Welcome to ELI5’s documentation!

ELI5 is a Python library which allows to visualize and debug various Machine Learning models using unified API. It has built-in support for several ML frameworks and provides a way to explain black-box models.

• Overview
  • Installation
  • Features
  • Basic Usage
  • Why?
  • Architecture
• Tutorials
  • Debugging scikit-learn text classification pipeline
  • TextExplainer: debugging black-box text classifiers
  • Explaining XGBoost predictions on the Titanic dataset
  • Named Entity Recognition using sklearn-crfsuite
  • Explaining Keras image classifier predictions with Grad-CAM
• Supported Libraries
  • scikit-learn
  • XGBoost
  • LightGBM
  • CatBoost
  • lightning
  • sklearn-crfsuite
  • Keras
• Inspecting Black-Box Estimators
  • LIME
  • Permutation Importance
• API
  • ELI5 top-level API
  • eli5.formatters
  • eli5.lightning
  • eli5.lime
  • eli5.sklearn
  • eli5.sklearn_crfsuite
  • eli5.xgboost
  • eli5.lightgbm
  • eli5.catboost
  • eli5.permutation_importance
  • eli5.keras
  • eli5.base
• Contributing
  • Making releases
• Changelog
  • 0.11.0 (2021-01-23)
  • 0.10.1 (2019-08-29)
  • 0.10.0 (2019-08-21)
  • 0.9.0 (2019-07-05)
  • 0.8.2 (2019-04-04)
  • 0.8.1 (2018-11-19)
  • 0.8 (2017-08-25)
  • 0.7 (2017-07-03)
  • 0.6.4 (2017-06-22)
  • 0.6.3 (2017-06-02)
  • 0.6.2 (2017-05-17)
  • 0.6.1 (2017-05-10)
  • 0.6 (2017-05-03)
  • 0.5 (2017-04-27)
  • 0.4.2 (2017-03-03)
  • 0.4.1 (2017-01-25)
  • 0.4 (2017-01-20)
  • 0.3.1 (2017-01-16)
  • 0.3 (2017-01-13)
  • 0.2 (2016-12-03)
  • 0.1.1 (2016-11-25)
  • 0.1 (2016-11-24)
  • 0.0.6 (2016-10-12)
  • 0.0.5 (2016-09-27)
  • 0.0.4 (2016-09-24)
  • 0.0.3 (2016-09-21)
  • 0.0.2 (2016-09-19)
  • 0.0.1 (2016-09-15)

License is MIT.