I am using Pytorch\'s custom dataset feature to create a custom dataset from 128 separate files in one folder. My issue is that the resources I\'ve come across cater to files in