Transcriptions are accurate. Having the same problem generating a British accent now, which seems weird.
That all makes sense. I have found more epochs leads to cleaner audio...but I'm still getting a smoothed out version of the voice when I generate. What I want is a thick accent like the dataset…
I did an additional 500 iterations with a new dataset of the same voice over the original model I trained as the Source Model... No change at all. Getting an accent to shine through is difficult.…
Thanks for that. I'll give it a shot. When you say "Source Model" is that the Autoregressive model under settings?
I am finding this very slow as well. It seems the same file using the same voice folder and model in DLAS generates significantly faster than here in the AI-Voice-Cloning using the same ultra-fast…
It was also missing /models/tortoise/train_diffusion_vocoder_22k_level.yml so I restored that as well... Now it seems to be training.
I did run update.bat. Will try the git restore...
I ran the initial setup script. "setup-cuda.bat"
Is there a different one?
Dangit. I just found that one and was going to post it here... Thanks.