Dataloader in PyTorch
xdtl opened this issue · 2 comments
xdtl commented
This may affect data augmentation: https://tanelp.github.io/posts/a-bug-that-plagues-thousands-of-open-source-ml-projects/
BloodAxe commented
This repo is not affected by a well-known behavior of np.random under multiprocessing in python.
xdtl commented
It's good to know that. Thanks!