Applio
Applio copied to clipboard
Rework batch inference
This PR changes how batch inference works by not reloading the hubert and voice model