implicit v0.5.0 Release Notes
-
💥 Breaking API Changes
The API for implicit has substantially changed in v0.5.0 - and any code written for the previous API will need to be rewritten:
- 🔄 Change model.fit to take a user_items sparse matrix #484
- Return numpy arrays from recommend methods #482
- Don't require empty rows in user_items and item_users parameters #526
- Unify API for rank_items/recommend/recommend_all #489
🐎 Performance Improvements
- Speedup evaluation by using batch recommend functions #520
- 👉 Use FAISS for GPU inference #506
- Multithreaded speedups for CPU models #517
- 👉 Use thrust::binary_search to verify negative samples on GPU #524
- 🚀 Release GIL on GPU calls #528
🆕 New Features
- ➕ Add incremental retraining support for ALS models #527
- Add filtering for similar_items and similar_users #488
- ➕ Add support for recalculate_users/items on the GPU #515
- ➕ Add methods for converting MF models to/from gpu #521
- ➕ Add a tutorial notebook for the lastfm example #529
- Approximate nearest neighbour for BPR/LMF and GPU models #487
- Dynamically detect CUDA availability #174