Audio augmentations library for PyTorch for audio in the time-domain, with support for stochastic data augmentations as used often in self-supervised / contrastive learning.