---
library_name: gr00t
tags:
- gr00t
- gr00t-n1.7
- robotics
- vla
base_model: nvidia/GR00T-N1.7-3B
---

# groot1.7_eyedrops_basket

Fine-tuned [GR00T N1.7-3B](https://huggingface.co/nvidia/GR00T-N1.7-3B) on SO101 data (`eyedrops_basket`).

| | |
|---|---|
| **Base model** | [nvidia/GR00T-N1.7-3B](https://huggingface.co/nvidia/GR00T-N1.7-3B) |
| **Dataset** | [pravsels/object_top_shelf_reset_remote](https://huggingface.co/datasets/pravsels/object_top_shelf_reset_remote) |
| **Task** | `eyedrops_basket` |
| **Training** | Isambard-AI GH200, batch 64, action horizon 16 |
| **W&B project** | [groot1.7_eyedrops_basket](https://wandb.ai/pravsels/groot1.7_eyedrops_basket) |
| **W&B run** | [0iv3tnd4](https://wandb.ai/pravsels/groot1.7_eyedrops_basket/runs/0iv3tnd4) |

## Checkpoints

| Step | Loss | Path |
|------|------|------|
| 10,000 | 0.0174 | `checkpoints/10000/pretrained_model/` |
| 8,000 | 0.0225 | `checkpoints/8000/pretrained_model/` |

## Usage

Load from a specific step:

```python
from pathlib import Path
# Use checkpoints/<step>/pretrained_model/ with the GR00T inference stack
ckpt = "pravsels/groot1.7_eyedrops_basket/checkpoints/10000/pretrained_model"
```