i currently having confusion on the 'MiniBatchSize' function offered under trainingOptions in deep learning. I put a scenario below for better understanding on the questions.
Dataset: 4500 Sample ( 9 categories with 500 sample each)
MiniBatchSize : 10
- Does it mean that i would have 10 samples in every batch or 500 samples in every batch (5000 sample/10 batch)?
- Does having more samples in one batch size increase the accuracy of the trained network (CNN) or vise versa ?
Wish someone could help me clarify on the confusion. Thank You very much.