DoctorPopi
  • Joined on 2023-06-06
DoctorPopi commented on issue mrq/ai-voice-cloning#400 2023-10-01 18:57:54 +00:00
Upscaled output creates bad quality?

Hey! Thank you for your answers! I'll investigate this more. I'm indeed not using the Voice Fixer, I find it a bit unstable and it messes with non verbal sounds.

DoctorPopi opened issue mrq/ai-voice-cloning#400 2023-09-30 16:52:29 +00:00
Upscaled output creates bad quality?
DoctorPopi commented on issue mrq/ai-voice-cloning#355 2023-09-29 06:54:46 +00:00
RAM memory leak

Hey, after prolonged usage, I’m having the same problem I think. I cannot generate anything else through Mrq without having the OOM Cuda allocation error. What’s even weirder, I just shut down…

DoctorPopi commented on issue mrq/ai-voice-cloning#316 2023-09-28 13:26:07 +00:00
getting error when using emotions

Nevermind, the issue came back with me as well, it seems quite random...

DoctorPopi commented on issue mrq/ai-voice-cloning#248 2023-09-26 09:03:17 +00:00
Overfitting with large datasets

Hey! Well it does sound pretty good, even though of course it’s a bit random, but I’d say it outputs good results 40% of the time. What is a bit frustrating though is that the results are NOT…

DoctorPopi commented on issue mrq/ai-voice-cloning#248 2023-09-25 11:41:46 +00:00
Overfitting with large datasets

Great news!! I've managed to get a very good validation curve over almost 200 epochs, only by changing the learning rate from 1e-05 to 1e-06. The training loss mel is still around 2.0, but I think…

DoctorPopi commented on issue mrq/ai-voice-cloning#248 2023-09-24 13:35:03 +00:00
Overfitting with large datasets

Hello!

Allow me to share my own experience about this matter. I, too, have been struggling for like a week on the problem of the validation curve going up almost immediately, like as soon as…

DoctorPopi commented on issue mrq/ai-voice-cloning#316 2023-09-23 16:41:10 +00:00
getting error when using emotions

Hey! I noticed the same problem, and I am also certain it used to work before!

DoctorPopi commented on issue mrq/ai-voice-cloning#396 2023-09-22 13:46:31 +00:00
Validation.txt shows nothing

Okay thank you for these clarifications!

I'm really struggling with the validation though, the validation yellow line starts going up almost as soon as the training starts :/ I've seen [this…

DoctorPopi closed issue mrq/ai-voice-cloning#396 2023-09-22 13:46:31 +00:00
Validation.txt shows nothing
DoctorPopi commented on issue mrq/ai-voice-cloning#396 2023-09-22 06:50:44 +00:00
Validation.txt shows nothing

Oh okay I’ll try working with the text length then for now. Write talking about the number of characters right, spaces included?

Thank you!

DoctorPopi closed issue mrq/ai-voice-cloning#384 2023-09-21 15:33:35 +00:00
Why so many models? And about a thousand of other questions :)
DoctorPopi opened issue mrq/ai-voice-cloning#396 2023-09-21 11:38:39 +00:00
Validation.txt shows nothing
DoctorPopi closed issue mrq/ai-voice-cloning#389 2023-09-18 05:34:56 +00:00
Recreate dataset after correcting whisper.json doesn't overwrite the train.txt
DoctorPopi commented on issue mrq/ai-voice-cloning#389 2023-09-18 05:34:52 +00:00
Recreate dataset after correcting whisper.json doesn't overwrite the train.txt

Okay I'm closing the issue because I figured out my process. For those who are interested, here's what I think you have to do:

Case 1 - You want to add more audio clips after having started…

DoctorPopi opened issue mrq/ai-voice-cloning#389 2023-09-17 19:37:16 +00:00
Recreate dataset after correcting whisper.json doesn't overwrite the train.txt
DoctorPopi commented on issue mrq/ai-voice-cloning#250 2023-09-17 15:05:01 +00:00
resume a training

Perfect! Thank you!

DoctorPopi commented on issue mrq/ai-voice-cloning#384 2023-09-17 15:04:46 +00:00
Why so many models? And about a thousand of other questions :)

Thank you very much for all that information! I think I have all I need for now, I'll make another post should more questions arise. A thousand thanks for your patience and for the great work you…

DoctorPopi commented on issue mrq/ai-voice-cloning#250 2023-09-17 12:09:06 +00:00
resume a training

Hi! Is it possible to resume a training and add new clips? Or should that rather be another finetuning altogether?

DoctorPopi commented on issue mrq/ai-voice-cloning#384 2023-09-16 17:31:31 +00:00
Why so many models? And about a thousand of other questions :)

After preparing the dataset but before actually training, under train.txt, add in the additional non-verbal terms into there.

Should I also add them in the actual whisper.json transcription?