litagin02

本家RVCやVC Clientでは、harvestに代わる学習・推論時のピッチ推定としてrmvpeが導入されており、 harvestと比べて速度の面でも質の面でも高くなっております。 それのサポートがここでもあればよいと思います。

Hello, and thanks for nice toolkit! When I tried to use MaskGCT and looked some codes in it, I realized that it seems that a code on Japanese g2p was...

PyTorch inference / lazy importing is 100x slower in callbacks on Windows (Gradio 6.0 vs 5.x)

6

### Describe the bug Transformers BERT model inference is approximately 100x slower when executed inside a Gradio 6.0 callback compared to Gradio 5.x on Windows. In the code below, on...