Tacotron gst branch

[JRMeyer]

>>> geneing
[July 10, 2019, 5:54pm]

What's the plan for the GST branch? Are you planning to implement
Predicting Expressive Speaking Style From Text In End-To-End Speech
Synthesis paper?

[This is an archived TTS discussion thread from discourse.mozilla.org/t/tacotron-gst-branch]

[JRMeyer]

>>> erogol
[July 10, 2019, 9:02pm]

I plan to learn speaker style with GST and direct it for inference.
However, so far what GST learns is quite random hence hard to
manipulate.

[Archived Post]