Training a UNet model with Pytorch

I am training a UNet model on a dataset of 100 images. The image size is 1024 x1024.
I am using NVIDIA Tesla k80, 56 RAM. My batch size is two images due to GPU memory restrictions.
The training process takes a lot of time, I am training four days so far, and it still doesn’t finish.
I am not sure if it normal process due to the large image size, or maybe I do something wrong. If anyone encounters with training images 1024x1024 and how much time does it take?
Thanks in advance :slight_smile: