{ "schema_version": 1, "repo_id": "wzlllong/value-model-benchmark-training-data", "release_format": "LeRobot Dataset v3.0", "documentation": { "en": "README.md", "zh": "README_zh.md" }, "domains": [ { "domain": "simulation", "name_en": "Simulation", "name_zh": "仿真世界", "subsets": [ "simulation-full-shot-data", "simulation-one-shot-data", "simulation-test-30" ], "platform": "ARX" }, { "domain": "real-world", "name_en": "Real-World", "name_zh": "真实世界", "subsets": [ "real-world-full-shot-data", "real-world-one-shot-data", "real-world-test-30" ], "platform": "AgiBot Genie02" } ], "available_subsets": [ { "subset_id": "simulation-full-shot-data", "domain": "simulation", "shot_regime": "full-shot", "availability": "available", "tasks": 15, "episodes_per_task": 100, "episodes": 1500, "path_in_repo": "training-data/simulation/full-shot", "name_en": "Simulation Full-Shot Training Data", "name_zh": "仿真 Full-Shot 训练集", "format": "LeRobot v3.0", "robot_type": "unified_robot", "frames": 789612, "fps": 25, "camera_streams_per_episode": 3, "state_dimensions": 14, "action_dimensions": 14, "splits": { "train": "0:1500" }, "reference_repo": "RoboDojo-Benchmark/RoboDojo", "video_codecs": [ "av1", "h264" ], "videos": 4500, "platform": "ARX", "source_description_en": "Based on open-source RoboDojo data reviewed by our team, with selected demonstrations recorded again.", "source_description_zh": "基于 RoboDojo 开源数据，经团队审核后，对部分示例重新采集。", "display_task_names": [ "Arrange Largest Number", "Fill Pen Holder", "Fold Clothes", "Hang Mugs", "Insert Tubes", "Make Toast", "Organize Table", "Pack Objects Into Box", "Play Stacking Toy", "Press By Number", "Put Bottles Into Dustbin", "Stack Blocks", "Store Laptop And Headphones", "Swap Blocks", "Sweep Blocks" ], "display_task_names_zh": [ "排列数字组成最大数", "装填笔筒", "折叠衣物", "悬挂马克杯", "插入管子", "制作吐司", "整理桌面", "将物品装入盒中", "操作堆叠玩具", "依照数字按按钮", "将瓶子放入垃圾桶", "堆叠积木", "收纳笔记本电脑与耳机", "交换积木位置", "清扫积木" ], "role": "training", "size_bytes": 45467875256, "size_scope": "entire subset directory, including annotations, documentation and integrity files", "subtasks": { "definitions": "training-data/simulation/full-shot/annotations/subtasks.json", "task_count": 15, "definition_count": 58, "temporal_annotations_available": false, "revision": "3f21eb1beeedaf1b1d5aaacda4a033174521d8c3" }, "task_keys": [ "arrange_largest_number", "fill_pen_holder", "fold_clothes", "hang_mugs", "insert_tubes", "make_toast", "organize_table", "pack_objects_into_box", "play_stacking_toy", "press_by_number", "put_bottles_into_dustbin", "stack_blocks", "store_laptop_and_headphones", "swap_blocks", "sweep_blocks" ], "conversion": "training-data/simulation/full-shot/meta/conversion.json" }, { "subset_id": "simulation-one-shot-data", "domain": "simulation", "shot_regime": "one-shot", "availability": "available", "tasks": 15, "episodes_per_task": 1, "episodes": 15, "path_in_repo": "training-data/simulation/one-shot", "name_en": "Simulation One-Shot Training Data", "name_zh": "仿真 One-Shot 训练集", "format": "LeRobot v3.0", "robot_type": "unified_robot", "frames": 8080, "fps": 25, "camera_streams_per_episode": 3, "state_dimensions": 14, "action_dimensions": 14, "splits": { "train": "0:15" }, "video_codecs": [ "av1" ], "videos": 45, "platform": "ARX", "source_description_en": "Based on open-source RoboDojo data reviewed by our team, with selected demonstrations recorded again.", "source_description_zh": "基于 RoboDojo 开源数据，经团队审核后，对部分示例重新采集。", "display_task_names": [ "Arrange Largest Number", "Fill Pen Holder", "Fold Clothes", "Hang Mugs", "Insert Tubes", "Make Toast", "Organize Table", "Pack Objects Into Box", "Play Stacking Toy", "Press By Number", "Put Bottles Into Dustbin", "Stack Blocks", "Store Laptop And Headphones", "Swap Blocks", "Sweep