跳转至

📇 机器人数据集登记表

这是索引层,不是推荐清单。 目的是让检索能回答「有没有 / 大概是什么 / 长什么样」, 而不是替代卡片里的判断。登记 ≠ 背书。

两个层级: - 🟢 已下卡 — 我们读过全文、有判断和批注,结论可引用。 - 🟡 已登记 — 只有机器抄来的客观元数据,没人读过,引用前必须回一手源核。

194 条(🟢 13 · 🟡 181)。 数据源:我们的卡片 + datasets.bot + HuggingFace API。生成于 2026-07-27。

🖐 带触觉模态的

主线相关,单独提出来。模态标注来自 datasets.bot,未逐条核实

  • 卡片-RH20T (110,000段 · custom)
  • Humanoid Everyday (27.8h · 10,300段 · ↓94,220 · lerobot) USC-PSI-Lab/humanoid-everyday
  • AgiBot World 2026 (2976.4h · ↓27,015 · lerobot) agibot-world/AgiBotWorld2026
  • CALVIN (24h · custom)
  • FTP-1 Dataset (3000h · zarr)
  • LESS (Local Encoder for Spatial Sensing) (800h · custom)

🟢 已下卡(有我们的判断)

🟡 已登记(未读,仅客观元数据)

  • ABC-130k (3598.2h · 130,919段 · ↓665,474 · custom) XDOF/ABC-130k
  • 2.1tbofdata (↓152,053) Whoisjutanlee/2.1tbofdata
  • Hy-Embodied-0.5-VLA-Data (↓132,072) tencent/Hy-Embodied-0.5-VLA-Data
  • PhysicalAI-Robotics-GR00T-X-Embodiment-Sim (↓125,372) Rooftech650/PhysicalAI-Robotics-GR00T-X-Embodiment-Sim
  • kuka_lerobot (↓119,289) introvoyz041/kuka_lerobot
  • xperience-10m (↓109,150) ropedia-ai/xperience-10m
  • fractal20220817_data_lerobot (↓106,932) IPEC-COMMUNITY/fractal20220817_data_lerobot
  • Humanoid Everyday (27.8h · 10,300段 · ↓94,220 · lerobot) USC-PSI-Lab/humanoid-everyday
  • kumagong (↓92,840) balatubs123/kumagong
  • bridge_orig_lerobot (↓84,775) arekborucki/bridge_orig_lerobot
  • bc_z_lerobot (↓84,680) IPEC-COMMUNITY/bc_z_lerobot
  • L2D (↓76,016) yaak-ai/L2D
  • agibot_alpha_v30 (↓74,078) cadene/agibot_alpha_v30
  • InternData-N1 (↓73,851) InternRobotics/InternData-N1
  • OpenFly-traj-converted (↓72,225) newniuuu/OpenFly-traj-converted
  • stera-10m (↓62,302) fpvlabs/stera-10m
  • trex_dataset (↓61,716) zekaiwang/trex_dataset
  • OmniWorld (↓61,434) InternRobotics/OmniWorld
  • MolmoAct2 Bimanual YAM Dataset (720h · 32,246段 · ↓57,816 · lerobot) allenai/MolmoAct2-BimanualYAM-Dataset
  • DOM (↓51,190) hzxie/DOM
  • PhysicalAI-Autonomous-Vehicle-Cosmos-Drive-Dreams (↓49,622) nvidia/PhysicalAI-Autonomous-Vehicle-Cosmos-Drive-Dreams
  • computer-use-large (↓49,388) REXX-NEW/computer-use-large
  • PhysicalAI-Robotics-Open-H-Embodiment (↓46,070) nvidia/PhysicalAI-Robotics-Open-H-Embodiment
  • InternData-A1 (↓43,338) InternRobotics/InternData-A1
  • WIYH (↓33,990) tars-robotics/WIYH
  • X-WAM-RoboTwin (↓33,891) sharinka0715/X-WAM-RoboTwin
  • xrobotwin_data (↓28,904) ssr-xembodiment/xrobotwin_data
  • OmniAction (↓28,423) OpenMOSS-Team/OmniAction
  • AgiBot World 2026 (2976.4h · ↓27,015 · lerobot) agibot-world/AgiBotWorld2026
  • droid_1.0.1 (↓25,007) lerobot/droid_1.0.1
