Lale.jl is a Julia wrapper of Python’s Lale library for semi-automated data science. Lale makes it easy to automatically select algorithms and tune hyperparameters of pipelines that are compatible with scikit-learn, in a type-safe fashion.
Lale provides easy pipeline composition using combinators and visualization of the pipeline for ease of understanding. Performing combined algorithm selection and hyperparameter tuning (CASH) on the given pipeline is just 3 lines of code. Here is an example notebook demonstrating some basic functionality of Lale.jl.
Lale.jl has a good coverage of machine learning operators from scikit-learn and also operators from the imbalanced-learn Python package. We are working on adding more operators.
It is great! I did not know lale library in Python, but your proposal seems simpler to use :-). Unfortunately, I miss more documentation, in addition to the notebook demo.
hi, it is still in the early stage of development. yes, the online documentation is still not finished. lale in python has extensive documentation in addition to the scikitlearn operators it is using which has its own documentation. you can pass similar arguments for scikitlearn operators by consulting scikitlearn and lale python package documentation.
For example, if a classifier does not support multi-label classification and there is an attempt to use it for multi-label, this would be type-unsafe. Lale performs validation of the hyperparameters as well as input/output types before any attempt is made to train/use the model, making it type-safe.