Downloads · 30 days
9
18% of all-time downloads
hi-space/GR00T-N1.7-3B-Pick-Banana-v2
GR00T-N1.7-3B-Pick-Banana-v2 is a robotics model from hi-space. Use it for the robotics task on the model card, and read the license before you ship it in a product. The card lists the license as other.
A fine-tuned version of nvidia/GR00T-N1.7-3B for fruit pick-and-place on a SO-ARM 101, trained on the jinseonylee/SO101PickAndPlaceFruit dataset.
Downloads · 30 days
9
18% of all-time downloads
All-time downloads
49
Public
Parameters
3.1B
25.5 GB on disk
Likes
0
Public
Click a slice to open those files.
.pt13 GB · 51%
From the Hugging Face model README
A fine-tuned version of nvidia/GR00T-N1.7-3B for fruit pick-and-place on a SO-ARM 101, trained on the jinseonylee/SO101_PickAndPlace_Fruit dataset.
GR00T-N1.7 (Gr00tN1d7) is a vision-language-action (VLA) model for robot manipulation. This checkpoint is fine-tuned for a pick-and-place task on a SO-ARM 101 arm.
| Parameter | Value |
|---|---|
| Tuned components | Diffusion model, projector (vision encoder + LLM backbone frozen) |
| Training steps | 13000 |
| Global batch size | 32 |
| Learning rate | 1e-4 (cosine, warmup ratio 0.05) |
| Weight decay | 1e-5 |
| State dropout prob | 0.2 |
| Optimizer | adamw_torch |
| Precision | bf16 |
| Final training loss | ~0.024 (avg train_loss 0.060) |
| Action representation | single_arm relative, gripper absolute |
| Attention | Flash Attention 2 |
new_embodiment)State inputs:
single_arm — arm joint positions (relative)gripper — gripper position (absolute)Action outputs: same as state inputs (16-step horizon)
Cameras: front, wrist
Language conditioning: annotation.human.task_description
Refer to the NVIDIA Isaac GR00T repository for full inference and deployment instructions. Open-loop evaluation:
uv run python gr00t/eval/open_loop_eval.py \
--dataset-path <path-to>/SO101_PickAndPlace_Fruit_lerobot \
--embodiment-tag NEW_EMBODIMENT \
--model-path <path-to>/checkpoint-13000 \
--traj-ids 0 --execution-horizon 16 --steps 400
This model is intended for SO-ARM 101 fruit pick-and-place tasks matching the training dataset's setup. It is not guaranteed to transfer zero-shot to different scenes, objects, or hardware without additional fine-tuning.
This model inherits the license from the base model nvidia/GR00T-N1.7-3B / nvidia/Cosmos-Reason2-2B. Please refer to NVIDIA's terms for usage restrictions.