CUDA memory error while training with other larger datasets

[gustmd0121]

Hello Adam, thanks for your pytorch implementation of Group-CNN. It helped me greatly for my research. I just had an issue while using your code to train a different dataset (AID dataset). When I crop the images to 224 by 224, and use a batch size of only 4~8, I seem to be getting a CUDA memory error. I have to crop to 128 by 128 and use a batch size of 4 in order to barely not get a CUDA error. Currently I am using GeForce RTX 2080 Ti GPU with 11gb of memory. Is there any solution to this problem that you might no of? I am fairly new to this field so if this is a simple fix I am sorry for the inconvenience.