Views
No views yet
Fast3R class from fast3r.models.fast3r (follow instructions from the Fast3R GitHub repo to install) and use it as a regular PyTorch model.1from fast3r.models.fast3r import Fast3R
2from fast3r.models.multiview_dust3r_module import MultiViewDUSt3RLitModule
3
4# Load the model from Hugging Face
5model = Fast3R.from_pretrained("jedyang97/Fast3R_ViT_Large_512")
6model = model.to("cuda")
7
8# [Optional] Create a lightweight lightning module wrapper for the model.
9# This provides functions to estimate camera poses, evaluate 3D reconstruction, etc.
10# See fast3r/viz/demo.py for an example.
11lit_module = MultiViewDUSt3RLitModule.load_for_inference(model)
12
13# Set model to evaluation mode
14model.eval()
15lit_module.eval()@InProceedings{Yang_2025_Fast3R,
title={Fast3R: Towards 3D Reconstruction of 1000+ Images in One Forward Pass},
author={Jianing Yang and Alexander Sax and Kevin J. Liang and Mikael Henaff and Hao Tang and Ang Cao and Joyce Chai and Franziska Meier and Matt Feiszli},
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
month={June},
year={2025},
}