Integration of classifier/regressors from Create ML?

[lagerard-gefip]

Is there any plan to bring to MLX algorithms from Create ML? (e.g. random forests, boosted trees, etc.)

Although it is not gradient-based, if it could leverage automatic parallelization + compilation of functions, it would give another edge to MLX, specifically against scikit-learn.

[awni]

There is no plan for that. Something like scikit-learn can be built on top of MLX to run faster on macOS. In fact there is some discussion around how to make that work since MLX mostly follows the array API.