Mayavoz Versions Save

Stable API

• Support for train & inference
• Available Model architectures: WaveUnet, Demucs, DCCRN
• Available loss & metric functions: mea, mse, si-snr, si-sdr, pesq, stoi.

Next Focus:

• documentations & examples.
• pretrained models : train on more datasets and improve the performance of already available models.
• benchmark the improvement attained in different audio tasks using mayavoz