# Quantized Visual Geometry Grounded Transformer
------
Weilun Feng1,2∗, Haotong Qin3∗, Mingqiang Wu1,2∗, Chuanguang Yang1†, Yuqi Li1, Xiangqi Li1,2, Zhulin An1†, Libo Huang1, Yulun Zhang4, Michele Magno3, Yongjun Xu1
*Equal Contribution †Corresponding Author
1.Institute of Computing Technology, Chinese Academy of Sciences, 2.University of Chinese Academy of Sciences,3.ETH Z¨ urich, 4.Shanghai Jiao Tong University
## 📰 News
- **[2026.01.26]** 😉 The paper has been accepted by ICLR 2026.
- **[2025.10.10]** 🎉 Paper and code released! Check out our [paper]([Quantized Visual Geometry Grounded Transformer](https://arxiv.org/pdf/2509.21302)).
## 🚀 Updates
- **[2026.02.09]** 😉The calibration dataset and quantized weights have been updated in the Hugging Face repository. Please [check](https://huggingface.co/wlfeng/QuantVGGT).
- **[2026.02.08]** 🎉Code for calibration training, evaluation on the 7-Scene and NRGBD datasets, and calibration set selection is now available.
- **[2025.10.10]** 🎉Evaluation code for reproducing our camera pose estimation results on Co3D is now available.
------
## 🌟Results
## 🛠️ Installation
First, clone this repository to your local machine, and install the dependencies (torch, torchvision, numpy, Pillow, and huggingface_hub).
```
git clone git@github.com:wlfeng0509/QuantVGGT.git
cd QuantVGGT
pip install -r requirements.txt
pip install -r requirements_demo.txt
```
Then download the pre trained weights provided by [VGGT](https://github.com/facebookresearch/vggt) and prepare Co3D dataset following [this](https://github.com/facebookresearch/vggt/tree/evaluation/evaluation).
Then download the pre trained W4A4 quantization parameters from [huggingface](https://huggingface.co/wlfeng/QuantVGGT/tree/main) and place the downloaded folder under ***evaluation\outputs\w4a4*** branch.
Then download the calibration set from [huggingface](https://huggingface.co/wlfeng/QuantVGGT/tree/main) and place the downloaded folder under ***evaluation\outputs*** branch.
## 📊 Quick start
We can now use the provided script for inference **(remember to change the data path within the script)**.
```
cd evaluation
bash make_calibation.sh # Filter and Save Calibration Set
bash run_co3d.sh # Calibration Training and Evaluation on Co3D
```
**Generate filtered Co3d calibration data**
```
python Quant_VGGT/vggt/evaluation/make_calibation.py \
--model_path VGGT-1B/model_tracker_fixed_e20.pt \
--co3d_dir co3d_datasets/ \
--co3d_anno_dir co3d_v2_annotations/ \
--seed 0 \
--cache_path all_calib_data.pt \ # Data to be filtered for calibration set
--save_path calib_data.pt \ # Save path for calibration set
--class_mode all \ # Category selection mode for calibration data
--kmeans_n 6 \ # Number of cluster centers
--kmeans_m 7 \ # Number of samples per category
```
**Quantize calibrate and evaluate on Co3d.**
```
python Quant_VGGT/vggt/evaluation/run_co3d.py \
--model_path Quant_VGGT/VGGT-1B/model_tracker_fixed_e20.pt \
--co3d_dir co3d_datasets/ \
--co3d_anno_dir co3d_v2_annotations/ \
--dtype quarot_w4a4\ # Quantization Bit Width
--seed 0 \
--lac \
--lwc \
--cache_path calib_data.pt \ # calibration data path
--class_mode all \ # Category selection mode for calibration data
--exp_name a44_uqant \
--resume_qs \ # Load quantized model from exp_name
```
**Quantize calibrate on Co3d, evaluate on 7 scenes or NRGBD dataset.**
```
python Quant_VGGT/vggt/evaluation/run_7andN.py\
--model_path Quant_VGGT/VGGT-1B/model_tracker_fixed_e20.pt \
--co3d_dir co3d_datasets/ \
--co3d_anno_dir co3d_v2_annotations/ \
--dtype quarot_w4a4\
--lwc \
--lac \
--exp_name quant_w4a4\
--cache_path calib_data.pt \
--class_mode all \
--output_dir "Quant_VGGT/vggt/eval_results" \
--kf 100 \ # Sample every keyframe.
--dataset nr \ # evaluation 7s or nr dataset
```
Also, you can use the quantized model for predicting other 3D attributes following the guidance [here](https://github.com/facebookresearch/vggt/tree/evaluation#detailed-usage).
## Comments
* Our codebase is heavily builds on [VGGT](https://github.com/facebookresearch/vggt) and [QuaRot](https://github.com/spcl/QuaRot). Thanks for open-sourcing!
## BibTeX
If you find *QuantVGGT* is useful and helpful to your work, please kindly cite this paper:
```
@article{feng2025quantized,
title={Quantized Visual Geometry Grounded Transformer},
author={Feng, Weilun and Qin, Haotong and Wu, Mingqiang and Yang, Chuanguang and Li, Yuqi and Li, Xiangqi and An, Zhulin and Huang, Libo and Zhang, Yulun and Magno, Michele and others},
journal={arXiv preprint arXiv:2509.21302},
year={2025}
}
```