This is a very interesting paper which presents ‘SHERPA’, a Python library for hyperparameter tuning of machine learning models. With Sherpa, researchers can easily optimize hyperparameters using a variety of powerful and interchangeable algorithms.

It provides:

  • Hyperparameter optimization for machine learning researchers.
  • It can be used with any Python machine learning library such as Keras, Tensorflow, PyTorch, or Scikit-Learn.
  • A choice of hyperparameter optimization algorithms such as Bayesian optimization via GPyOpt (example notebook), Asynchronous Successive Halving (aka Hyperband) (example notebook), and Population Based Training (example notebook).
  • Parallel computation that can be fitted to the user’s needs.
  • A live dashboard for the exploratory analysis of results.

Download Paper: https://arxiv.org/pdf/2005.04048.pdf

Github: https://github.com/sherpa-ai/sherpa