Authors: Hassan Hammoud, Andrew Yu, Venugopal Chillal
baseline.ipynb
- has baseline resultsALS.ipynb
- has ALS code and resultsSAR.ipynb
- has SAR code and resultslightgcn/lightgcn.ipynb
- has LightGCN code, results are listed in different results CSVsratings_Electronics.csv.bz2
- has the original data zipped via bz2
If necessary, spin up a new virtual env and seed it with our requirements.txt
.
You should be able to run the baseline notebook directly, given that your virtual environment has all necessary packages.
You should be able to run the SAR notebook directly, given that your virtual environment has all necessary packages.
To run ALS properly, you should set it up in Sagemaker or another Pyspark compatible environment.
You should be able to run the SAR notebook directly, given that your virtual environment has all necessary packages.
Note that the notebook will create three folders: yamls
to store .yaml files of hyperparameter combinations, embeddings
to store embeddings from models run, and models
, which stores models.