Error while generating / no audio

[Takinono]

Checks

• [X] This template is only for bug reports, usage problems go with 'Help Wanted'.
• [X] I have thoroughly reviewed the project documentation but couldn't find information to solve my problem.
• [X] I have searched for existing issues, including closed ones, and couldn't find a solution.
• [X] I confirm that I am using English to submit this report in order to facilitate communication.

Environment Details

Windows 11, RTX 3070

Steps to Reproduce

When I upload my audio sample along with my prompt and hit synthesize, nothing happens. I do have weird error messages:

but it remains stuck at 0% after "Prefix dict has been built successfully".

Edit, tried in CLI, same problem

I tried following these fixes, but it doesn't solve the problem. https://github.com/SWivid/F5-TTS/issues/356

Thanks for your help

✔️ Expected Behavior

No response

❌ Actual Behavior

No response

[SWivid]

What is your torch (cuda or cpu version installed?) and cuda version, and how it the gpu usage when inference. It would be appreciate if closely follow the issue template and provide detailed info to help figure out.

[Takinono]

Sorry, I'm using : Windows 11 RTX 3070 conda 24.9.2 torch Version: 2.5.1 CUDA Version: 12.3

When using CLI, CPU usage is between 70 and 80%, GPU 1% When using web interface, CPU between 70 and 80%, GPU 1 to 5%

[SWivid]

@Takinono so it seems like a torch-cpu version installed try pip install torch==2.5.1+cu121 torchaudio==2.5.1+cu121 --extra-index-url https://download.pytorch.org/whl/cu121 if you still want torch2.5.1 or just following installation instruction: pip install torch==2.3.0+cu118 torchaudio==2.3.0+cu118 --extra-index-url https://download.pytorch.org/whl/cu118

[Takinono]

This solved the problem ! pip install torch==2.5.1+cu121 torchaudio==2.5.1+cu121 --extra-index-url https://download.pytorch.org/whl/cu121

Thanks !

[SWivid]

will close as solved, feel free to open if further questions