How to use espnet/m4singer_svs_naive_rnn_dp with ESPnet:
unknown model type (must be text-to-speech or automatic-speech-recognition)
