Pytorch中几种新闻推荐方法的实现。.zip

# News Recommendation Implementation of several news recommendation methods in Pytorch. Support multi-GPU training and testing. > **WARNING**: Multi-GPU training and testing only work on Linux because currently `torch.distributed` only support Linux. ### Requirements - python - pytorch - numpy - scikit-learn - nltk - tqdm You can create a `conda` virtual environment named `nr` with the provided `env.yaml`. ```bash conda env create -f env.yaml # Maybe only work on Linux ``` ### Usage - **Clone this repository** ```bash git clone https://github.com/yflyl613/NewsRecommendation.git cd NewsRecommendation ``` - **Prepare data** A scirpt `download_data.sh` is provided to download and unzip all the required data. It will create a new folder `data/` under `NewsRecommendation/`. ```bash # In NewsRecommendation/ chmod +x download_data.sh ./download_data.sh ``` - **Start training** A script `demo.sh` is provied for model training and testing, in which you can modify parameters for the experiment. Please refer to `parameters.py` for more details. ```bash # In NewsRecommendation/data/ cd ../src chmod +x demo.sh # train ./demo.sh train # test ./demo.sh test <checkpoint name> # E.g. ./demo.sh test epoch-1.pt ``` ### Results on MIND-small validation set^[1] - **NAML^[2]** | News information | AUC | MRR | nDCG@5 | nDCG@10 | Configuration | | :------------------------------: | :---: | :---: | :----: | :-----: | :---------------------------------------------: | | subcategory<br>category<br>title | 66.24 | 32.08 | 35.36 | 41.56 | batch size 128 (32*4)<br> 5 epochs<br>lr 3e-4 | - **NRMS^[3]** | News information | AUC | MRR | nDCG@5 | nDCG@10 | Configuration | | :--------------: | :---: | :---: | :----: | :-----: | :-------------------------------------------: | | title | 66.61 | 31.86 | 35.19 | 41.46 | batch size 128 (32*4)<br> 4 epochs<br>lr 3e-4 | > **Please feel free to contact me by opening an [issue](https://github.com/yflyl613/NewsRecommendation/issues) if you have any problem or find a bug :)** ### Reference [1] Fangzhao Wu, Ying Qiao, Jiun-Hung Chen, Chuhan Wu, Tao Qi, Jianxun Lian, Danyang Liu, Xing Xie, Jianfeng Gao, Winnie Wu and Ming Zhou. [MIND: A Large-scale Dataset for News Recommendation](https://msnews.github.io/assets/doc/ACL2020_MIND.pdf). ACL 2020. [2] Chuhan Wu, Fangzhao Wu, Mingxiao An, Jianqiang Huang, Yongfeng Huang, and Xing Xie. [Neural News Recommendation with Attentive Multi-View Learning](https://www.ijcai.org/Proceedings/2019/0536.pdf). IJCAI. 2019. [3] Chuhan Wu, Fangzhao Wu, Suyu Ge, Tao Qi, Yongfeng Huang, and Xing Xie. [Neural News Recommendation with Multi-Head Self-Attention](https://www.aclweb.org/anthology/D19-1671.pdf). EMNLP-IJCNLP. 2019.