Batch size

The number of training examples processed together before a single weight update. A larger batch produces a more stable gradient estimate at the cost of more memory per step, a smaller batch updates more often on noisier estimates.

Why exams ask this

Tested as a memory-versus-stability tradeoff. A larger batch size gives a smoother, more accurate gradient estimate but costs more memory per step, and the exam wants that tradeoff named, not "bigger is always better."

Relevant to

Related concepts

Resources

No resources linked to this concept yet.