Multiprocesssing for deep learning cross-validation

Question

I am working on a research project which involves cross-validation for deep learning models. The data is pretty small (under 1000 matrices of size 200 x 200), as are models