ML Fundamentals
Hyperparameter
Configuration settings for training a model (learning rate, batch size, number of layers) that are set before training begins: not learned from data.
Use this term as a quick reference, then jump back into the part of the site where it matters.