---
datasets: staudi25/lego_pickup_mega_plus
library_name: lerobot
license: apache-2.0
model_name: pi05
pipeline_tag: robotics
tags:
- pi05
- lerobot
- robotics
---

# Model Card for pi05
lerobot-train --dataset.repo_id=staudi25/lego_pickup_mega_plus --job_name=pi05_so101_lego_pickup --policy.type=pi05 --policy.pretrained_path=staudi25/pi05_lego_pickup_mega --policy.device=cuda --wandb.enable=false --dataset.image_transforms.enable=true --log_freq=1 --steps=3000 --save_freq=3000 --policy.repo_id=staudi25/pi05_lego_mega_plus --policy.dtype=bfloat16 --output_dir=outputs/train/pi05_lego_mega_plus --policy.compile_model=true --policy.gradient_checkpointing=true --policy.freeze_vision_encoder=false --policy.train_expert_only=false --batch_size=32


<!-- Provide a quick summary of what the model is/does. -->


**π₀.₅ (Pi05) Policy**

π₀.₅ is a Vision-Language-Action model with open-world generalization, from Physical Intelligence. The LeRobot implementation is adapted from their open source OpenPI repository.

**Model Overview**

π₀.₅ represents a significant evolution from π₀, developed by Physical Intelligence to address a big challenge in robotics: open-world generalization. While robots can perform impressive tasks in controlled environments, π₀.₅ is designed to generalize to entirely new environments and situations that were never seen during training.

For more details, see the [Physical Intelligence π₀.₅ blog post](https://www.physicalintelligence.company/blog/pi05).


This policy has been trained and pushed to the Hub using [LeRobot](https://github.com/huggingface/lerobot).
See the full documentation at [LeRobot Docs](https://huggingface.co/docs/lerobot/index).

---

## How to Get Started with the Model

For a complete walkthrough, see the [training guide](https://huggingface.co/docs/lerobot/il_robots#train-a-policy).
Below is the short version on how to train and run inference/eval:

### Train from scratch

```bash
lerobot-train \
  --dataset.repo_id=${HF_USER}/<dataset> \
  --policy.type=act \
  --output_dir=outputs/train/<desired_policy_repo_id> \
  --job_name=lerobot_training \
  --policy.device=cuda \
  --policy.repo_id=${HF_USER}/<desired_policy_repo_id>
  --wandb.enable=true
```

_Writes checkpoints to `outputs/train/<desired_policy_repo_id>/checkpoints/`._

### Evaluate the policy/run inference

```bash
lerobot-record \
  --robot.type=so100_follower \
  --dataset.repo_id=<hf_user>/eval_<dataset> \
  --policy.path=<hf_user>/<desired_policy_repo_id> \
  --episodes=10
```

Prefix the dataset repo with **eval\_** and supply `--policy.path` pointing to a local or hub checkpoint.

---

## Model Details

- **License:** apache-2.0