New: 98,000+ skills indexed from GitHub repos with 100+ stars · every skill safety-checkedBrowse top skillsTop skills
Esc to closeBrowse all skills
Submit a skillBrowse skills
Submit a skillBrowse skills

Agent skill

OpenPI Fine-Tuning and Serving

by Orchestra-Research in Orchestra-Research/AI-Research-SKILLs

Fine-tunes and serves Physical Intelligence's pi0, pi0-fast and pi0.5 robot policies with JAX or PyTorch, including checkpoint conversion and policy servers.

MITAuto-check passedAI & LLM Engineering

Install OpenPI Fine-Tuning and Serving

skills CLI
$ npx skills add Orchestra-Research/AI-Research-SKILLs --skill fine-tuning-serving-openpi -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install Orchestra-Research/AI-Research-SKILLs fine-tuning-serving-openpi --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/Orchestra-Research/AI-Research-SKILLs.git skills-src && mkdir -p .claude/skills && cp -r skills-src/18-multimodal/openpi .claude/skills/fine-tuning-serving-openpi && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
fine-tuning-serving-openpi
GitHub stars
13k
Token cost
~3.6k tokens
SKILL.md length
821 words
Files
6 (incl. references)
Skills in repo
96
Repo updated
First seen
Licence
MIT

At a glance

Fine-tunes and serves Physical Intelligence's pi0, pi0-fast and pi0.5 robot policies with JAX or PyTorch, including checkpoint conversion and policy servers.

  • Adapting a pi0 model to a custom robot dataset
  • SKILL.md covers Quick start, Core concepts, Compute requirements and Workflow 0: Blank-machine setup, plus 8 more sections
  • Calls uv, git and pip; reaches github.com
  • Converting JAX checkpoints to PyTorch for deployment

What it does

This covers the whole loop for the public `openpi` repo: cloning with submodules, syncing the workspace with `uv`, computing normalization statistics, training in JAX (the primary, official backend) or PyTorch (community), converting JAX checkpoints to PyTorch, and serving a policy over a WebSocket API that a client reaches through `openpi_client`. Training and serving settings live in `src/openpi/training/config.py`, and every config needs normalization stats before training.

After any config or dataset change the cycle repeats: compute norm stats, train, serve the checkpoint, validate inference. The skill distinguishes pi0 (flow matching), pi0-fast (autoregressive action tokens, 2-5x faster) and pi0.5 (improved vision encoder), and gives memory guidance such as about 24 GB for serving pi0.5 and about 60 GB for JAX fine-tuning on an 80GB A100. Reference files cover checkpoints and environments for ALOHA, DROID and LIBERO, config recipes, PyTorch gotchas, a remote client pattern and training debugging.

When your agent uses it

  • Adapting a pi0 model to a custom robot dataset
  • Converting JAX checkpoints to PyTorch for deployment
  • Running a policy inference server for ALOHA, DROID or LIBERO
  • Debugging norm stats errors or GPU memory problems

Example prompts

  • “Compute norm stats for my dataset config and start a pi0.5 JAX fine-tune.”
  • “Convert this JAX checkpoint to PyTorch and serve it on port 8000.”
  • “My openpi client returns wrong actions after I changed the dataset, so find out why.”

Requirements

  • A GPU such as an A100 or H100
  • `uv` and the public `openpi` repository

What it can do on your machine

Read from SKILL.md and the folder at commit 773a529. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • uv
    • git
    • pip
    • gsutil

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    Hosts in commands or code, which the agent is likely to contact:

    • github.com

    Also links to:

    • physicalintelligence.company
    • huggingface.co

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

OpenPI Fine-Tuning and Serving loads about 3.6k tokens when it runs, and up to ~5.6k if it reads all its reference files. Until then it costs about 94 tokens; SKILL.md has 821 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~94
When it runs · the whole SKILL.md, loaded when a task matches
~3.6k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~5.6k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from Orchestra-Research/AI-Research-SKILLs at commit 773a529, republished under its MIT licence (© Orchestra-Research). 821 words, ~3,559 tokens.

