How to use ftshijt/espnet_codec_soundstream_large_v1.2_category with ESPnet:
unknown model type (must be text-to-speech or automatic-speech-recognition)
