g1-sonic-base / README.md
BlackCatRobotics's picture
Upload README.md with huggingface_hub
5ff8523 verified
|
Raw
History Blame Contribute Delete
2.71 kB
---
license: other
task_categories:
- robotics
- reinforcement-learning
tags:
- humanoid
- unitree-g1
- whole-body-control
- teleoperation
- vla
- onnx
- nvidia
- locomotion
- manipulation
size_categories:
- n<1K
---
# GEAR-SONIC: Whole-Body Humanoid Control
**SONIC (Supersizing Motion Tracking)** โ€” NVIDIA's humanoid behavior foundation model for natural whole-body control. Provides a unified policy for walking, running, jumping, manipulation, and more.
## Model Details
| Field | Value |
|---|---|
| **License** | NVIDIA Open Model License (commercial use โœ…) |
| **Robot** | Unitree G1 (29-DoF) |
| **Input** | Observations (joint states, IMU, motion references) |
| **Output** | 29-dim joint position targets |
| **Control Frequency** | 50Hz |
| **Format** | ONNX (encoder + decoder + planner) |
## Variants
| Variant | Use Case | Lookahead |
|---|---|---|
| **default** | General-purpose motion tracking | 200ms |
| **low_latency** | Teleoperation, VLA execution | 80ms |
| **sonic_v1_1** | Heading-normalized teleoperation | 200ms |
## Quick Start
```python
from huggingface_hub import snapshot_download
import onnxruntime as ort
import numpy as np
# Download model
snapshot_download(repo_id="BlackCatRoboticsAI/g1-sonic-base", local_dir="./g1-sonic")
# Load models
encoder = ort.InferenceSession("./g1-sonic/model_encoder.onnx")
decoder = ort.InferenceSession("./g1-sonic/model_decoder.onnx")
# Run inference
obs = np.random.randn(1, 1762).astype(np.float32)
tokens = encoder.run(None, {"obs_dict": obs})
action = decoder.run(None, {"obs_dict": obs})
```
## Features
- ๐Ÿค– **Unified Whole-Body Control**: Single policy handles walking, running, crawling, jumping, manipulation
- ๐ŸŽฎ **Real-Time Teleoperation**: VR-based whole-body teleoperation via PICO headset
- ๐Ÿš€ **Hardware Deployment**: C++ inference stack, runs on Jetson
- ๐ŸŽจ **Kinematic Planner**: Real-time locomotion with multiple movement styles
- ๐Ÿ”„ **Multi-Modal Control**: Keyboard, gamepad, VR, high-level planning
## Teleoperation
SONIC supports real-time whole-body teleoperation via PICO VR headset:
- Walking, running, sideways movement
- Kneeling, getting up, jumping
- Bimanual manipulation, object hand-off
## Citation
```bibtex
@article{luo2025sonic,
title={SONIC: Supersizing Motion Tracking for Natural Humanoid Whole-Body Control},
author={Luo, Zhengyi and Yuan, Ye and Wang, Tingwu and others},
journal={arXiv preprint arXiv:2511.07820},
year={2025}
}
```
## License
NVIDIA Open Model License โ€” commercial use permitted with attribution. See [NVIDIA Open Model License](https://developer.download.nvidia.com/licenses/NVIDIA-OneWay-Noncommercial-License-22Mar2022.pdf) for full terms.