Sorry for the mistake in Question,
Is the combination of high batch size and low epoch the same as that of low batch size and high epoch?
Hey! I was just wondering …
If I try to generate High-resolution(128,128,3) images from Low resolution(32,32,3) images using GAN on a dataset of 1000 Images by following two methods :
Result_A: I run 100,000 epoch and batch_size =1 then my model will train on a total of 100,000 images. And this one image per epoch is selected randomly.
Result_B: I run 100 epochs on batch_size = 1000, then also my model trains on a total of 100,000 images.
Would Result_A be similar to Result_B? If not, why?
The differentiating factor is that in Result_B my model goes through all images, but it might not go through all images. Will this create a significant difference?