Models
This section documents high-level models that wrap one or more learners to expose a complete API for training, evaluation, and gradient boosting operations in reinforcement learning and supervised settings.
Each model inherits from a common base interface (BaseGBT) and typically contains logic for managing boosting steps, SHAP explanations, device movement, and gradient tracking.