Here will be stored the "Benchmarking Vision-Language-Action Models on SO-101: Failure and Recovery Analysis" dataset.
🏗️ Building on HF
qxc
Qiu-Xinchuan
AI & ML interests
Embodied AI
Recent Activity
updated a collection 15 days ago
Dual-arm robot dataset (SO-101) updated a collection 15 days ago
Single-arm robot dataset (SO-101) updated a collection 15 days ago
Single-arm robot dataset (SO-101) Organizations
None yet
This dataset is the SO-101 single-arm dataset, containing a total of 966 demonstrations across 11 tasks, with a total duration of 7 hours, 32 minutes.
使用evo-rl项目对total数据进行了5轮强化学习训练
500 dual-arm demonstrations for color-based block placement, expanded from 200 demos through simple-to-complex data collection.
-
Qiu-Xinchuan/so101_place_block_next_to_same_color_4cam_rl_0
Viewer • Updated • 184k • 920 -
Qiu-Xinchuan/so101_place_block_next_to_same_color_4cam_rl_1
Viewer • Updated • 116k • 1.17k -
Qiu-Xinchuan/so101_double_4cam_rl_total_0
Viewer • Updated • 299k • 187 -
Qiu-Xinchuan/so101_place_block_next_to_same_color_4cam_rl_0_3epoch
4B • Updated • 5
The SO-101 dual-arm dataset includes 1,783 demonstrations (8 tasks, ~18h total).
pi05_total_40000：单臂，可以进行夹取方块、把笔放入笔筒等操作，基于aawangge/so101_total数据集训练
aaawangge/so101_towel_rl_2_pi05_acp_120000：双臂，可以完成叠毛巾任务，基于aaawangge/so101_towe
Here will be stored the "Benchmarking Vision-Language-Action Models on SO-101: Failure and Recovery Analysis" dataset.
500 dual-arm demonstrations for color-based block placement, expanded from 200 demos through simple-to-complex data collection.
-
Qiu-Xinchuan/so101_place_block_next_to_same_color_4cam_rl_0
Viewer • Updated • 184k • 920 -
Qiu-Xinchuan/so101_place_block_next_to_same_color_4cam_rl_1
Viewer • Updated • 116k • 1.17k -
Qiu-Xinchuan/so101_double_4cam_rl_total_0
Viewer • Updated • 299k • 187 -
Qiu-Xinchuan/so101_place_block_next_to_same_color_4cam_rl_0_3epoch
4B • Updated • 5
This dataset is the SO-101 single-arm dataset, containing a total of 966 demonstrations across 11 tasks, with a total duration of 7 hours, 32 minutes.
The SO-101 dual-arm dataset includes 1,783 demonstrations (8 tasks, ~18h total).
使用evo-rl项目对total数据进行了5轮强化学习训练
pi05_total_40000：单臂，可以进行夹取方块、把笔放入笔筒等操作，基于aawangge/so101_total数据集训练
aaawangge/so101_towel_rl_2_pi05_acp_120000：双臂，可以完成叠毛巾任务，基于aaawangge/so101_towe