Endgame Documentation

Endgame is a comprehensive machine learning toolkit providing 300+ estimators, transformers, and visualizers across tabular, time series, signal processing, CV, NLP, audio, and multimodal domains. It unifies state-of-the-art and classical methods under a consistent scikit-learn-compatible API.

import endgame as eg

# Quick model comparison
result = eg.quick.compare(X, y, task='classification')

# Full pipeline
model = eg.models.LGBMWrapper(preset='endgame')
model.fit(X_train, y_train)
predictions = model.predict(X_test)

Key Features

  • 100+ models with sklearn-compatible API

  • Polars-powered preprocessing for speed

  • Competition-winning defaults via preset system

  • Conformal prediction and probability calibration

  • Comprehensive signal processing (45 transforms)

  • AutoML framework matching AutoGluon’s simplicity

  • 42 interactive visualizations for model interpretation

Getting Started

User Guides

API Reference

Indices and tables