Blocks" ], "display_task_names_zh": [ "排列数字组成最大数", "装填笔筒", "折叠衣物", "悬挂马克杯", "插入管子", "制作吐司", "整理桌面", "将物品装入盒中", "操作堆叠玩具", "依照数字按按钮", "将瓶子放入垃圾桶", "堆叠积木", "收纳笔记本电脑与耳机", "交换积木位置", "清扫积木" ], "role": "training", "size_bytes": 499637016, "size_scope": "entire subset directory, including annotations, documentation and integrity files", "subtasks": { "definitions": "training-data/simulation/one-shot/annotations/subtasks.json", "task_count": 15, "definition_count": 58, "temporal_annotations_available": true, "revision": "3f21eb1beeedaf1b1d5aaacda4a033174521d8c3", "segments": "training-data/simulation/one-shot/annotations/subtask_segments.parquet", "frame_labels": "training-data/simulation/one-shot/annotations/frame_labels.parquet", "segment_count": 58, "frame_count": 8080, "join_keys": [ "episode_index", "frame_index" ], "interval_convention": "[start_frame,end_frame)", "annotation_status": "user_confirmed_release", "boundary_mask": "boundary_uncertain" }, "task_keys": [ "arrange_largest_number", "fill_pen_holder", "fold_clothes", "hang_mugs", "insert_tubes", "make_toast", "organize_table", "pack_objects_into_box", "play_stacking_toy", "press_by_number", "put_bottles_into_dustbin", "stack_blocks", "store_laptop_and_headphones", "swap_blocks", "sweep_blocks" ], "conversion": "training-data/simulation/one-shot/meta/conversion.json" }, { "subset_id": "real-world-full-shot-data", "domain": "real-world", "shot_regime": "full-shot", "name_en": "Real-World Full-Shot Training Data", "name_zh": "真机 Full-Shot 训练集", "availability": "available", "path_in_repo": "training-data/real-world/full-shot", "format": "LeRobot v3.0", "robot_type": "G2A", "tasks": 20, "episodes": 2000, "episodes_per_task": 100, "frames": 1534321, "fps": 30, "camera_streams_per_episode": 3, "state_dimensions": 290, "action_dimensions": 52, "splits": { "train": "0:2000" }, "action_representations": [ "joint-targets", "end-effector-target-poses" ], "end_effector_target_source": "forward kinematics from joint actions", "videos": 6000, "platform": "AgiBot Genie02", "display_task_names": [ "Transfer Blocks Using Plate", "Enter Keypad Number Sequence", "Set the Table", "Arrange Seasoning Jars", "Load Plastic Plates Into Rack", "Press Button Five Times", "Place Pot On Induction Cooker", "Organize Desktop Workstation", "Organize Books And Pen", "Arrange Family Tea Set", "Operate Drawers In Sequence", "Sort Blocks in Order", "Fill And Place Pen Holder", "Organize Cosmetics Storage Box", "Strike Drum Five Times", "Apply Dish Soap Repeatedly", "Season Meat Skewers Repeatedly", "Pack Items For Shipping", "Arrange Flowers and Reed Diffuser", "Prepare Breakfast" ], "display_task_names_zh": [ "用盘子转移积木", "输入键盘数字序列", "摆放餐具", "摆放调料罐", "将塑料盘装入盘架", "按按钮五次", "将小锅放上电磁炉", "整理桌面工作区", "整理书本和笔", "摆放家庭茶具", "按顺序打开与关闭抽屉", "按顺序摆放彩色积木", "装填并摆放笔筒", "整理化妆品收纳盒", "敲鼓五次", "反复按压洗洁精瓶", "反复给肉串撒料", "打包待寄送物品", "插花与摆放藤条香薰", "准备早餐" ], "size_bytes": 23125731832, "duration_seconds": 51144.03333333333, "role": "training", "size_scope": "entire subset directory, including annotations, documentation and integrity files", "subtasks": { "definitions": "training-data/real-world/full-shot/annotations/subtasks.json", "task_count": 20, "definition_count": 97, "temporal_annotations_available": false, "revision": "3f21eb1beeedaf1b1d5aaacda4a033174521d8c3" }, "conversion": "training-data/real-world/full-shot/meta/conversion.json" }, { "subset_id": "real-world-one-shot-data", "domain": "real-world", "shot_regime": "one-shot", "availability": "available", "tasks": 20, "episodes_per_task": 1, "episodes": 20, "path_in_repo": "training-data/real-world/one-shot", "name_en": "Real-World One-Shot Training Data", "name_zh": "真机 One-Shot 训练集", "format": "LeRobot v3.0", "robot_type": "G2A", "frames": 16362, "fps": 