[CVPR 2025 Best Paper Award] VGGT: Visual Geometry Grounded Transformer
[CVPR 2025 Best Paper Award] VGGT: Visual Geometry Grounded Transformer
@inproceedings{wang2025vggt,
title={VGGT: Visual Geometry Grounded Transformer},
author={Wang, Jianyuan and Chen, Minghao and Karaev, Nikita and Vedaldi, Andrea and Rupprecht, Christian and Novotny, David},
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
year={2025}
}
[May 18, 2026] The next step of VGGT — VGGT-Omega — has been released! Check it out at https://vggt-omega.github.io/.
[May 15, 2026] We fixed an implementation issue that was keeping redundant intermediate tensors in memory. With the same GPU memory budget, VGGT can now run on roughly 2-3x more input frames! See VGGT-Omega for more details.
[July 29, 2025] We've updated the license for VGGT to permit commercial use (excluding military applications). All code in this repository is now under a commercial-use-friendly license. However, only the newly released checkpoint VGGT-1B-Commercial is licensed for commercial usage — the original checkpoint remains non-commercial. Full license details are available here. Access to the checkpoint requires completing an application form, which is processed by a system similar to LLaMA's approval workflow, automatically. The new checkpoint delivers similar performance to the original model. Please submit an issue if you notice a significant performance discrepancy.
[July 6, 2025] Training code is now available in the training folder, including an example to finetune VGGT on a custom dataset.
[June 13, 2025] Honored to receive the Best Paper Award at CVPR 2025! Apologies if I’m slow to respond to queries or GitHub issues these days. If you’re interested, our oral presentation is available here. Another long presentation can be found here (Note: it’s shared in .pptx format with animations — quite large, but feel free to use it as a template if helpful.)
[June 2, 2025] Added a script to run VGGT and save predictions in COLMAP format, with bundle adjustment support optional. The saved COLMAP files can be directly used with gsplat or other NeRF/Gaussian splatting libraries.
[May 3, 2025] Evaluation code for reproducing our camera pose estimation results on Co3D is now available in the evaluation branch.
Visual Geometry Grounded Transformer (VGGT, CVPR 2025) is a feed-forward neural network that directly infers all key 3D attributes of a scene, including extrinsic and intrinsic camera parameters, point maps, depth maps, and 3D point tracks, from one, a few, or hundreds of its views, within seconds.
First, clone this repository to your local machine, and install the dependencies (torch, torchvision, numpy, Pillow, and huggingface_hub).
git clone [email protected]:facebookresearch/vggt.git
cd vggt
pip install -r requirements.txt
Alternatively, you can install VGGT as a package (click here for details).
Now, try the model with just a few lines of code:
…
The model weights will be automatically downloaded from Hugging Face. If you encounter issues such as slow loading, you can manually download them here and load, or:
model = VGGT()
_URL = "https://huggingface.co/facebook/VGGT-1B/resolve/main/model.pt"
model.load_state_dict(torch.hub.load_state_dict_from_url(_URL))
You can also optionally choose which attributes (branches) to predict, as shown below. This achieves the same result as the example above. This example uses a batch size of 1 (processing a single scene), but it naturally works for multiple scenes.
…
Furthermore, if certain pixels in the input frames are unwanted (e.g., reflective surfaces, sky, or water), you can simply mask them by setting the corresponding pixel values to 0 or 1. Precise segmentation masks aren't necessary - simple bounding box masks work effectively (check this issue for an example).
We provide multiple ways to visualize your 3D reconstructions. Before using these visualization tools, install the required dependencies:
pip install -r requirements_demo.txt
Please note: VGGT typically reconstructs a scene in less than 1 second. However, visualizing 3D points may take tens of seconds due to third-party rendering, independent of VGGT's processing time. The visualization is slow especially when the number of images is large.
Our Gradio-based interface allows you to upload images/videos, run reconstruction, and interactively explore the 3D scene in your browser. You can launch this in your local machine or try it on Hugging Face.
python demo_gradio.py
Run the following command to run reconstruction and visualize the point clouds in viser. Note this script requires a path to a folder containing images. It assumes only image files under the folder. You can set --use_point_map to use the point cloud from the point map branch, instead of the depth-based point cloud.
python demo_viser.py --image_folder path/to/your/images/folder
We also support exporting VGGT's predictions directly to COLMAP format, by:
…
Please ensure that the images are stored in /YOUR/SCENE_DIR/images/. This folder should contain only the images. Check the examples folder for the desired data structure.
The reconstruction result (camera parameters and 3D points) will be automatically saved under /YOUR/SCENE_DIR/sparse/ in the COLMAP format, such as:
SCENE_DIR/
├── images/
└── sparse/
├── cameras.bin
├── images.bin
└── points3D.bin
The exported COLMAP files can be directly used with gsplat for Gaussian Splatting training. Install gsplat following their official instructions (we recommend gsplat==1.3.0):
An example command to train the model is:
cd gsplat
python examples/simple_trainer.py default --data_factor 1 --data_dir /YOUR/SCENE_DIR/ --result_dir /YOUR/RESULT_DIR/
Our model shows surprisingly good performance on single-view reconstruction, although it was never trained for this task. The model does not need to duplicate the single-view image to a pair, instead, it can directly infer the 3D structure from the tokens of the single view image. Feel free to try it with our demos above, which naturally works for single-view reconstruction.
We did not quantitatively test monocular depth estimation performance ourselves, but @kabouzeid generously provided a comparison of VGGT to recent methods here. VGGT shows competitive or better results compared to state-of-the-art monocular approaches such as DepthAnything v2 or MoGe, despite never being explicitly trained for single-view tasks.
Our work builds upon a series of previous research projects. If you're interested in understanding how our research evolved, check out our previous works:
Deep SfM Revisited ──┐ PoseDiffusion ─────► VGGSfM ──► VGGT CoTracker ──┘Thanks to these great repositories: PoseDiffusion, VGGSfM, CoTracker, DINOv2, Dust3r, Moge, PyTorch3D, Sky Segmentation, Depth Anything V2, Metric3D and many other inspiring works in the community.
See the LICENSE file for details about the license under which this code is made available.
Please note that only this model checkpoint allows commercial usage. This new checkpoint achieves the same performance level (might be slightly better) as the original one, e.g., AUC@30: 90.37 vs. 89.98 on the Co3D dataset.