Training & Learning
Batch Size
How many training examples the model looks at before it updates its weights once.
Batch size sets how many examples are processed together before the model makes one update. Larger batches give smoother, more stable updates and use hardware more efficiently but need more memory, while smaller batches add useful noise and fit on smaller machines. It interacts closely with the learning rate, so the two are usually tuned together.
Related terms
Learning Rate
The size of the steps a model takes while learning, and one of the most important training settings to get right.
Gradient Descent
The optimization method that gradually tweaks a model to lower its error, like walking downhill to the lowest point.
Hyperparameter
A setting you choose before training that controls how the model learns, as opposed to what it learns.
Epoch
One complete pass of the training process through the entire training dataset.