  • HIW-500-LeRobot (↓23,614) BitRobot/HIW-500-LeRobot
  • deform360 (↓23,460) brownu/deform360
  • Retargeted_AMASS_for_bxi_elf2 (↓22,885) fleaven/Retargeted_AMASS_for_bxi_elf2
  • bridgev2 (↓20,670) Qu3tzal/bridgev2
  • FreeTacMan (↓20,143) OpenDriveLab/FreeTacMan
  • rbo_oxe_base_language_table_lerobot (↓19,070) saaduddinM/rbo_oxe_base_language_table_lerobot
  • gs_scenes (↓17,812) RukawaY/gs_scenes
  • Wanda (↓17,634) LeCAR-Lab/Wanda
  • XVLA-Soft-Fold (↓17,567) Facebear/XVLA-Soft-Fold
  • Retargeted_AMASS_for_FourierN1 (↓17,115) fleaven/Retargeted_AMASS_for_FourierN1
  • PhysicalAI-Robotics-Manipulation-SingleArm (↓16,581) nvidia/PhysicalAI-Robotics-Manipulation-SingleArm
  • real_robot_data (↓15,503) Hoshipu/real_robot_data
  • Kai0 (↓14,999) OpenDriveLab-org/Kai0
  • OctoSense (↓14,652) anthonytec2/OctoSense
  • pusht (↓14,495) lerobot/pusht
  • roboverse_data (↓14,438) RoboVerseOrg/roboverse_data
  • SynthVerse (↓13,631) InternRobotics/SynthVerse
  • pack_toothbrush_Nov19-advantages (↓13,626) villekuosmanen/pack_toothbrush_Nov19-advantages
  • OSCAR_human (↓13,081) zywu2115/OSCAR_human
  • OSCAR_robot (↓12,753) zywu2115/OSCAR_robot
  • nuReasoning (↓12,075) qixuewei/nuReasoning
  • MolmoAct Dataset (22h · 10,488段 · ↓11,800 · lerobot) allenai/MolmoAct-Dataset
  • RealSource-World (↓11,772) RealSourceData/RealSource-World
  • grand_tour_dataset (↓11,211) leggedrobotics/grand_tour_dataset
  • comet-1.5k (↓11,100) delinqu/comet-1.5k
  • GoPro-Raw-Videos (↓10,493) Fanqi-Lin/GoPro-Raw-Videos
  • 2025-challenge-demos (↓10,390) elonelonelon/2025-challenge-demos
  • build_block_tower-advantages (↓10,152) villekuosmanen/build_block_tower-advantages
  • bridge_v2_lerobot (↓10,146) jesbu1/bridge_v2_lerobot
  • datagen-clutter-v1-joint-5cam (↓10,055) IDEAS-Lab-Northwestern/datagen-clutter-v1-joint-5cam
  • Retargeted_AMASS_for_robotics (↓9,581) fleaven/Retargeted_AMASS_for_robotics
  • PhysicalAI-Robotics-Manipulation-Kitchen-Demos (↓9,182) nvidia/PhysicalAI-Robotics-Manipulation-Kitchen-Demos
  • PH2D (↓8,980) RogerQi/PH2D
  • DILLO-LIBERO-dataset (↓8,977) Sapienza/DILLO-LIBERO-dataset
  • UrbanVerse-Training-Scenes (↓8,798) UCLA-VAIL/UrbanVerse-Training-Scenes
  • Cobot_Magic_make_hamburger (↓8,699) RoboCOIN/Cobot_Magic_make_hamburger
  • HumanEgo (↓8,573) Leo-TX/HumanEgo
  • reachy-mini-emotions-library (↓8,495) pollen-robotics/reachy-mini-emotions-library
  • high_quality_folding (↓8,283) lerobot/high_quality_folding
  • InternData-A1-LeRobot-v3.0-by-embodiment (↓8,212) griffinlabs/InternData-A1-LeRobot-v3.0-by-embodiment
