groot1.7_fold_tea_towel_20k

Fine-tuned GR00T N1.7-3B for fold_tea_towel on SO101 data.

Base model	nvidia/GR00T-N1.7-3B
Dataset	pravsels/groot_fold_tea_towel_v2_h264
Task	fold_tea_towel
Embodiment	bimanual SO-101
Training	20k steps, 4x RTX 5090, global_batch_size=32
W&B project	groot1.7_fold_tea_towel
W&B run	o729haau

Checkpoints

Final checkpoint (step 20000) lives at the repository root for direct loading.

Usage

Load with the GR00T inference stack from this repo root (weights + processor configs).