STAGE

Listen to audio samples from STAGE, our STemmed Accompaniment GEneration Model.
All generated samples are straight from the model, with no post-processing.
All contexts are taken from an isolated test dataset.
Read the arXiv.