Which will hopefully let me create lots of voices faster

It's just too time consuming now to create a single voice.

@User I am guessing fastpitch multispeaker pretraining doesn't work well for you ?

I actually haven't tried this

Do you have a link where i could check it out?

oh gotcha

You have to dig through the issues to find how to set it up with multispeaker

and probably look through the code, I have not had time because I am building a Interactive Waifu AI Platform that just uses these models. To raise money for my wife's and son's gene therapy over the next couple of years

oh wow. that's really cool. would love to follow along with your project.

and thanks for the tip. will take a look.

i try to clone someones a voice this time rn

@User yeah let me know how it goes, I am gonna try it in a few weeks, I don't have links to the project right now, ill send it to you when I make it more public 🙂

i cloned donald trumps voice

Tonald Drump

Michael stevens

@User are you using taco2 to try to replicate zero shot?

the modified tacotron2 described in https://arxiv.org/abs/1910.10838, yeah

from reading the paper seems like the main purpose of the additional attention layers was for japanese, but I guess I'll see how it goes with english 🙃

@everyone i found a app called pydroid its like python

Dead chat

what am i supposed to do with this information?

chat's not dead

chat's just resting

How good is michael Stevens voice?. If you havent still Trained yet. no problem

I haven't trained since the first time with all numbers. Training my zero-shot model now and will hopefully have some initial results soon.

Ok

Did you know the Michael stevens memes?

I do not

they use sentence mixing and that doesnt make any sense