cmwaters/pi05_go_board_joint_v1

Model Card for pi05

π₀.₅ (Pi05) is a Vision-Language-Action model from Physical Intelligence designed for open-world generalization: it evolves π₀ to generalize to entirely new environments and situations that were never seen during training. The LeRobot implementation is adapted from their open-source OpenPI repository.

This policy has been trained and pushed to the Hub using LeRobot.

Learn how to train and run it in the LeRobot pi05 guide, or browse the full documentation.

---

Model Details

• License: apache-2.0
• Fine-tuned from: lerobot/pi05_base
• Robot type: so101
• Cameras: overhead, wrist

Inputs & Outputs

The policy consumes these observation features and produces these action features.

Inputs

Feature	Type	Shape
observation.images.overhead	VISUAL	(3, 224, 224)
observation.images.wrist	VISUAL	(3, 224, 224)
observation.state	STATE	(6,)

Outputs

Feature	Type	Shape
action	ACTION	(6,)

Training Dataset

• Repository: cmwaters/go_board_joint_v1
• Episodes: 169
• Frames: 33341
• Frame rate: 10 FPS
• Task(s): "place black stone at D4", "place white stone at D10", "place black stone at D16", "place white stone at K4", "place black stone at K10", "place white stone at K16", "place black stone at Q4", "place white stone at Q10", "place black stone at Q16", "place white stone at F7", "place black stone at T12", "place white stone at J6", "place black stone at H7", "place white stone at F6", "place black stone at L10", "place white stone at M9", "place black stone at D17", "place white stone at E12", "place black stone at G7", "place black stone at J10", "place white stone at M10", "place black stone at H5", "place black stone at L7", "place white stone at R8", "place black stone at R9", "place white stone at F13", "place black stone at T18", "place black stone at S9", "place black stone at G6", "place white stone at M7", "place black stone at M6", "place white stone at E14", "place black stone at E13", "place white stone at K11", "place black stone at F12", "place white stone at M1", "place black stone at Q11", "place white stone at L18", "place black stone at G5", "place white stone at L8", "place white stone at O13", "place black stone at O14", "place black stone at F17", "place white stone at M5", "place white stone at N5", "place black stone at O5", "place black stone at N7", "place white stone at O6", "place black stone at O7", "place black stone at N6", "place white stone at N14", "place black stone at T1", "place white stone at E7", "place black stone at F16", "place white stone at F18", "place black stone at N13", "place white stone at C2", "place black stone at F8", "place white stone at P8", "place black stone at J14", "place black stone at J5", "place black stone at A10", "place white stone at C12", "place black stone at B2", "place white stone at K8", "place black stone at C15", "place white stone at D14", "place black stone at E9", "place white stone at D9", "place black stone at Q14", "place white stone at T5", "place black stone at E2", "place white stone at F2", "place black stone at F1", "place white stone at M15", "place black stone at T6", "place white stone at P15", "place black stone at M12", "place white stone at L10", "place white stone at S11", "place black stone at J11", "place white stone at H16", "place black stone at P8", "place white stone at B5", "place black stone at A6", "place white stone at L11", "place black stone at L2", "place black stone at P3", "place white stone at N2", "place black stone at H6", "place white stone at O18", "place white stone at A19", "place black stone at A18", "place white stone at A12", "place white stone at H5", "place black stone at E18", "place white stone at L12", "place black stone at L13", "place black stone at O3", "place white stone at P7", "place black stone at L6", "place white stone at K10", "place white stone at Q7", "place black stone at K9", "place white stone at J12", "place black stone at K12", "place white stone at O7", "place white stone at D16", "place black stone at N3", "place white stone at M2", "place black stone at T14", "place white stone at R14", "place white stone at P16", "place black stone at Q15", "place white stone at N17", "place black stone at O17", "place white stone at O11", "place black stone at N12", "place white stone at O12", "place black stone at O1", "place white stone at Q4", "place black stone at Q5", "place white stone at R5", "place black stone at C8", "place white stone at A10", "place black stone at B14", "place white stone at B15", "place white stone at C14", "place black stone at H10", "place white stone at G10", "place white stone at H15", "place black stone at E14", "place white stone at J1", "place black stone at H2", "place white stone at G19", "place black stone at N10", "place white stone at H17", "place black stone at J16", "place black stone at G12", "place white stone at G13", "place black stone at M5", "place white stone at E17", "place black stone at P18", "place white stone at C5", "place black stone at D6", "place white stone at Q1", "place white stone at N11", "place black stone at C17", "place white stone at C18", "place black stone at J8", "place white stone at H8", "place black stone at T10", "place black stone at K13", "place white stone at J19", "place black stone at H15", "place white stone at G14", "place white stone at P6", "place black stone at R15"

Training Configuration

Setting	Value
Training steps	30000
Batch size	1
Optimizer	adamw
Learning rate	2.5e-05
Seed	1000
LeRobot version	0.5.2

---

How to Get Started with the Model

New to LeRobot? These guides cover the full workflow:

• Install LeRobot — set up the lerobot package.
• Hardware setup — assemble, wire, and calibrate your robot and cameras.
• Record data & train a policy — the end-to-end imitation-learning walkthrough.
• CLI cheat-sheet — quick reference for the lerobot-* commands.

The short version to run and train this policy:

Run the policy on your robot

lerobot-rollout \
  --strategy.type=base \
  --robot.type=so101 \
  --robot.port=<your_robot_port> \
  --robot.cameras="{ <camera_1>: {type: opencv, index_or_path: <index_or_path>, width: 640, height: 480, fps: 30}, <camera_2>: {type: opencv, index_or_path: <index_or_path>, width: 640, height: 480, fps: 30}}" \
  --policy.path=cmwaters/pi05_go_board_joint_v1 \
  --task="place black stone at D4" \
  --duration=60

Replace the remaining <...> placeholders with your own values: --robot.port and the camera names/indices are specific to your machine, and the camera names must match the observation keys this policy was trained on.

When --strategy.type=base is used the script doesn't record the episodes. Skipping duration will make the policy run indefinitely. For more information look at rollout documentation.

Train your own policy

This policy type is usually fine-tuned from the pretrained base model lerobot/pi05_base:

lerobot-train \
  --dataset.repo_id=${HF_USER}/<dataset> \
  --policy.path=lerobot/pi05_base \
  --output_dir=outputs/train/<policy_repo_id> \
  --job_name=lerobot_training \
  --policy.device=cuda \
  --policy.repo_id=${HF_USER}/<policy_repo_id> \
  --wandb.enable=true

Writes checkpoints to outputs/train/<policy_repo_id>/checkpoints/.

---

Evaluation

No evaluation results have been provided for this policy yet.

---

Citation

If you use this policy, please cite the method linked in the description above, along with LeRobot:

@misc{cadene2024lerobot,
    author = {Cadene, Remi and Alibert, Simon and Soare, Alexander and Gallouedec, Quentin and Zouitine, Adil and Palma, Steven and Kooijmans, Pepijn and Aractingi, Michel and Shukor, Mustafa and Aubakirova, Dana and Russi, Martino and Capuano, Francesco and Pascal, Caroline and Choghari, Jade and Moss, Jess and Wolf, Thomas},
    title = {LeRobot: State-of-the-art Machine Learning for Real-World Robotics in Pytorch},
    howpublished = "\url{https://github.com/huggingface/lerobot}",
    year = {2024}
}