YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
LIBERO LeRobot post-training data
Repository ID: Mark-ZJTang/libero_real.
LeRobot v2-format downstream fine-tuning dataset with 1,693 episodes, 273,465 frames, and 40 task IDs. It contains agent-view and wrist-view images, state, and 7D actions. LIBERO simulator evaluation uses bundled BDDL/init/assets and does not read these demonstrations.
- Split:
train(0:1693), 10 FPS, two chunks. imageandwrist_image: image,[256, 256, 3]each.state: float32,[8].actions: float32,[7].- Task, episode, frame, and global indices are retained.
Restore it to the ALAM repository's loader-compatible relative path and verify both downstream datasets with:
.venvs/publish/bin/python workflows/publishing/download_huggingface.py \
--artifact libero_real_lerobot
.venvs/pi0/bin/python workflows/pi0_post_training/tests/test_datasets.py
The resulting root is data/lerobot/libero_real; the public fine-tuning wrapper
supplies its parent as HF_LEROBOT_HOME.
Before publication, verify LIBERO dataset redistribution rights and upstream attribution, then add the correct license and data-generation details.
- Downloads last month
- 461