Learning rate is a setting that controls how much an artificial intelligence or machine learning model changes its internal parameters during training. It determines the size of each adjustment made when the model learns from errors. A high learning rate allows larger, faster updates but may cause the model to miss the best solution. A low learning rate makes smaller, more careful updates, but training can take longer to reach a useful result.
For example, when training an image-recognition model, an appropriate learning rate helps it gradually improve its predictions without making unstable changes. Choosing a suitable value can improve training speed, accuracy, and overall model performance.
