/serenRec

Baselines for sequential recommendation experiments

Primary LanguagePythonApache License 2.0Apache-2.0

Overview

serenRec is a python toolkit developed for sequential-/session-based recommendation baselines and experiments.

How to Run

python main.py -use_cuda -gpu_id=0 -model=sasrec

Implemented Algorithms

Model Publication
Session-Pop A re-visit of the popularity baseline in recommender systems
Session-MF Matrix factorization techniques for recommender systems
Caser Personalized Top-N Sequential Recommendation via Convolutional Sequence Embedding
SASRec Self-Attentive Sequential Recommendation.
SRGNN Session-based Recommendation with Graph Neural Networks
STAMP STAMP: Short-Term Attention/Memory Priority Model for Session-based Recommendation
GRU4Rec Improved Recurrent Neural Networks for Session-based Recommendations
FMLP Filter-enhanced MLP is All You Need for Sequential Recommendation

TODO List

  • More baselines
  • interface for more datasets