compatible with master version acoustic pre-train

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

compatible with master version acoustic pre-train #141

Closed nestyme closed 10 months ago

nestyme commented 10 months ago

Hello! I am trying to find working with latest master pretrain in Releases, but cannot find any. Could you please help me and refer more directly where I can find acoustic pretrain for the latest master version? Thank you

yqzhishen commented 10 months ago

Former model releases of MIDI-less mode are still compatible, but they need to be converted to acoustic models before they can be loaded. See release notes of v2.0.0 for migration tips.

nestyme commented 10 months ago

thank you!