Fundamentals
Hyperparameter
A hyperparameter is a setting that controls model structure or training.
How it works
The training algorithm does not directly learn hyperparameters as model weights. Teams choose them through reasoning, sweeps, or optimization studies.
A concrete example
Illustrative: compare learning rates 0.001, 0.0003, and 0.0001 over three runs.
Do not confuse it with Parameter
a hyperparameter configures learning; a parameter is produced by learning.