{ "codebase_version": "v3.0", "fps": 20, "features": { "observation.images.robot0_eye_in_hand": { "dtype": "video", "shape": [ 256, 256, 3 ], "names": [ "height", "width", "channel" ], "video_info": { "video.fps": 20, "video.codec": "h264", "video.pix_fmt": "yuv420p", "video.is_depth_map": false, "has_audio": false }, "info": { "video.height": 256, "video.width": 256, "video.codec": "h264", "video.pix_fmt": "yuv420p", "video.is_depth_map": false, "video.fps": 20, "video.channels": 3, "has_audio": false, "video.fast_decode": null, "video.preset": null, "video.g": null, "video.extra_options": null, "video.video_backend": "pyav", "video.crf": null } }, "observation.images.robot0_agentview_left": { "dtype": "video", "shape": [ 256, 256, 3 ], "names": [ "height", "width", "channel" ], "video_info": { "video.fps": 20, "video.codec": "h264", "video.pix_fmt": "yuv420p", "video.is_depth_map": false, "has_audio": false }, "info": { "video.height": 256, "video.width": 256, "video.codec": "h264", "video.pix_fmt": "yuv420p", "video.is_depth_map": false, "video.fps": 20, "video.channels": 3, "has_audio": false, "video.fast_decode": null, "video.preset": null, "video.g": null, "video.extra_options": null, "video.video_backend": "pyav", "video.crf": null } }, "observation.images.robot0_agentview_right": { "dtype": "video", "shape": [ 256, 256, 3 ], "names": [ "height", "width", "channel" ], "video_info": { "video.fps": 20, "video.codec": "h264", "video.pix_fmt": "yuv420p", "video.is_depth_map": false, "has_audio": false }, "info": { "video.height": 256, "video.width": 256, "video.codec": "h264", "video.pix_fmt": "yuv420p", "video.is_depth_map": false, "video.fps": 20, "video.channels": 3, "has_audio": false, "video.fast_decode": null, "video.preset": null, "video.g": null, "video.extra_options": null, "video.video_backend": "pyav", "video.crf": null } }, "annotation.human.task_description": { "dtype": "int64", "shape": [ 1 ], "fps": 20 }, "annotation.human.task_name": { "dtype": "int64", "shape": [ 1 ], "fps": 20 }, "observation.state": { "dtype": "float64", "shape": [ 16 ], "fps": 20 }, "action": { "dtype": "float64", "shape": [ 12 ], "fps": 20 }, "next.reward": { "dtype": "float32", "shape": [ 1 ], "fps": 20 }, "next.done": { "dtype": "bool", "shape": [ 1 ], "fps": 20 }, "timestamp": { "dtype": "float32", "shape": [ 1 ], "names": null }, "frame_index": { "dtype": "int64", "shape": [ 1 ], "names": null }, "episode_index": { "dtype": "int64", "shape": [ 1 ], "names": null }, "index": { "dtype": "int64", "shape": [ 1 ], "names": null }, "task_index": { "dtype": "int64", "shape": [ 1 ], "names": null }, "language_persistent": { "dtype": "language", "shape": [ 1 ], "names": null }, "language_events": { "dtype": "language", "shape": [ 1 ], "names": null } }, "total_episodes": 2031, "total_frames": 559517, "total_tasks": 11, "chunks_size": 1000, "data_files_size_in_mb": 100, "video_files_size_in_mb": 200, "data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet", "video_path": "videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4", "robot_type": "PandaOmron", "splits": { "train": "0:2031" }, "tools": [ { "type": "function", "function": { "name": "say", "description": "Speak a short utterance to the user via the TTS executor.", "parameters": { "type": "object", "properties": { "text": { "type": "string", "description": "The verbatim text to speak." } }, "required": [ "text" ] } } } ] }