Learning Rate

ML

A hyperparameter controlling the step size during model training as parameters are adjusted. A learning rate that is too high may cause instability; too low causes slow convergence. Finding the right learning rate is crucial for training success.