1. Try synthesize with short text (ex. "test" or "테스트") 2. Sometimes success, sometimes returns audio with ref text (not "test" that i requested) This can re-produce with predefined model mika. my model was kr model with ref text "테스트"
This can re-produce with predefined model mika. my model was kr model with ref text "테스트"