This repository contains the official PyTorch implementation of our paper:
Our codebase is built upon the BEVDet (v1) and StreamPETR codebases. Please refer to their original repos for instractions about setting up datsets and environments.
@article{yang2024widthformer,
title={WidthFormer: Toward Efficient Transformer-based BEV View Transformation},
author={Yang, Chenhongyi and Lin, Tianwei and Huang, Lichao and Crowley, Elliot J},
journal={arXiv preprint arXiv:2401.03836},
year={2024}
}