Batch Size
The number of training examples a model looks at before it updates its weights once. Bigger batches give steadier updates but need more memory; smaller batches update more often and add helpful noise.
Think of It Like This
It's how many flashcards you review before pausing to update your notes.
Rather than update after every single example (slow and jittery) or after the whole dataset (memory-hungry and sluggish), models process a batch at a time and update once per batch. Larger batches smooth out each step but demand more memory; smaller batches update more often, and the extra noise in their estimates can actually help the model generalize.