|
Download README.md from drwahl/so101_block_stack: direct link, hf CLI and curl.
- Browser
- Download file 1.88 kB
-
https://huggingface.co/datasets/drwahl/so101_block_stack/resolve/main/README.md
- Command line
-
hf download hf://datasets/drwahl/so101_block_stack/README.md
-
curl -L -o README.md https://huggingface.co/datasets/drwahl/so101_block_stack/resolve/main/README.md
1.88 kB
metadata
license: apache-2.0
task_categories:
- robotics
tags:
- LeRobot
- so101
- pi0.5
so101_block_stack
The block stack in simulation and on the arm, in one dataset. lerobot trains on a single dataset, so learning from both means writing both into one. Built by sim/scripts/combine.py in vla-test from drwahl/so101_block_stack_sim and drwahl/so101_block_stack_hw.
468 episodes in LeRobot v3.0 format at 10 Hz: the 360 sim episodes, and the 36 hardware episodes written three times. How many times the hardware set is written is what sets how much of a batch comes off the arm.
meta/train_layouts.jsonl and meta/eval_layouts.jsonl come from the sim set. meta/stats.json is recomputed over the whole.
drwahl/pi05_so101_block_stack is trained on this.
Citation
@misc{wahl2026so101blockstack,
title = {{SO-101 block stacking demonstrations in simulation and on hardware}},
author = {Wahl, Daniel},
year = {2026},
publisher = {Hugging Face},
doi = {10.57967/hf/10750},
url = {https://huggingface.co/datasets/drwahl/so101_block_stack}
}
@misc{wahl2026so101blockstacksim,
title = {{SO-101 block stacking demonstrations in simulation}},
author = {Wahl, Daniel},
year = {2026},
publisher = {Hugging Face},
doi = {10.57967/hf/10748},
url = {https://huggingface.co/datasets/drwahl/so101_block_stack_sim}
}
@misc{wahl2026so101blockstackhw,
title = {{SO-101 block stacking demonstrations on hardware}},
author = {Wahl, Daniel},
year = {2026},
publisher = {Hugging Face},
doi = {10.57967/hf/10749},
url = {https://huggingface.co/datasets/drwahl/so101_block_stack_hw}
}