---
library_name: lerobot
tags:
- robotics
- pi05
- lerobot
datasets:
- RDLwicked/plug_and_unplug_socket_oneEP_May_6th
---

# plug_and_unplug_socket_oneEP_May_6th_4gpu

Pi0.5 policy fine-tuned with LeRobot.

- **Training step:** 80000
- **Dataset:** [`RDLwicked/plug_and_unplug_socket_oneEP_May_6th`](https://huggingface.co/datasets/RDLwicked/plug_and_unplug_socket_oneEP_May_6th)
- **Pretrained from:** `lerobot/pi05_base`

## Files

Standard LeRobot pi0.5 checkpoint layout:

| File | Purpose |
|---|---|
| `model.safetensors` | Policy weights |
| `config.json` | Policy config |
| `policy_preprocessor.*` | Input normalization stats |
| `policy_postprocessor.*` | Output unnormalization stats |
| `train_config.json` | Training config snapshot |

## Loading

```python
from lerobot.policies.pi05.modeling_pi05 import PI05Policy
policy = PI05Policy.from_pretrained("mcarroll101/plug_and_unplug_socket_oneEP_May_6th_4gpu")
```

_(Manually pushed from `/lus/lfs1aip2/projects/u6ed/milo/data/gim_arm/plug_and_unplug_socket_oneEP_May_6th_4gpu/checkpoints/080000/pretrained_model` via `scripts/misc/push_pi05_checkpoint.py`.)_