  • ActionNet (↓8,150) FourierIntelligence/ActionNet
  • trex-visualizer (↓8,132) Beakerman0101/trex-visualizer
  • simlingo (↓7,961) RenzKa/simlingo
  • lead (↓7,919) ln2697/lead
  • HABIT (Human Aware Behavior and Interaction Training dataset) (164h · 10,563段 · ↓7,648 · custom/lerobot) configinc/HABIT
  • FinErva (↓7,499) jerrychi/FinErva
  • agent-reward-bench (↓7,403) McGill-NLP/agent-reward-bench
  • PhysicalAI-Robotics-Manipulation-Kitchen (↓7,335) nvidia/PhysicalAI-Robotics-Manipulation-Kitchen
  • metaworld_mt50 (↓7,282) lerobot/metaworld_mt50
  • RoboBench (↓7,281) LeoFan01/RoboBench
  • MetaFold (↓7,191) chenhn02/MetaFold
  • AMASS_Retargeted_for_G1 (↓6,861) ember-lab-berkeley/AMASS_Retargeted_for_G1
  • D2E-480p (↓6,759) open-world-agents/D2E-480p
  • pusht_image (↓6,746) lerobot/pusht_image
  • roboreal_lerobot (↓6,547) JackLiu0406/roboreal_lerobot
  • AI2_Alphabot_2_organize_cosmetics_1 (↓6,411) RoboCOIN/AI2_Alphabot_2_organize_cosmetics_1
  • fmb_dataset_lerobot (↓6,397) IPEC-COMMUNITY/fmb_dataset_lerobot
  • 10Kh-RealOmin-OpenData (↓6,246) ad1t7a/10Kh-RealOmin-OpenData
  • libero (↓6,107) lerobot/libero
  • bridge_v2_lerobot_pathmask (↓6,091) jesbu1/bridge_v2_lerobot_pathmask
  • Uni-GUI-Desktop-1 (↓6,071) UI-MOPD/Uni-GUI-Desktop-1
  • xarm_lift_medium (↓5,948) lerobot/xarm_lift_medium
  • X-WAM-RoboCasa (↓5,899) sharinka0715/X-WAM-RoboCasa
  • ArtVIP (↓5,847) X-Humanoid/ArtVIP
  • libero_90_no_noops_lerobot (↓5,776) IPEC-COMMUNITY/libero_90_no_noops_lerobot
  • MOSAIC_Dataset (↓5,682) BAAI-Humanoid/MOSAIC_Dataset
  • PhysicalAI-SimReady-Warehouse-01 (↓5,623) nvidia/PhysicalAI-SimReady-Warehouse-01
  • libero_plus_rlds (↓5,605) Sylvest/libero_plus_rlds
  • UrbanVerse-100K (↓5,510) Oatmealliu/UrbanVerse-100K
  • behavior-1k-mp-collected-turning-on-radio (↓5,427) Hoshipu/behavior-1k-mp-collected-turning-on-radio
  • AI2_Alphabot_2_store_toys_5 (↓5,417) RoboCOIN/AI2_Alphabot_2_store_toys_5
  • data (↓5,309) dual-sim/data
  • MindCraft (↓5,241) TMarcus/MindCraft
  • 2026-humanoid-ikea-assembly-challenge (↓5,196) BitRobot/2026-humanoid-ikea-assembly-challenge
  • Cobot_Magic_pull_zipper (↓5,188) RoboCOIN/Cobot_Magic_pull_zipper
  • GWAM_Data (↓5,166) ChangChrisLiu/GWAM_Data
  • Cobot_Magic_pour_drink (↓5,115) RoboCOIN/Cobot_Magic_pour_drink
  • Uni-GUI-OpenMobile (↓5,073) UI-MOPD/Uni-GUI-OpenMobile
  • droid_3d (↓5,045) ZibinDong/droid_3d
  • GR1_robot (↓5,043) Physis-AI/GR1_robot
  • community_dataset_v3 (↓5,004) HuggingFaceVLA/community_dataset_v3
  • something_something_v2_lerobot (↓4,984) JUNTAO123/something_something_v2_lerobot
