | |
| ], | |
| "num_obs_steps": 6 | |
| } | |
| }, | |
| { | |
| "registry_name": "g05_embodiment_projection", | |
| "config": { | |
| "embodiment": "galaxea_r1lite", | |
| "policy_state_dim": 27, | |
| "policy_action_dim": 27, | |
| "camera_order": [ | |
| "observation.images.head_rgb", | |
| "observation.images.left_wrist_rgb", | |
| "observation.images.right_wrist_rgb" | |
| ] | |
| } | |
| }, | |
| { | |
| "registry_name": "normalizer_processor", | |
| "config": { | |
| "eps": 1e-08, | |
| "features": { | |
| "observation.state": { | |
| "type": "STATE", | |
| "shape": [ | |
| 27 | |
| ] | |
| }, | |
| "observation.images.head_rgb": { | |
| "type": "VISUAL", | |
| "shape": [ | |
| 3, | |
| 256, | |
| 256 | |
| ] | |
| }, | |
| "observation.images.left_wrist_rgb": { | |
| "type": "VISUAL", | |
| "shape": [ | |
| 3, | |
| 256, | |
| 256 | |
| ] | |
| }, | |
| "observation.images.right_wrist_rgb": { | |
| "type": "VISUAL", | |
| "shape": [ | |
| 3, | |
| 256, | |
| 256 | |
| ] | |
| }, | |
| "action": { | |
| "type": "ACTION", | |
| "shape": [ | |
| 27 | |
| ] | |
| } | |
| }, | |
| "norm_map": { | |
| "STATE": "MEAN_STD", | |
| "ACTION": "MEAN_STD", | |
| "VISUAL": "IDENTITY" | |
| } | |
| }, | |
| "state_file": "policy_preprocessor_step_7_normalizer_processor.safetensors" | |
| }, | |
| { | |
| "registry_name": "g05_normalization_clamp", | |
| "config": { | |
| "minimum": -5.0, | |
| "maximum": 5.0 | |
| } | |
| }, | |
| { | |
| "registry_name": "device_processor", | |
| "config": { | |
| "device": "cuda", | |
| "float_dtype": null | |
| } | |
| }, | |
| { | |
| "registry_name": "g05_tokenizer", | |
| "config": { | |
| "processor_dir": "hf_processor", | |
| "action_tokenizer_path": "action_tokenizer.safetensors" | |
| }, | |
| "artifacts": { | |
| "processor_dir": "hf_processor", | |
| "action_tokenizer_path": "action_tokenizer.safetensors" | |
| } | |
| } | |
| ] | |
| } |