Here's how it did, it sounded pretty damn good I must say.

@User Could you link the fixed talknet notebook?

The voice im working on is having a bumpy ride right now. It's going from this

to this constantly

https://discord.com/channels/768215836665446480/820353681567907901/867013758835752991 Your alignments look basically perfect. Just ignore the padded area.

Here's how the Stanley Parrable Narrator training is going

Attention graph: Loss is still ~0.35, gonna train it for another day

why did you reverse the y axis

Yes train more!

why did I reverse the y axis?

that's how SV2TTS's architecture reports local attention graphs

Currently updating on the Ed Sullivan model, I added more voice clips weeks ago but yknow.

How would one use a string variables value as an argument in a command?

like If I had to give a path for a command, "command file.txt" how would I set it so that instead of writing the filename I could just pass a variable to it?

Here's a test of my updated Ed Sullivan model.

Well, it didn't say Ed, but yknow.

wow right off the bat with guru ant

the very first epoch

this is how my voice thing im working on sounds, but it just doesn't words properly...

"Why is it not working?" is what i typed

that's just which character the ai is looking at at any given time, which should absolutely start out diagonal because you're starting from a pretrained model that already knows how to speak (but in the wrong voice) personally i'm surprised my model starts out like this instead

This is what the voice sounds like, it's better but how do I make it better and fix it not saying certain words properly?

did you enable raw input?

yes, it doesn't work otherwise

it says like three words kinda decent, but then goes to gibberish

It sounds like it says "the voice is working, but chesrglasrlyn"

Has anyone used Synthbots tool?

hmm

idk what synthbot is