I was able to run the code on a machine with huge memory. My question is why during training you calculate the accuracy (besides the loss). This is because my model does not use the accuracy metric. Would my result still be correct if I calculated only the loss.
I was able to run the code on a machine with huge memory. My question is why during training you calculate the accuracy (besides the loss). This is because my model does not use the accuracy metric. Would my result still be correct if I calculated only the loss.