Closed piaohai closed 3 years ago
A way to synthesize gestures from a plain text is not available in this repository, because the code is assumed to be run with the GENEA workshop dataset. I recommend you to modify SubtitleWrapper
class to support your own transcripts. Note that you should provide timestamps for each word.
You might be interested in our new work Gesture-Generation-from-Trimodal-Context which is a follow-up work of this repository and provides what you want, synthesizing gestures from a plain text, so please check it out.
hi,guys, would you pleases provide some TRANSCRIPT demo,i want to run test, tks.