Machine learning optimization technique that uses. Used to find the optimal weighting for Hyperparameters.
- A high-learning rate can cause model to converge too quickly.
- Low-learning rate means model can behave in too closely to the prompt, perhaps missing a key solution early.