Download SKILL.mdSave it as .claude/skills/fine-tuning-serving-openpi/SKILL.md (or your agent's skills folder). This skill also uses 5 other files; get the full folder from GitHub.
name
fine-tuning-serving-openpi
description
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.
version
1.0.0
author
Orchestra Research
license
MIT
tags
OpenPI, Physical Intelligence, VLA, Robotics, JAX, PyTorch, Fine-Tuning, Policy Serving, ALOHA, DROID, LIBERO, pi0
dependencies
uv>=0.4.0, jax>=0.4.30, torch>=2.1.0, transformers>=4.53.2

OpenPI Fine-Tuning and Serving

End-to-end workflows for fine-tuning and serving Physical Intelligence's OpenPI models (pi0, pi0-fast, pi0.5) on robot manipulation tasks from the public openpi repository. Covers blank-machine setup, JAX training, PyTorch training, checkpoint conversion, and policy inference serving.

Quick start

Clone the public repo, install the workspace, then serve a pretrained policy:

bash
git clone --recurse-submodules https://github.com/Physical-Intelligence/openpi.git
cd openpi
GIT_LFS_SKIP_SMUDGE=1 uv sync
GIT_LFS_SKIP_SMUDGE=1 uv pip install -e .
uv run scripts/serve_policy.py --env DROID
python
from openpi_client import websocket_client_policy

client = websocket_client_policy.WebsocketClientPolicy(host="localhost", port=8000)
result = client.infer(observation)
actions = result["actions"]  # numpy array of shape (chunk_size, action_dim)

Core concepts

Model family: OpenPI implements three model variants from Physical Intelligence:

ModelArchitectureSpeedQualityTypical use
pi0Flow-matching VLABaselineHighestResearch, complex tasks
pi0-fastAutoregressive action tokens2-5x fasterGoodReal-time control
pi0.5pi0 + improved vision encoderBaselineBestLatest default

Key design choices:

  • Dual backend: JAX (primary, official training) and PyTorch (community, deployment-friendly)
  • Config-driven: All training/serving parameters defined in src/openpi/training/config.py
  • Norm stats: Every config requires precomputed normalization statistics before training
  • WebSocket serving: Policy servers expose a WebSocket API for low-latency inference

Training loop invariant: After every config or dataset change, always re-run this cycle:

  1. Compute norm stats → 2. Train → 3. Serve checkpoint → 4. Validate inference

Compute requirements

TaskGPUVRAMNotes
Serve pi0.5 (inference)1x A100/H100~24 GBSingle GPU sufficient
Fine-tune pi0.5 (JAX)1x A100 80GB~60 GBUse fsdp_devices for multi-GPU
Fine-tune pi0 (JAX)1x A100 80GB~40 GBSmaller model footprint
Fine-tune (PyTorch DDP)1-8x A100~40 GB/GPUtorchrun launcher
Compute norm statsCPU or 1x GPU~8 GBFast, can run on login node

Workflow 0: Blank-machine setup

Copy this checklist and track progress:

text
Setup Progress:
- [ ] Step 1: Clone the public openpi repo with submodules
- [ ] Step 2: Install uv and sync the workspace
- [ ] Step 3: Install the editable package
- [ ] Step 4: Verify core imports and serving entrypoint

Step 1: Clone repo

bash
git clone --recurse-submodules https://github.com/Physical-Intelligence/openpi.git
cd openpi

If you already cloned without submodules:

bash
git submodule update --init --recursive

Step 2: Sync dependencies

bash
GIT_LFS_SKIP_SMUDGE=1 uv sync

Step 3: Install editable package

bash
GIT_LFS_SKIP_SMUDGE=1 uv pip install -e .

