How to use espnet/visinger2-zh-jp-multisinger-svs with ESPnet:
unknown model type (must be text-to-speech or automatic-speech-recognition)
