still gives me FileNotFoundError: [Errno 2] No such file or directory: 'wavs/\ufeff1.wav'

This keeps happening, and I don't know why.

I have the list like Audioses said.

I have the audios all there and accounted for.

I'm lost at what to do.

Are

\

even allowed in folder paths

does your zip file contain the wavs folder?

yes

I even tried it without and get the same error.

I think I succesfully created a voice, but it slurs alot. Like I mean alot. It's not able to pronounce anything, sometimes fails entirely and so on. Does talknet2 support other languages than english? If so did I need to do something else? The voice I tried was in Turkish and it is half trying to speak english, half turkish. Any advise?

Finally got the hang of it.

On the Hifigan training now (step 11)

How long should I train it to?

Gosmokeless28 was saying 3100 epochs while the training itself says 2k steps.

I trained it to 2k steps to test. But I may retrain it not sure

I just want to be sure if it is worth doing for other languages than English. I hope someone gives some insight into this lol

both of those are reccomendations

what are your suggested parameter settings?

Epoch amount: 200 Batch size: 32 Learning rate: 1e-4 Minimum learning rate: 3e-7

oh someone said that to set the learning rates high. Shoulnd I do that then?

The learning rates that I've suggested are actually lower than the default ones. They're negative numbers.

oh that makes perfect sense lol

I got a slurring model with an other language than English. I am hoping to get better results by retraining the spect model that's why I asked.

You know TalkNet only supports English speech, right?

No. I've been trying to get an answer for that question for a while 😄

that's very unfortunate. I loved the reference audio feature of that. Guess I'Ll have to scrap that completely 😦

What's talknet?

The voice models that support reference audio

ur mums bagina nah jkjk its tacotron's ugly twin

Is anyone else getting this on the controllable?