GitHub

Enhancing Disentanglement of Popularity Bias for Recommendation with Contrastive Learning

We provide three datasets: Ciao, LastFM, and Movielens-10M

Environment: I have tested this code with python3.8 Pytorch=1.7.1 CUDA=11.0
Run DECL
- first start the visdom server: visdom -port 33336
- then run model on each dataset by running: python app.py --flagfile ./config/ciao_decl.cfg

Name		Name	Last commit message	Last commit date
Latest commit History 6 Commits
data		data
src		src
README.md		README.md
metrics.py		metrics.py
model.py		model.py
recommender.py		recommender.py
tester.py		tester.py
trainer.py		trainer.py
utils.py		utils.py