Step 4: Verify installation

bash
uv run python -c "from openpi.training import config as _config; print(_config.get_config('pi05_droid').name)"
uv run scripts/serve_policy.py --help

When to use vs alternatives

Use this skill when:

  • Fine-tuning pi0, pi0-fast, or pi0.5 on LeRobot or RLDS datasets
  • Serving OpenPI policies for ALOHA, DROID, or LIBERO evaluation
  • Converting JAX checkpoints to PyTorch format
  • Debugging OpenPI training issues (norm stats, memory, config)

Use fine-tuning-openvla-oft instead when:

  • Fine-tuning OpenVLA with continuous action heads and LoRA
  • Reproducing OpenVLA-OFT paper results on LIBERO or ALOHA

Use evaluating-cosmos-policy instead when:

  • Evaluating NVIDIA Cosmos Policy on simulation benchmarks

Workflow 1: JAX fine-tuning on LeRobot data

Copy this checklist and track progress:

text
JAX Fine-Tuning Progress:
- [ ] Step 1: Select and copy closest training config
- [ ] Step 2: Update dataset mapping and base checkpoint
- [ ] Step 3: Compute normalization statistics
- [ ] Step 4: Launch JAX training
- [ ] Step 5: Serve checkpoint and run inference sanity check

Step 1: Select config

Copy the closest config from src/openpi/training/config.py:

ConfigUse case
pi05_liberopi0.5 LIBERO fine-tuning
pi0_liberopi0 full fine-tuning on LIBERO
pi0_fast_liberopi0-fast on LIBERO
pi0_aloha_pen_uncapALOHA custom data
pi05_droid_finetuneSmall custom DROID dataset (LeRobot format)
pi05_full_droid_finetuneFull DROID RLDS large-scale training

Step 2: Update dataset and transforms

python
# In src/openpi/training/config.py, modify your config:
TrainConfig(
    name="my_custom_config",
    model_type="pi05",
    data=LeRobotDataConfig(
        repo_id="your-org/your-dataset",
        # Adjust transforms to match your data format
    ),
    weight_loader=Pi05WeightLoader(),  # Match model type
)

Set repo_id for your dataset and ensure weight_loader matches the model type (pi0 vs pi0.5).

Step 3: Compute normalization statistics

bash
uv run scripts/compute_norm_stats.py --config-name <config_name>

This must run before every training launch when config, dataset, or transforms change.

Step 4: Launch JAX training

bash
XLA_PYTHON_CLIENT_MEM_FRACTION=0.9 uv run scripts/train.py <config_name> \
  --exp-name=<run_name> \
  --overwrite

For full DROID RLDS training, add the rlds dependency group:

bash
uv run --group rlds scripts/compute_norm_stats.py \
  --config-name pi05_full_droid_finetune \
  --max-frames 10000000

XLA_PYTHON_CLIENT_MEM_FRACTION=0.9 uv run --group rlds scripts/train.py \
  pi05_full_droid_finetune \
  --exp-name=<run_name> --overwrite

Step 5: Serve and validate

bash
uv run scripts/serve_policy.py policy:checkpoint \
  --policy.config=<config_name> \
  --policy.dir=checkpoints/<config_name>/<run_name>/<step>

Verify with a test client:

python
from openpi_client import websocket_client_policy

client = websocket_client_policy.WebsocketClientPolicy(host="localhost", port=8000)
# Build observation matching your config's expected keys
obs = {"image": img_array, "state": state_array, "prompt": "pick up the cup"}
result = client.infer(obs)
print(f"Action shape: {result['actions'].shape}")  # (chunk_size, action_dim)

Workflow 2: PyTorch training and checkpoint conversion

Copy this checklist and track progress:

text
PyTorch Setup Progress:
- [ ] Step 1: Sync dependencies and verify transformer version
- [ ] Step 2: Apply OpenPI transformer patches
- [ ] Step 3: Convert JAX checkpoint to PyTorch format
- [ ] Step 4: Launch PyTorch training or serve converted checkpoint

