Skip to content

Welcome to Lazy Predict

Lazy Predict helps you build and benchmark a large number of machine learning models with minimal code, so you can quickly discover which algorithms work best on your dataset — without any parameter tuning.

PyPI version Downloads CodeFactor

Quick Start

from lazypredict.Supervised import LazyClassifier
from sklearn.datasets import load_breast_cancer
from sklearn.model_selection import train_test_split

data = load_breast_cancer()
X_train, X_test, y_train, y_test = train_test_split(
    data.data, data.target, test_size=0.2, random_state=42
)

clf = LazyClassifier(verbose=0, ignore_warnings=True)
models, predictions = clf.fit(X_train, X_test, y_train, y_test)
print(models)

from lazypredict.Supervised import LazyRegressor
from sklearn.datasets import load_diabetes
from sklearn.model_selection import train_test_split

data = load_diabetes()
X_train, X_test, y_train, y_test = train_test_split(
    data.data, data.target, test_size=0.2, random_state=42
)

reg = LazyRegressor(verbose=0, ignore_warnings=True)
models, predictions = reg.fit(X_train, X_test, y_train, y_test)
print(models)

import numpy as np
from lazypredict.TimeSeriesForecasting import LazyForecaster

t = np.arange(200)
y = 10 + 0.05 * t + 3 * np.sin(2 * np.pi * t / 12) + np.random.normal(0, 1, 200)

fcst = LazyForecaster(verbose=0, ignore_warnings=True)
scores, predictions = fcst.fit(y[:180], y[180:])
print(scores)

What Can It Do?

Task Class Models
Classification LazyClassifier 40+ scikit-learn classifiers + XGBoost, LightGBM, CatBoost, cuML
Regression LazyRegressor 40+ scikit-learn regressors + XGBoost, LightGBM, CatBoost, cuML
Time Series LazyForecaster 20+ statistical, ML, deep learning, and foundation models
Distributed LazySparkClassifier / LazySparkRegressor PySpark MLlib models

Feature Highlights

  • Zero-boilerplate benchmarking — one call trains and evaluates all models
  • 40+ classifiers and regressors from scikit-learn plus optional boosting libraries
  • 20+ time series forecasting models — Naive, ETS, ARIMA, Theta, ML, LSTM/GRU, TimesFM
  • GPU acceleration — XGBoost, LightGBM, CatBoost, cuML (RAPIDS), LSTM/GRU, TimesFM
  • Automatic seasonal period detection via autocorrelation analysis
  • Exogenous variables support in time series forecasting (SARIMAX, AutoARIMA, ML models)
  • Multiple categorical encoders — OneHot, Ordinal, Target, Binary
  • MLflow integration — automatic experiment tracking via environment variable
  • Cross-validation — both train/test split and k-fold CV supported
  • Per-model timeout — skip slow models automatically
  • Hyperparameter tuning — Optuna, scikit-learn, or FLAML backends
  • Explainability — permutation importance and optional SHAP values
  • Ensemble methods — simple average, weighted average, stacking
  • Visualization — forecast plots, model comparison, residuals, error heatmaps
  • Intel Extension acceleration — opt-in CPU speedup via sklearnex.patch_sklearn()
  • Python 3.10 – 3.14 support
  • MIT licensed, free software

Getting Started

  • Installation

    Install from PyPI, conda-forge, or source. Learn about optional extras.

  • Usage Guide

    Core usage patterns for classification, regression, time series, and the CLI.

  • Time Series Forecasting

    Deep dive into LazyForecaster — all 20+ models, cross-validation, ensembles, and plots.

  • Examples

    Detailed worked examples for every feature with copy-paste code.

  • Advanced Features

    Explainability, hyperparameter tuning, visualization, distributed computing, and Spark.

  • API Reference

    Complete auto-generated API docs for every public class and function.