|
Download README.md from padiredd/Ablation_Block_visual: direct link, hf CLI and curl.
- Browser
- Download file 1.56 kB
-
https://huggingface.co/datasets/padiredd/Ablation_Block_visual/resolve/main/README.md
- Command line
-
hf download hf://datasets/padiredd/Ablation_Block_visual/README.md
-
curl -L -o README.md https://huggingface.co/datasets/padiredd/Ablation_Block_visual/resolve/main/README.md
1.56 kB
| license: mit | |
| tags: | |
| - LeRobot | |
| - robotics | |
| - simulation | |
| - cyclo_lab | |
| task_categories: | |
| - robotics | |
| # Ablation_Block_visual | |
| Ablation_Block_visual: 2,000 successful sim block-pick demos for the visual vs physical domain-randomization ablation (data_collection/ablation_vp, 2026-11-07). Code-as-policy program cap/programs/block_pick_v1_nocomp.py (the 2k_withoutpush recipe: no pushes, |yaw| 0-45 deg, full left-arm region, seeds 40001-40040, gripper open/close randomization). **Visual randomization only**: table look (nominal / flat colour / library texture), backdrop (real lab backgrounds / library textures / grid), block colour jitter, dr2 room lights, 0-3 spot lights. Start pose, table height (29.25 in) and cameras are nominal. Block pose, yaw, motion parameters and gripper values match padiredd/Ablation_Block_physical episode for episode (same pair_key = seed, batch, env; --split_rng). | |
| FFW-SG2 (ROBOTIS) in Isaac Lab, 15 fps, 3 cameras (head, left wrist, right wrist) stretched to 240x240, 16-dim actions / 22-dim state in the real datasets' layout. Collected in the cyclo_lab real2sim pipeline (padiredd). | |
| | | | | |
| |---|---| | |
| | episodes | 2000 | | |
| | frames | 337435 | | |
| | fps | 15 | | |
| | LeRobot format | v3.0 | | |
| | cameras | observation.images.rgb.cam_left_wrist, observation.images.rgb.cam_right_wrist, observation.images.rgb.cam_left_head | | |
| | other features | observation.state, action, timestamp, frame_index, episode_index, index, task_index | | |
| Uploaded 2026-11-07 04:06 UTC from the lab server with data_collection/hf_datasets/upload_datasets.py. | |