What SmartML Benchmarks Tell You Before Hyperparameter Tuning

In most ML blogs and tutorials, benchmarks look like this: Tune forever → squeeze the last 0.1% →...