machinelearningnuremberg/HPO-B

This is a benchmark for machine learning researchers and practitioners to systematically compare and evaluate the effectiveness of different hyperparameter optimization (HPO) algorithms. It provides a standardized collection of past machine learning model evaluations, allowing you to test how well a new HPO method selects optimal model settings. You feed in your new HPO algorithm, and it returns a list of maximum accuracies achieved over a series of trials, showing how your algorithm performs against established benchmarks.