  • AIRoA MoMa Dataset (94h · 25,469段 · lerobot) airoa-org/airoa-moma
  • ALFA (AirLab Failure and Anomaly Dataset) (1.3h · 47段 · custom/rosbag)
  • ALOHA Static (2.5h · 800段 · lerobot)
  • CALVIN (24h · custom)
  • DexMimicGen (21,000段 · hdf5) MimicGen/dexmimicgen_datasets
  • DexYCB (0.83h · 1,000段 · custom)
  • DreamDojo GR-1 Post-Training (88.4h · lerobot) nvidia/PhysicalAI-Robotics-GR00T-Teleop-GR1
  • EPIC-KITCHENS-100 (100h · 700段 · custom)
  • Ego-Exo4D (1286.3h · 5,035段 · custom)
  • Ego4D (3670h · custom)
  • EgoDex (829h · 338,000段 · hdf5)
  • EgoInfinity (1,150段 · custom) Rice-RobotPI-Lab/egoinfinity
  • FMB (Functional Manipulation Benchmark) (31.3h · 22,550段 · custom/rlds)
  • FTP-1 Dataset (3000h · zarr)
  • FurnitureBench (219.6h · 5,100段 · custom)
  • Galaxea Open-World Dataset (500h · 100,000段 · lerobot) OpenGalaxea/Galaxea-Open-World-Dataset
  • HIW-500: Humanoids In-the-Wild (500h · 23,000段 · rosbag/lerobot) BitRobot/HIW-500
  • HumanPlus (40h · hdf5)
  • In-flight Positional and Energy-Use Dataset of Package-Delivery Quadcopter UAVs (10.75h · 209段 · custom)
  • LESS (Local Encoder for Spatial Sensing) (800h · custom)
  • Language-Table (3865h · 442,226段 · rlds)
  • LeRobot Shirt-Folding Dataset (131h · 5,688段 · lerobot) lerobot/full_folding
  • ManiSkill2 (hdf5) haosulab/ManiSkill2
  • ManipArena (188h · 10,812段 · lerobot) ManipArena/maniparena-dataset
  • MimicGen (48,000段 · hdf5) amandlek/mimicgen_datasets
  • Mobile ALOHA (276段 · hdf5/rlds)
  • MolmoAct2 SO-100/SO-101 Dataset (lerobot) allenai/MolmoAct2-SO100_101-Dataset
  • MolmoBot (5704h · 1,700,000段 · custom) allenai/MolmoBot-Data
  • NVIDIA GR00T X-Embodiment Sim (lerobot) nvidia/PhysicalAI-Robotics-GR00T-X-Embodiment-Sim
  • RT-1 Robot Action Dataset (350.6h · 87,212段 · rlds)
  • RoboArena (10,783段 · custom)
  • RoboCasa (2227h · hdf5/lerobot)
  • RoboDojo (38.57h · 5,300段 · hdf5/lerobot) RoboDojo-Benchmark/RoboDojo
  • RoboNet (162,417段 · hdf5/rlds)
  • RoboSet (RoboAgent) (28,500段 · hdf5) jdvakil/RoboSet-Teleoperation
  • Robotic Interestingness Dataset (SubTF) (7.8h · 7段 · custom/rosbag)
  • SubT-MRS (2000h · 25段 · custom/rosbag)
  • TartanAir (27.8h · 1,037段 · custom)
  • TartanAir V2 (custom)
  • TartanAviation (3374h · 550段 · custom)
  • TartanDrive (5h · custom/rosbag)
  • TartanDrive 2.0 (7h · custom/rosbag)
  • TrajAir (2442h · custom)
  • WIT-UAS (Wildland-fire Infrared Thermal UAS Dataset) (custom/rosbag)
  • Wire Detection Dataset (custom)
  • Yamaha-CMU Off-Road Dataset (YCOR) (custom)
  • robomimic (hdf5/rlds) amandlek/robomimic

