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

# groot1.7_fold_tea_towel

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

| | |
|---|---|
| **Base model** | [nvidia/GR00T-N1.7-3B](https://huggingface.co/nvidia/GR00T-N1.7-3B) |
| **Dataset** | [villekuosmanen/armnetbench_fold_tea_towel](https://huggingface.co/datasets/villekuosmanen/armnetbench_fold_tea_towel) |
| **Task** | `fold_tea_towel` |
| **Training** | Isambard-AI GH200, batch 64, action horizon 16 |
| **W&B project** | [groot1.7_fold_tea_towel](https://wandb.ai/pravsels/groot1.7_fold_tea_towel) |
| **W&B run** | [qp91sjl0](https://wandb.ai/pravsels/groot1.7_fold_tea_towel/runs/qp91sjl0) |

## Checkpoints

| Step | Loss | Path |
|------|------|------|
| 6,000 | 0.0262 | `checkpoints/6000/pretrained_model/` |
| 4,000 | 0.0432 | `checkpoints/4000/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_fold_tea_towel/checkpoints/6000/pretrained_model"
```