TPOT

TPOT is a Python AutoML library for data scientists and researchers who want to automate machine learning pipeline selection on their own hardware. It uses genetic programming to search candidate pipelines, reducing the manual work of comparing models and tuning their parameters. It's free and open source under LGPL-3.0.

The search covers pipeline design as well as model settings. Genetic feature selection lets it explore which input features to retain, while customizable search spaces let you control which components and parameter ranges it considers. Multi-objective optimization supports evaluating candidates against more than one goal, including custom objectives.

Its graph-based framework gives researchers room to customize the evolutionary algorithm itself. TPOT uses scikit-learn and includes XGBoost and LightGBM in its Python stack. Dask handles parallel processing, and the library works in Python scripts and Jupyter notebooks. The documentation also addresses Apple Silicon Macs and other Arm-based CPUs.

TPOT can preprocess data by filling missing values, encoding categorical features and standardizing inputs. You'll still need to choose components suited to your data: it doesn't check that every candidate operator can handle the input format. Its optional preprocessing fits the entire training set before cross-validation splits, rather than fitting separately within each fold.