SSE-PT-Deployment

Demo for deploying SSE-PT(SSE-PT: Temporal Collaborative Ranking Via Personalized Transformer) https://github.com/wuliwei9278/SSE-PT/tree/master

1. install dependencies from rquirements

2. download MovieLens 1M dataset, and put movies.dat and users.dat into the same folder of model.py

3. dowanlod the processed dataset of ml1m.txt from SSE-PT repo, and put it into the same folder of model.py

4. train the SSE-PT model by ml1m.txt and save the model by tensorflow saver (model.py may need to be changed according to the required tensor and operation, refer to the 'opcode' file)

   • repo: https://github.com/wuliwei9278/SSE-PT/tree/master
   • traning environment: Aliyun Cloud: 12Core, 92G RAM, GPU Nvidia-V100 32G, cuda 10.13.0, nvidia 450.80.02
   • refer to the revised main.py file in this repo

5. streamlit run model.py