• Joined on 2023-11-01
elbow commented on issue mrq/ai-voice-cloning#147 2023-11-10 19:04:54 +00:00
Discussion about Fine Tuning on a different language.

Trained a new model with the Japanese tokenizer, and after ~55 epochs (~825000 samples processed),…

elbow commented on issue mrq/ai-voice-cloning#435 2023-11-02 17:44:11 +00:00
[Discussion] What is a large dataset?

Would say that the minimum would need to be around 5000 to avoid overfitting then?

That is really hard to say. It really depends on dataset entropy and variance... I usually don't go bellow…

elbow commented on issue mrq/ai-voice-cloning#435 2023-11-02 16:52:34 +00:00
[Discussion] What is a large dataset?

I agree, a LR finder would be awesome. I indeed tried very low learning rates but the model remains stuck at pretty high values (around 2) and doesn't go down anymore (I've read somewhere that…

elbow commented on issue mrq/ai-voice-cloning#435 2023-11-02 12:14:47 +00:00
[Discussion] What is a large dataset?

Hey! Thank you, yes indeed I'm just trying to refine an english speaking model for a game character. Thank you for the finetuned model, if I understand correctly, James Becker took the feminine…

elbow commented on issue mrq/ai-voice-cloning#433 2023-11-01 21:00:07 +00:00
Training multiple finetunes on a single dataset in GUI?

Is there a supported method in the GUI for testing out different training settings and switching between the final finetune?

No not really. It would be nice to have learning rate finder, but…

elbow commented on issue mrq/ai-voice-cloning#435 2023-11-01 20:50:00 +00:00
[Discussion] What is a large dataset?

Depends what are you trying to achieve - when I am trying to finetune to new language I have a dataset of cca 500k clips (about 1000hours of audio) - James Becker (the original creator of tortoise…