Transformers4Rec: Building Session-Based Recommendations with an NVIDIA Merlin Library

Originally published at:

Transformers4Rec makes it easy to use SOTA NLP architectures for sequential and session-based recommendation by leveraging HuggingFace Transformers.

We have been excited on the effectiveness of Transformers even for short sessions compared to other state-of-the-art session-based recommendation models, as we observed in our research experiments and also for the competitions we won using the library.
Looking forward to hear from you if you have already tried or plan to try Transformers4rec to your own use case.