Variational AutoEncoder For Collaborative Filtering

An Implementation of Variational Autoencoders for Collaborative Filtering (Liang et al. 2018) in PyTorch.

Dependencies

• python 3.8

pip install -r requirements.txt

How to run

python main.py

Results

Test Result

Diff	Model(unofficial)	Official Code
Epoch	50	200
Loss	472.55275	-
NDCG@100 (mean/sd)	0.42109 / 0.00209	0.42592 / 0.00211
Recall@20 (mean/sd)	0.39145 / 0.00269	0.39535 / 0.00270
Recall@50 (mean/sd)	0.53414 / 0.00285	0.53540 / 0.00284

Reference

Official Code Implementation