lerobot-data-collection/level2_final_quality2

1---2datasets: lerobot-data-collection/level2_final_quality23library_name: lerobot4license: apache-2.05model_name: pi06pipeline_tag: robotics7tags:8- lerobot9- robotics10- pi011---12 13# Model Card for pi014 15<!-- Provide a quick summary of what the model is/does. -->16 17 18**π₀ (Pi0)**19 20π₀ is a Vision-Language-Action model for general robot control, from Physical Intelligence. The LeRobot implementation is adapted from their open source OpenPI repository.21 22**Model Overview**23 24π₀ represents a breakthrough in robotics as the first general-purpose robot foundation model developed by Physical Intelligence. Unlike traditional robots that are narrow specialists programmed for repetitive motions, π₀ is designed to be a generalist policy that can understand visual inputs, interpret natural language instructions, and control a variety of different robots across diverse tasks.25 26For more details, see the [Physical Intelligence π₀ blog post](https://www.physicalintelligence.company/blog/pi0).27 28 29This policy has been trained and pushed to the Hub using [LeRobot](https://github.com/huggingface/lerobot).30See the full documentation at [LeRobot Docs](https://huggingface.co/docs/lerobot/index).31 32---33 34## How to Get Started with the Model35 36For a complete walkthrough, see the [training guide](https://huggingface.co/docs/lerobot/il_robots#train-a-policy).37Below is the short version on how to train and run inference/eval:38 39### Train from scratch40 41```bash42lerobot-train \43  --dataset.repo_id=${HF_USER}/<dataset> \44  --policy.type=act \45  --output_dir=outputs/train/<desired_policy_repo_id> \46  --job_name=lerobot_training \47  --policy.device=cuda \48  --policy.repo_id=${HF_USER}/<desired_policy_repo_id>49  --wandb.enable=true50```51 52_Writes checkpoints to `outputs/train/<desired_policy_repo_id>/checkpoints/`._53 54### Evaluate the policy/run inference55 56```bash57lerobot-record \58  --robot.type=so100_follower \59  --dataset.repo_id=<hf_user>/eval_<dataset> \60  --policy.path=<hf_user>/<desired_policy_repo_id> \61  --episodes=1062```63 64Prefix the dataset repo with **eval\_** and supply `--policy.path` pointing to a local or hub checkpoint.65 66---67 68## Model Details69 70- **License:** apache-2.0