---
library_name: lerobot
tags:
- act
- robotics
- lerobot

---

# act_eyedrops_basket_20k

LeRobot ACT policy for `eyedrops_basket` on SO101 data.

| | |
|---|---|
| **Policy** | ACT (`policy.type=act`) |
| **Dataset** | [pravsels/object_top_shelf_reset_remote](https://huggingface.co/datasets/pravsels/object_top_shelf_reset_remote) |
| **Task** | `eyedrops_basket` |
| **Action dim** | 6 (single-arm) |
| **Cameras** | `top`, `wrist`, `front` |
| **Training** | 20k total: 10k scratch + 10k warm-start @ bs16 on GCloud A100 (phase-1: pravsels/act_eyedrops_basket) |
| **Phase-1 checkpoint** | [pravsels/act_eyedrops_basket](https://huggingface.co/pravsels/act_eyedrops_basket) |
| **W&B project** | [act_eyedrops_basket](https://wandb.ai/pravsels/act_eyedrops_basket) |
| **W&B run** | [w3rw19it](https://wandb.ai/pravsels/act_eyedrops_basket/runs/w3rw19it) |

## Checkpoints

The final checkpoint (local step `010000`, 20k total steps) lives at the **repository root** for direct loading.

## Usage

```python
from lerobot.policies.act.modeling_act import ACTPolicy
policy = ACTPolicy.from_pretrained("pravsels/act_eyedrops_basket_20k")
```