Back to Gap Map Infrastructure / Classic ML & computer vision

Dask-ML

Dask
open source / Overall score: 2.4

Dask-ML scales machine learning in Python across cores and clusters with Dask. It provides parallel versions of common scikit-learn estimators, preprocessing and model selection, including incremental and Hyperband hyperparameter search, and wrappers that let scikit-learn, XGBoost and LightGBM models train or predict on Dask arrays and dataframes. It is maintained within the Dask project.

Openness

5 high confidence
5.0
license
BSD-3-Clause (OSI)
source
public (github.com/dask/dask-ml)
core features withheld
no — no paid edition; the repository has no enterprise directory

BSD-3-Clause and built from the public repository. Dask-ML is a community project of the Dask ecosystem with no paid edition, so nothing in it is held back.

Adoption

2 high confidence
2.0

Measured on monthly PyPI downloads of the dask-ml package.

Capability

3 medium confidence
3.0

Dask-ML spreads scikit-learn-style training, preprocessing and hyperparameter search across a cluster or data larger than memory. It leaves the choice of model to the user, as scikit-learn does, and implements fewer learners of its own.

  • https://ml.dask.org/ recorded 2026-11-12

    Docs landing: "Dask-ML provides scalable machine learning in Python using Dask alongside popular machine learning libraries like Scikit-Learn, XGBoost, and others"; integrations for Scikit-Learn & Joblib, XGBoost & LightGBM, PyTorch, Keras and TensorFlow.

  • https://ml.dask.org/modules/api.html recorded 2026-11-12

    API reference: dask_ml.model_selection (GridSearchCV, RandomizedSearchCV, IncrementalSearchCV, HyperbandSearchCV), ensemble, linear_model (LinearRegression, LogisticRegression, PoissonRegression), naive_bayes, wrappers (ParallelPostFit, Incremental), cluster (KMeans, SpectralClustering), decomposition, preprocessing, impute, metrics and xgboost.

Verified 2026-11-12