Cornac

New improvements

- Improve the efficiency of KNN methods (331)
- Use `tqdm.auto` for compatibility on Jupyter notebook env (332)

New models

- MTER model is faster with Cython (320)
- Neighborhood-based methods (UserKNN and ItemKNN) (329)

New features and improvements

- Refactor code in examples (317)
- CI tools support Python 3.7
- Update and test all tutorials and examples

New models and datasets

- Weighted Bayesian Personalized Ranking (WBPR) model (309)
- Maximum Margin Matrix Factorization (MMMF) model (310)

New features and improvements

- Reset random number generator for reproducibility (301)
- Fix issue in NCRR metric (313)
- Use C++ Boost Random library for reproducibility across platforms (315)
- Support model saving and loading (316)

New features and improvements

- Add default attributes `total_users` and `total_items` to `Dataset` (300)

New models and datasets

- MovieLens 10M and 20M datasets (291)

New features and improvements

- Standardize datasets `load_feedback()` API (278)
- Add hyperopt for hyper-parameter tuning (286)
- Show validation results (optional) if exists (289)
- Update `Recommender.rank()` to support unknown item scores (283)
- Support multiple values of K for ranking metrics (297)
- Tutorial on how to work with auxiliary data (264)
- Tutorial on hyperparameter search for VAECF (290)
- Examples for VAECF, VMF, and SoRec (272, 276, 287)