← All skills
fine-tuning-serving-openpi
80 / 100 · Low risk
Fine to try. Some of what it can do is expected: a coding skill usually runs build and test commands.
By Orchestra-Research/AI-Research-SKILLs (GitHub) · coding · 14,447 stars · scanned at commit 773a529
Fine-tune and serve Physical Intelligence OpenPI models (pi0, pi0-fast, pi0.5) using JAX or PyTorch backends for robot policy inference across ALOHA, DROID, and LIBERO environments. Use when adapting pi0 models to custom datasets, converting JAX checkpoints to PyTorch, running policy inference servers, or debugging norm stats and GPU memory issues.
Works with any agent that supports SKILL.md skills.
What this skill can do
Runs commands on your computer
The author does not mention this. Expected for a coding skill.
Found in 26 places.
Show 17 more
- 18-multimodal/openpi/SKILL.md, line 92
GIT_LFS_SKIP_SMUDGE=1 uv sync - 18-multimodal/openpi/SKILL.md, line 98
GIT_LFS_SKIP_SMUDGE=1 uv pip install -e . - 18-multimodal/openpi/SKILL.md, line 104
uv run python -c "from openpi.training import config as _config; print(_config.get_config('pi05_droid').name)" - 18-multimodal/openpi/SKILL.md, line 171
uv run scripts/compute_norm_stats.py --config-name <config_name> - 18-multimodal/openpi/SKILL.md, line 179
XLA_PYTHON_CLIENT_MEM_FRACTION=0.9 uv run scripts/train.py <config_name> \ - 18-multimodal/openpi/SKILL.md, line 187
uv run --group rlds scripts/compute_norm_stats.py \ - 18-multimodal/openpi/SKILL.md, line 199
uv run scripts/serve_policy.py policy:checkpoint \ - 18-multimodal/openpi/SKILL.md, line 233
uv sync - 18-multimodal/openpi/SKILL.md, line 242
cp -r ./src/openpi/models_pytorch/transformers_replace/* \ - 18-multimodal/openpi/SKILL.md, line 249
uv run examples/convert_jax_model_to_pytorch.py \ - 18-multimodal/openpi/SKILL.md, line 260
uv run scripts/train_pytorch.py <config_name> --exp_name <run_name> - 18-multimodal/openpi/SKILL.md, line 266
uv run torchrun --standalone --nnodes=1 --nproc_per_node=<num_gpus> \ - 18-multimodal/openpi/SKILL.md, line 314
uv run scripts/serve_policy.py --env ALOHA - 18-multimodal/openpi/SKILL.md, line 320
uv run scripts/serve_policy.py policy:checkpoint \ - 18-multimodal/openpi/SKILL.md, line 330
uv run examples/simple_client/main.py --env DROID - 18-multimodal/openpi/SKILL.md, line 338
pip install "openpi-client @ git+https://github.com/Physical-Intelligence/openpi.git#subdirectory=packages/openpi-client" - 18-multimodal/openpi/SKILL.md, line 404
GIT_LFS_SKIP_SMUDGE=1 uv sync
and 6 more places
Downloads and installs other software
The author does not mention this.
Found in 10 places.
Show 7 more
Install
Paste this into your terminal. It downloads exactly the version we scanned, not whatever is newest.
if [ -e ~/.agents/skills/'fine-tuning-serving-openpi' ] || [ -L ~/.agents/skills/'fine-tuning-serving-openpi' ]; then echo 'A skill folder named fine-tuning-serving-openpi already exists — not replacing it' >&2; false; else d=$(mktemp -d) && git clone --quiet --filter=blob:none --no-checkout 'https://github.com/Orchestra-Research/AI-Research-SKILLs' "$d" && git -C "$d" checkout --quiet '773a52944ba4747a18bd4ae9ade53fff041adcbc' -- '18-multimodal/openpi' && mkdir -p ~/.agents/skills && cp -R "$d"/'18-multimodal/openpi' ~/.agents/skills/'fine-tuning-serving-openpi' && rm -rf "$d"; fi
Codex reads ~/.agents/skills directly. For Claude Code, Cursor, Copilot, Gemini CLI, Antigravity or OpenCode, also link it. For Claude Code: ln -s ~/.agents/skills/'fine-tuning-serving-openpi' ~/.claude/skills/'fine-tuning-serving-openpi' (other agents: ~/.cursor/skills, ~/.copilot/skills, ~/.gemini/skills, ~/.gemini/config/skills, ~/.config/opencode/skills). Or install through the finder skill, which does both.
Badge for your README
The badge for Orchestra-Research/AI-Research-SKILLs shows the score of its lowest-scoring skill.

[](https://zalavijaysinh.github.io/skilltrust/repos/Orchestra-Research/AI-Research-SKILLs/)
Score history
| Date | Version | Score |
|---|
| 2026-11-08 | 773a529 | 80 |
| 2026-11-07 | 773a529 | 80 |