How to use espnet/mixdata_svs_visinger2_spkemb_lang_pretrained with ESPnet:
unknown model type (must be text-to-speech or automatic-speech-recognition)
