Fundamentals
Parameter
A parameter is a learned numeric value inside a model.
How it works
Optimization changes parameters to reduce loss. Weights and biases are parameters because training estimates them from data.
A concrete example
Illustrative: a 3-input linear layer with 2 outputs has 3×2 weights plus 2 biases, or 8 parameters.
Do not confuse it with Hyperparameter
parameters are learned; hyperparameters are chosen before or around training.