Pytorch implementation for "Enhancing Dyadic Relations with Homogeneous Graphs for Multimodal Recommendation" -ECAI'23 arxiv

Data could be download from DropBox: Baby/Sports/Clothing

1. Download the data from the data link we provided above, then put the download data to the ./data folder
2. Use conda env create -f dragon.yml to create the enviroment with correct dependencies
3. Run generate-u-u-matrix.py on the dataset you want to generate the user co-occurrence graph
4. Enter the src folder and run with python main.py -m DRAGON -d dataset_name

@article{zhou2023enhancing,
  title={Enhancing Dyadic Relations with Homogeneous Graphs for Multimodal Recommendation},
  author={Zhou, Hongyu and Zhou, Xin and Shen, Zhiqi},
  journal={arXiv preprint arXiv:2301.12097},
  year={2023}
}