---
license: apache-2.0
library_name: lerobot
tags:
  - robotics
  - imitation-learning
  - behavior-cloning
  - manipulation
  - lerobot
datasets:
  - xinjiehu76/so101-pick-place-dataset
---

# SO-ARM101 ACT Pick-and-Place (LeRobot)

ACT (behavior cloning) policy trained with the LeRobot framework on a SO-ARM101 pick-and-place task.

## Task
Pick a red cube and place it into a gray bowl.

## Dataset
- https://huggingface.co/datasets/xinjiehu76/so101-pick-place-dataset  
  Subsets: `view_above`, `view_side`

## Files
- `model.safetensors`: trained weights
- `config.json`: model/config metadata
- `train_config.json`: training configuration

## Inference (LeRobot)
```bash
python -m lerobot.scripts.control_robot \
  --robot.type=so101_follower \
  --policy.path=xinjiehu76/so101-act-pick-place_ACT