I have a series of images which shows the evolution of a system in time and my aim is to predict the next n-steps forward in time. My issue is how to return a set of say 10 images to predict the next 10 and then continue. With a Pytorch dataloader using say a batch of 30, once I get to image 20 I need to generate the next set of images from the dataloader to perform prediction check.

Plus I need to iterate over the set of images eg 0-10, 1-11, 2-12 etc

This is for an assignment and I am only looking any ideas/ links to similar implementations as I want to write the code myself.

Thanks in advance