Step 1: Sync dependencies

bash
uv sync
uv pip show transformers

Step 2: Apply required patches

OpenPI PyTorch requires custom modifications to the installed transformers package:

bash
cp -r ./src/openpi/models_pytorch/transformers_replace/* \
  .venv/lib/python3.11/site-packages/transformers/

Step 3: Convert JAX checkpoint

bash
uv run examples/convert_jax_model_to_pytorch.py \
  --checkpoint_dir <jax_checkpoint_dir> \
  --config_name <config_name> \
  --output_path <pytorch_checkpoint_dir>

Step 4: Train or serve

Single GPU training:

bash
uv run scripts/train_pytorch.py <config_name> --exp_name <run_name>

Multi-GPU distributed training:

bash
uv run torchrun --standalone --nnodes=1 --nproc_per_node=<num_gpus> \
  scripts/train_pytorch.py <config_name> --exp_name <run_name>

Programmatic inference with converted checkpoint:

python
from openpi.training import config as _config
from openpi.policies import policy_config

config = _config.get_config("pi05_droid")
policy = policy_config.create_trained_policy(config, "<pytorch_checkpoint_dir>")
result = policy.infer(example)
actions = result["actions"]  # numpy array

Checkpoints follow the convention: checkpoints/<config_name>/<exp_name>/<step>/.

Show full SKILL.md (310 more words)Show less

Workflow 3: Policy inference serving

Copy this checklist and track progress:

text
Inference Server Progress:
- [ ] Step 1: Choose target environment and checkpoint
- [ ] Step 2: Start policy server
- [ ] Step 3: Confirm server is reachable
- [ ] Step 4: Integrate client into robot or simulation code

Step 1: Choose environment

Default environment presets:

EnvironmentConfigDefault checkpoint
ALOHApi05_alohags://openpi-assets/checkpoints/pi05_base
ALOHA_SIMpi0_aloha_simgs://openpi-assets/checkpoints/pi0_aloha_sim
DROIDpi05_droidgs://openpi-assets/checkpoints/pi05_droid
LIBEROpi05_liberogs://openpi-assets/checkpoints/pi05_libero

Step 2: Start server

Default mode (uses preset checkpoint):

bash
uv run scripts/serve_policy.py --env ALOHA

Explicit checkpoint mode (custom or local model):

bash
uv run scripts/serve_policy.py policy:checkpoint \
  --policy.config=pi05_libero \
  --policy.dir=checkpoints/pi05_libero/my_run/20000

Add --default_prompt "task description" when runtime observations omit a prompt.

Step 3: Verify connectivity

bash
uv run examples/simple_client/main.py --env DROID

Step 4: Embed remote client in robot code

Install the lightweight client in your robot environment:

bash
pip install "openpi-client @ git+https://github.com/Physical-Intelligence/openpi.git#subdirectory=packages/openpi-client"

Full integration example:

python
from openpi_client import websocket_client_policy
import numpy as np

# Connect to remote policy server
client = websocket_client_policy.WebsocketClientPolicy(
    host="gpu-server.local", port=8000
)

# Build observation (keys must match policy transforms)
observation = {
    "image": np.random.rand(224, 224, 3),  # RGB image
    "state": np.zeros(7),                   # Joint positions
    "prompt": "pick up the red block",
}

# Get actions
result = client.infer(observation)
actions = result["actions"]  # shape: (action_chunk_size, action_dim)

# Execute first action on robot
robot.step(actions[0])

Common issues

Issue: Missing norm stats error

Fix: run scripts/compute_norm_stats.py --config-name <config_name> before training.

