This repository contains code for the CVNet framework as described in our CVPR 2022 paper.
- Python 3 (we used v3.6.5)
- PyTorch (we used v1.2.0)
- PIL
- scipy and associated packages
- tqdm
- ChamferDistance
- Download datasets from here
- Unzip datasets into your desired directory
datasets
- Modify these directories in
package/config/config.ini
, and the directory where you intend to keep results - Complete train/val/test split by
split_datasets.py
, our divided txt files are placed inpackage/datasets
- Run experiments with
runner.py
- Download models from here
- Darnet (https://github.com/dcheng-utoronto/darnet)
- ChamferDistance (https://github.com/ThibaultGROUEIX/ChamferDistancePytorch)
Please contact me at xuziqiang@njust.edu.cn for questions.