Resources for learning current best-practices for structuring PyTorch code?

Does anyone have any recommendations for a recent style/best practices guide for writing pytorch code?

From my understanding, many deep learning projects end up sharing a lot of similar structure (data loading, model instantiation, train/test loops, analyzing training / loss dynamics). Is there anywhere I can learn the most common practices for structuring these methods? I’d like to prepare myself for collaborating with other ML/DL engineers in an industry context.


1 Like