Issue: Out of memory during JAX training

Fix: set XLA_PYTHON_CLIENT_MEM_FRACTION=0.9, lower batch size, or configure fsdp_devices:

python
# In config: use model-parallel sharding
TrainConfig(
    ...
    fsdp_devices=4,  # Shard across 4 GPUs
)

Issue: OOM while loading PyTorch checkpoints

Fix: export PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True

Issue: Config not found

Fix: ensure config name exists in src/openpi/training/config.py (exact match from _CONFIGS dict).

Issue: PyTorch training diverges after library changes

Fix: reapply the transformer patch. Run uv cache clean transformers to reset, then reapply.

Issue: serve_policy.py crashes with ModuleNotFoundError

Fix: resync the public workspace first:

bash
GIT_LFS_SKIP_SMUDGE=1 uv sync
GIT_LFS_SKIP_SMUDGE=1 uv pip install -e .

If the missing module is simulator-related, install the extra runtime dependencies called for by that example:

bash
uv pip install pytest robosuite==1.4.0 gym bddl easydict matplotlib

Issue: uv sync fails with rerun-sdk wheel mismatch

Fix:

bash
uv sync --no-dev
# or
uv sync --no-dev --no-install-package rerun-sdk

Issue: Checkpoint download times out

Fix: install gsutil and prefetch manually:

bash
pip install gsutil
gsutil -m cp -r gs://openpi-assets/checkpoints/pi05_libero /local/cache/

Remove stale .lock files if a previous download was interrupted.

Issue: Policy server exits with code 137

Fix: OOM kill. Set JAX memory variables:

bash
export XLA_PYTHON_CLIENT_PREALLOCATE=false
export XLA_PYTHON_CLIENT_ALLOCATOR=platform

For HPC/cluster users

On Slurm-managed clusters, wrap commands with resource allocation:

bash
srun --partition=gpu --gpus-per-node=1 --mem=64G --cpus-per-task=8 --pty bash

Route caches to scratch to avoid filling /home:

bash
export HF_HOME=/scratch/$USER/.cache/huggingface
export XDG_CACHE_HOME=/scratch/$USER/.cache
export PIP_CACHE_DIR=/scratch/$USER/.cache/pip
export UV_CACHE_DIR=/scratch/$USER/.cache/uv

Avoid stacking cluster Python modules when using uv-managed environments. Typically module load cuda is sufficient.

Advanced topics

Config recipes and baselines: See references/config-recipes.md Training debugging guide: See references/training-debugging.md Checkpoint and environment mapping: See references/checkpoints-and-env-map.md Remote client integration: See references/remote-client-pattern.md PyTorch precision and patching gotchas: See references/pytorch-gotchas.md

Resources

© Orchestra-Research, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 5 other files (references) in 18-multimodal/openpi of Orchestra-Research/AI-Research-SKILLs.

  • SKILL.md
  • references/checkpoints-and-env-map.md
  • references/config-recipes.md
  • references/pytorch-gotchas.md
  • references/remote-client-pattern.md
  • references/training-debugging.md

Open the folder on GitHubat commit 773a529

Compare with similar skills

OpenPI Fine-Tuning and Serving next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

OpenPI Fine-Tuning and Serving compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
OpenPI Fine-Tuning and Serving this skillOrchestra-Research/AI-Research-SKILLs13k—~3.6kAutomated safety check: PassMIT
Coreweave Core Workflow Bjeremylongshore/tons-of-skills-marketplace2.8k—~1.2kAutomated safety check: PassMIT
MUSA GPU Training Optimizeropen-infra-skills/infra-skills141—~1.7kAutomated safety check: PassApache-2.0
PyTorch Lightning Training Setupdavila7/claude-code-templates33k11 repos~1.7kAutomated safety check: PassMIT
Model ScaffoldAperivue/medsci-skills333—~3.1kAutomated safety check: PassMIT
GPU OptimizerMathews-Tom/armory329—~3.5kAutomated safety check: NotesMIT

Similar skills

  • Coreweave Core Workflow B

    jeremylongshore/tons-of-skills-marketplace

    Run distributed GPU training jobs on CoreWeave with multi-node PyTorch.