30, "camera_streams_per_episode": 3, "state_dimensions": 290, "action_dimensions": 52, "splits": { "train": "0:20" }, "video_codecs": [ "hevc" ], "videos": 60, "platform": "AgiBot Genie02", "size_bytes": 215659523, "duration_seconds": 545.4, "display_task_names": [ "Transfer Blocks Using Plate", "Enter Keypad Number Sequence", "Set the Table", "Arrange Seasoning Jars", "Load Plastic Plates Into Rack", "Press Button Five Times", "Place Pot On Induction Cooker", "Organize Desktop Workstation", "Organize Books And Pen", "Arrange Family Tea Set", "Operate Drawers In Sequence", "Sort Blocks in Order", "Fill And Place Pen Holder", "Organize Cosmetics Storage Box", "Strike Drum Five Times", "Apply Dish Soap Repeatedly", "Season Meat Skewers Repeatedly", "Pack Items For Shipping", "Arrange Flowers and Reed Diffuser", "Prepare Breakfast" ], "display_task_names_zh": [ "用盘子转移积木", "输入键盘数字序列", "摆放餐具", "摆放调料罐", "将塑料盘装入盘架", "按按钮五次", "将小锅放上电磁炉", "整理桌面工作区", "整理书本和笔", "摆放家庭茶具", "按顺序打开与关闭抽屉", "按顺序摆放彩色积木", "装填并摆放笔筒", "整理化妆品收纳盒", "敲鼓五次", "反复按压洗洁精瓶", "反复给肉串撒料", "打包待寄送物品", "插花与摆放藤条香薰", "准备早餐" ], "role": "training", "size_scope": "entire subset directory, including annotations, documentation and integrity files", "subtasks": { "definitions": "training-data/real-world/one-shot/annotations/subtasks.json", "task_count": 20, "definition_count": 97, "temporal_annotations_available": false, "revision": "3f21eb1beeedaf1b1d5aaacda4a033174521d8c3" }, "conversion": "training-data/real-world/one-shot/meta/conversion.json" }, { "subset_id": "real-world-test-30", "domain": "real-world", "role": "evaluation", "semantic_split": "test", "availability": "available", "name_en": "Small-Scale Real-World Public Diagnostic Test Set", "name_zh": "小规模真机公开诊断测试集", "path_in_repo": "evaluation/real-world/test-30", "format": "LeRobot v3.0", "tasks": 4, "episodes": 30, "frames": 27498, "fps": 30, "camera_streams_per_episode": 3, "videos": 90, "size_bytes": 382100550, "duration_seconds": 916.6, "splits": { "train": "0:30" }, "native_split_note": "train is the LeRobot container split; the collection is intended for evaluation.", "evaluation_scope": "small-scale real-world In-Domain diagnostic subset with partial solution coverage", "model_inputs": [ "images", "instruction" ], "annotations": "evaluation/real-world/test-30/annotations", "documentation": { "en": "evaluation/real-world/test-30/README.md", "zh": "evaluation/real-world/test-30/README_zh.md" }, "hf_config": "real-world-test-30", "hf_split": "test", "episode_catalog": "evaluation/real-world/test-30/episode_catalog.parquet", "checksums": "evaluation/real-world/test-30/checksums.sha256", "metric_keys": [ "tga_easy", "tga_hard", "voc", "cycle_voc", "vs", "sa", "sia", "csvc", "trr", "fpl", "voc_mem" ], "display_task_names": [ "Organize Desktop Workstation", "Arrange Flowers and Reed Diffuser", "Strike Drum Five Times", "Press Button Five Times" ], "display_task_names_zh": [ "整理书桌工作区", "插花与摆放藤条香薰", "敲鼓五次", "按按钮五次" ], "index_scope": "task_index and episode_index are local to this subset and independent of training subset indices.", "task_episode_counts": [ { "task_index": 0, "episodes": 12 }, { "task_index": 1, "episodes": 12 }, { "task_index": 2, "episodes": 3 }, { "task_index": 3, "episodes": 3 } ], "primary_camera": "observation.images.top_head", "primary_video_column": "video_primary", "data_revision": "578593ef28ee7a2d861ebf38162c928291b0d4a9", "size_scope": "entire subset directory, including annotations, documentation and integrity files", "subtasks": { "definitions": "evaluation/real-world/test-30/annotations/subtasks.json", "task_count": 4, "definition_count": 21, "temporal_annotations_available": true, "revision": "3f21eb1beeedaf1b1d5aaacda4a033174521d8c3", "segments": "evaluation/real-world/test-30/annotations/subtask_segments.parquet", "segment_count": 