Batch Size
In plain terms
How many training examples a model processes before updating.
Going deeper
Batch size sets how many samples are used to compute each gradient update, balancing speed, memory, and stability. It affects how training proceeds.
Under the hood
Large batches are faster/stabler but need more memory and can generalize worse; small batches add useful noise.
Examples
Mini-batch training, GPU memory limits, gradient descent, tuning.
Think of it like...
Like grading a stack of papers before adjusting your teaching, versus one at a time.