    2.8k GitHub stars~1.2k tokensUpdated today
    AI & LLM EngineeringAuto-check passed

  • MUSA GPU Training Optimizer

    open-infra-skills/infra-skills

    Profiles, benchmarks and tunes AI training workloads on Moore Threads MUSA GPUs with a measurement-first process that keeps model behavior unchanged.

    141 GitHub stars~1.7k tokensUpdated 3 mo ago
    AI & LLM EngineeringAuto-check passed

  • PyTorch Lightning Training Setup

    davila7/claude-code-templates

    Organizes PyTorch training code into LightningModules, DataModules and Trainers, with multi-GPU strategies, callbacks and logging configured.

    33k GitHub starsUsed in 11 repos~1.7k tokens
    AI & LLM EngineeringAuto-check passed

  • Model Scaffold

    Aperivue/medsci-skills

    A skill your agent uses when you need a runnable PyTorch training repo for a medical-imaging task (segmentation, classification, detection, synthesis, self-supervised, or fine-tuning a pretrained…

    333 GitHub stars~3.1k tokensUpdated 5 days ago
    AI & LLM EngineeringAuto-check passed

  • GPU Optimizer

    Mathews-Tom/armory

    GPU optimization for consumer NVIDIA GPUs (8-24GB VRAM) covering mixed precision, gradient checkpointing, XGBoost GPU, CuPy/cuDF migration, and torch.compile.

    329 GitHub stars~3.5k tokensUpdated 4 days ago
    AI & LLM EngineeringAuto-check: notes

  • Cuda Index Width

    pytorch/pytorch

    Choose 32-bit vs 64-bit index math in PyTorch CUDA kernels. An agent skill from pytorch/pytorch.

    104k GitHub stars~1.6k tokensUpdated today
    AI & LLM EngineeringAuto-check passed

More from Orchestra-Research/AI-Research-SKILLs

All 96 skills in this repo

  • AudioCraft Audio Generation

    Orchestra-Research/AI-Research-SKILLs

    Generates music from text descriptions with MusicGen and sound effects with AudioGen, using Meta's AudioCraft PyTorch library with melody and style conditioning.

    13k GitHub starsUsed in 8 repos~3.9k tokens
    Auto-check passed

  • LLM Benchmarking with lm-evaluation-harness

    Orchestra-Research/AI-Research-SKILLs

    Runs lm-evaluation-harness to benchmark language models on academic suites such as MMLU, GSM8K and HumanEval, compare models and track training checkpoints.

    13k GitHub starsUsed in 8 repos~3k tokens
    Auto-check passed

  • Segment Anything Model Guide

    Orchestra-Research/AI-Research-SKILLs

    Guide to using Meta's Segment Anything Model for zero-shot image segmentation with point, box or mask prompts, or automatic mask generation.

    13k GitHub starsUsed in 8 repos~3.3k tokens
    Auto-check passed

  • Chroma Vector Database

    Orchestra-Research/AI-Research-SKILLs

    Shows how to store documents and embeddings in Chroma, query them by similarity with metadata filters, and persist them to disk for RAG and semantic search projects.

    13k GitHub starsUsed in 7 repos~2.3k tokens
    Auto-check passed

  • CLIP Image-Text Matching

    Orchestra-Research/AI-Research-SKILLs

    Explains OpenAI's CLIP model for zero-shot image classification, image-text similarity, semantic image search and content moderation, with install steps and code patterns.

    13k GitHub starsUsed in 7 repos~1.7k tokens
    Auto-check passed

  • Whisper Speech Recognition

    Orchestra-Research/AI-Research-SKILLs

    Transcribes audio with OpenAI's Whisper: 99 languages, translation to English, language detection, six model sizes and word-level timestamps, from Python or the CLI.

    13k GitHub starsUsed in 7 repos~1.9k tokens
    Auto-check: notes

Works with

Categories

Questions about OpenPI Fine-Tuning and Serving

What does OpenPI Fine-Tuning and Serving do?

