seq2seq-EVC
seq2seq-EVC copied to clipboard
KunZhou9646
This is the implementation of our Interspeech 2021 paper: Limited data emotional voice conversion leveraging text-to-speech: two-stage sequence-to-sequence training.
Hello dear thanks for your work. I am working on your repository. in pretrain step. at first hundred interations (300, 400), I get Nan or Inf validation loss. It is...
Hello, May I ask which script is for calculating Mel-cepstral distortion(MCD) and average absolute differences of the utterance duration (DDUR)?
Hello, may I ask if this process is speaker independent? Does it mean that I can continue to use multi speaker data for the second stage of training after the...
Metadata
Owner
Metadata
This is the implementation of our Interspeech 2021 paper: Limited data emotional voice conversion leveraging text-to-speech: two-stage sequence-to-sequence training.