Support training variance models from DS files

openvpi / DiffSinger

An advanced singing voice synthesis system with high fidelity, expressiveness, controllability and flexibility based on DiffSinger: Singing Voice Synthesis via Shallow Diffusion Mechanism

Apache License 2.0

2.62k stars 275 forks source link

Support training variance models from DS files #132

Closed yqzhishen closed 11 months ago

yqzhishen commented 11 months ago

Implementation of #131.