Fine-tunes and serves Physical Intelligence's pi0, pi0-fast and pi0.5 robot policies with JAX or PyTorch, including checkpoint conversion and policy servers. This covers the whole loop for the public `openpi` repo: cloning with submodules, syncing the workspace with `uv`, computing normalization statistics, training in JAX (the primary, official backend) or PyTorch (community), converting JAX checkpoints to PyTorch, and serving a policy over a WebSocket API that a client reaches through `openpi_client`.py`, and every config needs normalization stats before training.

When should I use OpenPI Fine-Tuning and Serving?

OpenPI Fine-Tuning and Serving fits situations like: adapting a pi0 model to a custom robot dataset; converting JAX checkpoints to PyTorch for deployment; running a policy inference server for ALOHA, DROID or LIBERO; debugging norm stats errors or GPU memory problems.

How do I install OpenPI Fine-Tuning and Serving in Claude Code?

Run `npx skills add Orchestra-Research/AI-Research-SKILLs --skill fine-tuning-serving-openpi -a claude-code`. Or copy the skill folder (18-multimodal/openpi in Orchestra-Research/AI-Research-SKILLs) into .claude/skills/fine-tuning-serving-openpi in your project. Claude Code loads it when a task matches its description.

How do I install OpenPI Fine-Tuning and Serving in Codex?

Run `npx skills add Orchestra-Research/AI-Research-SKILLs --skill fine-tuning-serving-openpi -a codex`. Or copy the skill folder (18-multimodal/openpi in Orchestra-Research/AI-Research-SKILLs) into .agents/skills/fine-tuning-serving-openpi in your project. Codex loads it when a task matches its description.

Can I use OpenPI Fine-Tuning and Serving in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add Orchestra-Research/AI-Research-SKILLs --skill fine-tuning-serving-openpi -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/fine-tuning-serving-openpi, .gemini/skills/fine-tuning-serving-openpi, .github/skills/fine-tuning-serving-openpi and .opencode/skills/fine-tuning-serving-openpi in your project.

What does OpenPI Fine-Tuning and Serving need to run?

Going by SKILL.md and its folder, OpenPI Fine-Tuning and Serving needs the command-line tools its instructions call (uv, git, pip and gsutil). Our summary lists: A GPU such as an A100 or H100; `uv` and the public `openpi` repository.

Does OpenPI Fine-Tuning and Serving access the network?

SKILL.md names 3 domains. In commands or code: github.com; the agent is likely to contact it when it follows the instructions. As links in the text: physicalintelligence.company and huggingface.co. This is read from the text; nothing was executed.

Is OpenPI Fine-Tuning and Serving safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does OpenPI Fine-Tuning and Serving use?

OpenPI Fine-Tuning and Serving is published under the MIT licence (declared in SKILL.md). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does OpenPI Fine-Tuning and Serving use?

About 3.6k tokens (SKILL.md is roughly 14k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 2.1k tokens, read only when the agent opens those files.

What are the alternatives to OpenPI Fine-Tuning and Serving?

Skills that share tags, products or a category with OpenPI Fine-Tuning and Serving: Coreweave Core Workflow B (jeremylongshore/tons-of-skills-marketplace, 2.8k stars), MUSA GPU Training Optimizer (open-infra-skills/infra-skills, 141 stars), PyTorch Lightning Training Setup (davila7/claude-code-templates, 33k stars) and Model Scaffold (Aperivue/medsci-skills, 333 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains OpenPI Fine-Tuning and Serving?

Orchestra-Research (a GitHub organization) maintains it in Orchestra-Research/AI-Research-SKILLs, which has 13,405 GitHub stars. The repository holds 96 skills in this directory. The repository was last updated on June 16, 2026.

Source: Orchestra-Research/AI-Research-SKILLs on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.