Train, tune, and deploy LightGBM 4.6.0 gradient boosting models for regression, binary/multiclass classification, ranking, and custom objectives. Use when the user asks about LightGBM, gradient boosting on tabular data, tree-based ensemble models, or needs fast accurate predictions on structured data. Covers native API (Dataset/Booster/train/cv), scikit-learn API (LGBMClassifier/LGBMRegressor/LGBMRanker), callbacks, custom objectives/metrics, SHAP-style contributions, GPU/CUDA acceleration, and distributed training.
