The Tacotron2 model seems to be severely fragmented.

[kbuwel]

This is like the official Tacotron2 developer forum, so we post improvements.
I wish they officially supported multiple languages.
I speak Korean, and the Korean Tacotron2 also seems to be very fragmented.
I think it is similar to the old Android operating system.
Wouldn't fragmentation be reduced if Tacotron2 officially supported multiple languages?
There are currently too many Tacotron2 models.
I don't know which one to use.
It's confusing because the dataset folders, dataset files, and commands are all different.
I think this problem would be solved if multiple languages were officially supported.