MPSBoost Documentation

MPSBoost is a tree-model learning library for Apple Silicon. The current public release is 1.0.0, with native CPU/MPS backends, sklearn-style estimators, native CPU softmax multiclass classification, forests, advanced regression objectives, explanations, anomaly detection, ranking entry points, optional SHAP/portable-backend diagnostics, and real-world acceptance reporting.

Start Here

• Installation and environment
• Quick start
• Backend policy
• Estimator API
• Model families
• Release overview
• 1.0.0 release

The Chinese source documentation remains available through the language selector.