The audios aren't good or the transcription isn't right

fuck

Guess I'm just gonna have to find another character to voice the current version of my OC Carly Pressman

question for u guys: about how many wav files/minutes of audio go into a relatively accurate/pretty passable model? i saw online that its 800-1000 wavs or like 30min of audio

Thout would be recommended for anything good, but with Pipeline, you don't need as much.

oh dang thanks you just saved me a ton of time

Try Using Pipeline, It's Easier

Actually, I've decided to make a model of Penny from WarioWare

I'm getting this error when converting to mels. I'm not sure if it's a problem with too few wav files or what?

Check for any emtpy lines at the end of the transcript.

Any Entropy lines in general. A chance you might have hit enter twice

thanks

that worked

I always keep getting this error even tho I already ran cell 1 Exception Traceback (most recent call last) in 55 56 if not os.path.isdir("/content/tacotron2/wavs"): ---> 57 raise Exception("You must run step 1 before running this.") 58 59 wavs_path = "/content/tacotron2/wavs" Exception: You must run step 1 before running this.

If you have trouble using any notebooks, make sure you use the correct notebooks, which are linked in the #notebooks channel, as those notebooks are the most up-to-date ones.

My model doesn't appear under the models folder in the Tacotron folder after training it, what's going on?

Try Asking EthanRhys Why

There ARE other contributors than Ethan...

Best guess would be drive is full, or there's some problem within the Nb itself that errored out

Does anybody have a testing notebook that is configured with updates for tacotron2, or a working legacy notebook for testing models?

No, I actually have plenty of Drive space available

I guess the notebook is broken

It Needs To Say Saved To Drive In Order For It To Appear

Even after it says "saved to drive," nothing happens

what's better less audio clips but longer in time (between 7-10 seconds) or more audio clips but shorter in time (better 2-5 seconds)?

less with longer length

especially if there’s multiple sentences

okay thank you!

Try Pipeline