---
datasets:
- lerobot/aloha_sim_transfer_cube_scripted
base_model:
- lerobot/pi0
---

LeRobot version of Pi0 fine-tuned for 20k steps on the `lerobot/aloha_sim_transfer_cube_scripted_image` dataset.

Usage:
```python
policy = PI0Policy.from_pretrained("BrunoM42/pi0_aloha_transfer_cube")
action = policy.select_action(batch)
```

Run evaluation:
```python
python lerobot/scripts/eval.py  --policy.path=BrunoM42/pi0_aloha_transfer_cube  --env.type=aloha  --env.task=AlohaTransferCube-v0  --env.episode_length=600  --eval.n_episodes=10  --eval.batch_size=10  --seed=42
```