Hi all,
I was wondering why there is a difference in batch sizes (train=128, validation=32) in the flow_from_directory step.
Thanks
The main reason I could think of is because validation dataset is smaller than training dataset. So in order to loop at the same speed needs to smaller.
1 Like