衍生条目(折叠)

另有 22 条是已有条目的衍生 repo。其中「格式移植」那类往往正是实际要下的那个 (HF 上常见把原数据集转成 LeRobot v3.0 再发一个 repo);「单任务分片」则多为噪声。 全部仍在 JSON 里可被检索,只是不占人读版面。

  • language_table_lerobot IPEC-COMMUNITY/language_table_lerobot —— 格式移植,源自 Language-Table
  • droid_lerobot IPEC-COMMUNITY/droid_lerobot —— 格式移植,源自 DROID
  • AgiBotWorld-Beta agibot-world/AgiBotWorld-Beta —— 子集/变体,源自 AgiBot-World
  • MolmoAct-Midtraining-Mixture allenai/MolmoAct-Midtraining-Mixture —— 子集/变体,源自 MolmoAct Dataset
  • EgoDex-LeRobot-v3.0 griffinlabs/EgoDex-LeRobot-v3.0 —— 格式移植,源自 EgoDex
  • abc_130k_v3_train lerobot/abc_130k_v3_train —— 子集/变体,源自 ABC-130k
  • AgiBot-World-Beta-no-torso-movement griffinlabs/AgiBot-World-Beta-no-torso-movement —— 子集/变体,源自 AgiBot-World
  • droid_720p XinxuanLuM/droid_720p —— 子集/变体,源自 DROID
  • AgiBotWorld-Alpha agibot-world/AgiBotWorld-Alpha —— 子集/变体,源自 AgiBot-World
  • Galaxea-Open-World-Dataset-LeRobot-v3.0 griffinlabs/Galaxea-Open-World-Dataset-LeRobot-v3.0 —— 格式移植,源自 Galaxea Open-World Dataset
  • aloha_sim_transfer_cube_human lerobot/aloha_sim_transfer_cube_human —— 子集/变体,源自 ALOHA
  • droid_filter_lerobot periphanes/droid_filter_lerobot —— 格式移植,源自 DROID
  • aloha_sim_insertion_human lerobot/aloha_sim_insertion_human —— 子集/变体,源自 ALOHA
  • aloha_pen_uncap Voxel51/aloha_pen_uncap —— 子集/变体,源自 ALOHA
  • aloha_sim_insertion_scripted lerobot/aloha_sim_insertion_scripted —— 子集/变体,源自 ALOHA
  • calvin-task-ABCD-D-lerobot fywang/calvin-task-ABCD-D-lerobot —— 格式移植,源自 CALVIN
  • AgiBot-g1_left_capture_part RoboCOIN/AgiBot-g1_left_capture_part —— 单任务分片
  • aloha_sim_transfer_cube_scripted lerobot/aloha_sim_transfer_cube_scripted —— 子集/变体,源自 ALOHA
  • AgiBot-g1_box_storage_cardboard_box_c RoboCOIN/AgiBot-g1_box_storage_cardboard_box_c —— 单任务分片
  • AgiBot-g1_storage_item_e RoboCOIN/AgiBot-g1_storage_item_e —— 单任务分片
  • AgiBot-g1_box_storage_cardboard_box_a RoboCOIN/AgiBot-g1_box_storage_cardboard_box_a —— 单任务分片
  • AgiBot-g1_picks_up_battery_a RoboCOIN/AgiBot-g1_picks_up_battery_a —— 单任务分片

怎么用:机器可读全量在同目录 _数据集登记表.json(不随站点发布)。 刷新:python scripts/gen_datasets_registry.py --refresh