153, "join_keys": [ "episode_index", "task_index" ], "interval_convention": "[start_frame,end_frame_exclusive)" } }, { "subset_id": "simulation-test-30", "domain": "simulation", "role": "evaluation", "semantic_split": "test", "availability": "available", "name_en": "Small-Scale Simulation Public Diagnostic Test Set", "name_zh": "小规模仿真公开诊断测试集", "path_in_repo": "evaluation/simulation/test-30", "format": "LeRobot v3.0", "tasks": 4, "episodes": 30, "frames": 17149, "fps": 25, "camera_streams_per_episode": 3, "videos": 90, "duration_seconds": 685.96, "splits": { "train": "0:30" }, "native_split_note": "train is the LeRobot container split; the collection is intended for evaluation.", "evaluation_scope": "small-scale simulation In-Domain diagnostic subset with partial solution coverage", "model_inputs": [ "images", "instruction" ], "annotations": "evaluation/simulation/test-30/annotations", "documentation": { "en": "evaluation/simulation/test-30/README.md", "zh": "evaluation/simulation/test-30/README_zh.md" }, "hf_config": "simulation-test-30", "hf_split": "test", "episode_catalog": "evaluation/simulation/test-30/episode_catalog.parquet", "checksums": "evaluation/simulation/test-30/checksums.sha256", "metric_keys": [ "tga_easy", "tga_hard", "voc", "cycle_voc", "vs", "sa", "sia", "csvc", "trr", "voc_mem" ], "display_task_names": [ "Fill Pen Holder", "Hang Mugs", "Press By Number", "Swap Blocks" ], "display_task_names_zh": [ "装填笔筒", "悬挂马克杯", "依照数字按按钮", "交换积木" ], "index_scope": "task_index and episode_index are local to this subset and independent of training and real-world test subset indices.", "task_episode_counts": [ { "task_index": 0, "episodes": 12 }, { "task_index": 1, "episodes": 12 }, { "task_index": 2, "episodes": 3 }, { "task_index": 3, "episodes": 3 } ], "primary_camera": "observation.images.front", "primary_video_column": "video_primary", "annotation_evidence_note": "ED/MS terminal labels and semantic alignment are user-attested; TRR/MEM terminal-label evidence remains pending independent verification.", "size_bytes": 594599072, "data_revision": "578593ef28ee7a2d861ebf38162c928291b0d4a9", "size_scope": "entire subset directory, including annotations, documentation and integrity files", "subtasks": { "definitions": "evaluation/simulation/test-30/annotations/subtasks.json", "task_count": 4, "definition_count": 19, "temporal_annotations_available": true, "revision": "3f21eb1beeedaf1b1d5aaacda4a033174521d8c3", "segments": "evaluation/simulation/test-30/annotations/subtask_segments.parquet", "segment_count": 54, "join_keys": [ "episode_index", "task_index" ], "interval_convention": "[start_frame,end_frame_exclusive)" } } ], "planned_subsets": [], "data_revision": "main", "publication_authorization": "All data were authorized before public release.", "pretty_name": "RoboValue Training and Small-Scale Evaluation Data", "training_data_revision": "main", "revision_note": "Training metadata and documentation use main; pin the current repository commit for reproducibility. Evaluation payloads retain evaluation_data_revision.", "evaluation_data_revision": "578593ef28ee7a2d861ebf38162c928291b0d4a9", "training_path_revision": "main", "metric_documentation": { "english": "README.md#metric-summary", "chinese": "README_zh.md#metric-summary", "score_direction": { "default": "higher_is_better", "FPL": "lower_is_better" }, "reference_section": "4.3", "reference_pages": [ 9, 10, 11, 12 ], "metric_aliases": { "Memory-VOC": { "dataset_label": "VOC-MEM", "membership_key": "voc_mem" } } }, "subtask_documentation": { "en": "SUBTASKS.md", "zh": "SUBTASKS_zh.md", "training_tasks": 35, "distinct_definitions": 155, "revision": "3f21eb1beeedaf1b1d5aaacda4a033174521d8c3" } }