Original implementation for paper Sparse-Interest Network for Sequential Recommendation.
Qiaoyu Tan, Jianwei Zhang, Jiangchao Yao, Ninghao Liu, Jingren Zhou, Hongxia Yang, Xia Hu
Accepted to WSDM 2021
- Python 3.6
- TensorFlow-GPU == 1.15.0rc1
- Faiss-GPU == 1.6.4
-
Install TensorFlow-GPU 1.15.0rc1
-
Install Faiss-GPU based on the instructions here: https://github.com/facebookresearch/faiss/blob/master/INSTALL.md
-
Clone this repo
git clone https://github.com/Qiaoyut/SINE.git.
-
Original links of datasets are:
-
Two preprocessed datasets (MovieLens and Taobao) are included.
You can use python main.py --dataset {dataset_name} to train SINE on a dataset. Other hyperparameters can be found in the code.
For example, you can use python main.py --dataset ml1m to train SINE model on movieLens dataset.
If you want to train models on your own dataset, you should prepare the following three files:
- train/valid/test file: Each line represents an interaction, which contains three numbers
<user_id>,<item_id>,<time_stamp>.
The structure of our code is based on MIMN.
Please cite our paper if you find this code useful for your research:
@inproceedings{tan2021sparse,
title={Sparse-interest network for sequential recommendation},
author={Tan, Qiaoyu and Zhang, Jianwei and Yao, Jiangchao and Liu, Ninghao and Zhou, Jingren and Yang, Hongxia and Hu, Xia},
booktitle={Proceedings of the 14th ACM International Conference on Web Search and Data Mining},
pages={598--606},
year={2021}
}