An un-official PyTorch implementation of Video2Shop: Exact Matching Clothes in Videos to Online Shopping Images
Conda environment is suggested.
- Python 3.6 or +
- PyTorch 1.6 with Cuda
- torchvision
- tensorboardX
- easydict
- tqdm
To train with one GPU
python train.py -tph <training_phase: 0 first part - 1 fusion nodes - 2 both> -e <num_epochs> -lr <learning-rate> --dataset <dataset> --gpus 0
python evaluation.py -ckpt <ckpt>
If you use this code in your research, please use the following BibTeX entry.
@inproceedings{cheng2017video2shop,
title={Video2shop: Exact matching clothes in videos to online shopping images},
author={Cheng, Zhi-Qi and Wu, Xiao and Liu, Yang and Hua, Xian-Sheng},
booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition},
pages={4048--4056},
year={2017}
}