A Python library for audio data augmentation. Inspired by albumentations. Useful for machine learning.

multi-task and multi-track music transcription for everyone

Code for the paper Hybrid Spectrogram and Waveform Source Separation

An unofficial PyTorch implementation of Music Source Separation with Band-split RNN for MDX-23 ("Label Noise" Track)

Robust Speech Recognition via Large-Scale Weak Supervision