v1.0.0

💥 Breaking Changes

• The parameter type of hidden_units and device has been changed (#205).
• Data construction in model retrain use new functions (#212).

🐛 Bug Fixes

• Fix incompatible batch_normalization and initializer layer from TensorFlow2 (#206).

📚 Documentation

• Add formal documentation (#208).
• Add docstrings (#207).

🧻 Miscellaneous

• Use pyproject.toml for building project according to PEP 621, which is also officially recommended in PyPA, pip and Setuptools (#209).
• Since currently Python3.6 can't be fully built from pyproject.toml in Setuptools, we read related metadata from pyproject.toml into setup.py (#210).

Full Changelog: v0.12.6...v1.0.0

v0.12.6

🐛 Bug Fixes

• Fix sequence data sampling bug (#195)

📚 Documentation

• Add some caveats on duplicate samples in data, see User Guide (#196)

Full Changelog: v0.12.4...v0.12.6

v0.12.4

🐛 Bug Fixes

• Fix dense feature bugs in DIN & GraphSage & PinSage (#184)

🧻 Miscellaneous

• Replace pytest-cov with coveragepy in CI (#185)
• Add Sanic serving test (#187)
• Lint with ruff and isort (#188)
• Limit num of requests with semaphore in tf model serving (#190)

v0.12.2

💥 Breaking Changes

• The recommendation result will return a dict, with users as keys and numpy.array of items as values.

🚀 Features

• Add recommend strategy (#170). See User Guide.
• Add batch recommend (#171).

🐛 Bug Fixes

• Fix UserCF & ItemCF inner_id bugs (#172)

📚 Documentation

• Add Implementation Details (#174)
• Update User Guide (#174)

v0.12.0

🚀 Features

• Support train & save & retrain PyTorch models (#141)
• Add GraphSage, GraphSageDGL (#161)
• Add PinSage (#146), PinSageDGL (#150)
• Add serving models with Rust (#157)
• Add serving using Docker Compose (#156)

🐛 Bug Fixes

• Fix retraining item2vec and deepwalk models (#163)

📚 Documentation

• Add Rust Serving Guide (#157)
• Refactor docs (#155)

v0.10.2

Bug Fixes

• Fix mistakenly preserving feature state after running (#132)

API Changes

• remove reset_state parameter in DatasetFeat.build_trainset and DatasetPure.build_trainset (#132)

v0.10.0

• Add focal loss.
• Add embedding APIs.
• Refactor the serving part. The original serving module is renamed to libserving.

v0.8.4

Add graph algorithms: DeepWalk, NGCF, LightGCN

v0.8.2

Merge pull request #108 from massquantity/dev

update README

v0.8.0

Merge pull request #92 from massquantity/dev

v0.8.0