Amazing model - Onnx model for low resource
#2
by
sunnyglow
- opened
Hi Team, I tested this model and it is absolutely amazing with amazing quality output. However, this model requires some high-end machine and significant amount of power. Can you please convert this model to onnx version or some low resource version. Here is a link that already converts F5-tts model into multiple variations so that everybody can benefit. once again very good model and amazing work.
https://github.com/DakeQQ/F5-TTS-ONNX
Or a quantized model also will help alot.
i am also getting noise only, any one got the good output, if yes what kind of fix we have to do?
Even i got noisy output , i mean it's able to clone the voice but along with it much background noise comes