RecBole has included more than 100 widely used recommendation algorithms.
In the latest version, we implement RecBole recommendation models covering general recommendation, sequential recommendation, context-aware recommendation and knowledge-based recommendation. In addition, eight subpackage toolkits in version 2.0 based on RecBole framework have implemented 65 recommendation system models.
We summarize all the models in the following table:
NOTE: For some recommendation algorithms, we cannot find the original implementation, and we implemented them according to our understanding of the original paper. Besides, parameter tuning or learning algorithm may not be the optimal way as originally expected. If this case was found, please contact us to help improve the algorithm implementation.
We constructed preliminary experiments to test the time and memory cost on three different-sized datasets (small, medium and large). For detailed information, you can click the following links.
NOTE: Our test results only gave the approximate time and memory cost of our implementations in the RecBole library (based on our machine server). Any feedback or suggestions about the implementations and test are welcome. We will keep improving our implementations, and update these test results.