-
Notifications
You must be signed in to change notification settings - Fork 152
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
How to do voice cloning with Tortoise TTS? #242
Comments
Tortoise has more than one way to do it. |
The cheapest way |
Get some mp3 files of the voice you want, then put them in a folder. Next you will need to put these in the tortoise voices folder. I can't access my workstation at the minute but I think it was voices-tortoise inside the tts-generation-webui directory. |
Yes, but I can't see if you selected a voice Default means it's not a fine-tuned version |
I mean the time it takes to generate autoregressive samples. Is it supposed to take that long? |
That's slow. Might be using CPU not GPU. Also might have a lot of autoregressive samples too. You can check the task manager to see if GPU is even moving from this. |
CPU usage was at 70%. How do I make it use GPU? |
Might be an installation issue. I'll need my workstation, talk to you
tomorrow.
The core issue is usually pytorch being the wrong version.
…On Wed, Jan 10, 2024, 12:14 AM Gushousekai195 ***@***.***> wrote:
That's slow. Might be using CPU not GPU. Also might have a lot of
autoregressive samples too.
You can check the task manager to see if GPU is even moving from this.
CPU usage was at 70%.
How do I make it use GPU?
—
Reply to this email directly, view it on GitHub
<#242 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ABTRXI5WKOSC2DWN3WGKMHDYNW6K7AVCNFSM6AAAAABBTYYDGSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQOBTHA4DMNRTHA>
.
You are receiving this because you modified the open/close state.Message
ID: ***@***.***>
|
So if it's pytorch being the wrong version, how do I install the right version then? |
Did you use the one click installer? And select NVidia GPU? |
Yes, I believe I did. |
Right, then it sounds like an installation issue. |
Like this? |
Ok, I can see some errors and confirm that it's a cpu only install:
First time seeing this error, so this might just be an unlucky/interrupted download or a corrupt file cache etc. |
I get to this point and it just stops
And then after a while this happens:
I even redownloaded the installers. |
That's quite something.. it freezes at 99% and then crashes? Wow.
|
Still having problems
|
Sounds like something is wrong with conda, their project is here: https://github.com/conda/conda/issues |
Do you know how it can be done using pip then? |
This might work, but I am not 100% sure:
I'm not sure if 2.0.0 will install the right torch in this case. It might be 2.0.0+cu118 |
Now I cannot start the webui at all:
|
try
then update_windows.bat |
installing pytorch is a pain, even more so with pip. I really wish conda would fix their issue or provide you with a solution. |
I just want a webui I can voice clone with tortoise-tts with!
|
I just want conda to work without mysterious errors for an user of my webui. |
I think there are some youtube tutorials for installing other tortoise-tts projects like MRQ, those installations might just work. |
Anything with a ui? |
MRQ has UI afaik |
Can you give me a link to it please? |
For training a voice (I assume this is similar to training a LoRA a la Stable Diffusion?), you mentioned that there might be a UI somewhere... Pointers? A lot of this stuff is still word-of-mouth it seems. EDIT: I think this is what @rsxdalv was referring to: https://git.ecker.tech/mrq/ai-voice-cloning |
How can I clone a voice using Tortoise TTS in this webui? Please help?
The text was updated successfully, but these errors were encountered: