(lerobot) rlawlstjd@rlawlstjd:~/lerobot$ python -m lerobot.scripts.lerobot_train
--dataset.repo_id asdl-unist/pick_bluecup_and_remove-480p
--dataset.repo_id asdl-unist/arrange_cup
--output_dir="outputs/pi0_clean_training"
--policy.type=pi0
--policy.pretrained_path="lerobot/pi0_base"
--policy.dtype=bfloat16
--policy.device=cuda
--policy.gradient_checkpointing=true
--steps=10000
--batch_size=1
--log_freq=50
--save_checkpoint=true
--save_freq=500
--peft.method_type="lora"
--peft.r=32
--peft.target_modules="all-linear"
--policy.push_to_hub=true
--policy.repo_id="asdl-unist/pi0_arrange_cup_and_remove_blue_clean"
--wandb.enable=true
--wandb.project="lerobot-pi0_arrange_cup_and_remove_blue"

π₀ is a Vision-Language-Action model for general robot control, from Physical Intelligence. The LeRobot implementation is adapted from their open source OpenPI repository.

π₀ 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.

For more details, see the Physical Intelligence π₀ blog post.

This policy has been trained and pushed to the Hub using LeRobot. See the full documentation at LeRobot Docs.

For a complete walkthrough, see the training guide. Below is the short version on how to train and run inference/eval:

Prefix the dataset repo with eval_ and supply --policy.path pointing to a local or hub checkpoint.