How to train on dataset of high resolution images?

There is the same question but with more answers here: How to train on dataset of large images