Error generating speech

[TeeFReUnD]

Hello everyone,

I´m quite new to ComfyUI and i´m sure, this is a quite easy problem to solve, but i´m not sure, what i did wrong.
I downloaded all tokenizer files, created the subfolders and cloned both VibeVoice models via git. After that, i try to recreate the demo workflow, but when i try to run it, i get following error:

VibeVoiceSingleSpeakerNode
Error generating speech: Model loading failed: Failed to load model from C:\ComfyUI\ComfyUI\models\vibevoice\VibeVoice-1.5B.
Please ensure the model files are complete and properly downloaded.
Required files: config.json, pytorch_model.bin or model safetensors
Error: Expected a torch.device with a specified index or an integer, but got:cuda

Anyone got an idea, what i did wrong?
If you need any additional information, please let me know.

Looking forward to your help.

Best regards,
Simon

[agusliem]

maybe you need download all the necessary files, such as config.json and others on same directory which is in your case C:\ComfyUI\ComfyUI\models\vibevoice\

follow the format like this:
├── VibeVoice-1.5B/ # Model folder
│ ├── config.json
│ ├── model-00001-of-00003.safetensors
│ ├── model-00002-of-00003.safetensors
│ └── ... (other model files)

download the files from : https://huggingface.co/microsoft/VibeVoice-1.5B/tree/main and put them all on C:\ComfyUI\ComfyUI\models\vibevoice\VibeVoice-1.5B\

this should be solve your problems.

[TeeFReUnD]

Strange, i tried to manually download all files from above mentioned and put them in the mentioned directory but still receive the same error:

Anyone got another idea, what i could have done wrong?

[agusliem]

check your torch, perhaps you not install the pytorch espesially the audio torch

if you use comfyui portable then go to python embeded than type cmd, than type this command:

python.exe -m pip show torchaudio

if you use comfyui installation into windows system than type this command directly at C: directory :

pip show torchaudio

here the example of my torchaudio

[TeeFReUnD]

I got the portable edition too and when i run the command, i get this:

Seems like i have torchaudio installed, but it is an older version. Could this be problematic?

[TeeFReUnD]

Hm, i think, i can see the problem. When i start my portable ComfyUI, i get following screen:

C:\ComfyUI>.\python_embeded\python.exe -s ComfyUI\main.py --windows-standalone-build
[START] Security scan
[DONE] Security scan
'## ComfyUI-Manager: installing dependencies done.
** ComfyUI startup time: 2025-10-09 08:53:34.626
** Platform: Windows
** Python version: 3.11.8 (tags/v3.11.8:db85d51, Feb 6 2024, 22:03:32) [MSC v.1937 64 bit (AMD64)]
** Python executable: C:\ComfyUI\python_embeded\python.exe
** ComfyUI Path: C:\ComfyUI\ComfyUI
** ComfyUI Base Folder Path: C:\ComfyUI\ComfyUI
** User directory: C:\ComfyUI\ComfyUI\user
** ComfyUI-Manager config path: C:\ComfyUI\ComfyUI\user\default\ComfyUI-Manager\config.ini
** Log path: C:\ComfyUI\ComfyUI\user\comfyui.log

Prestartup times for custom nodes:
0.0 seconds: C:\ComfyUI\ComfyUI\custom_nodes\rgthree-comfy
0.0 seconds: C:\ComfyUI\ComfyUI\custom_nodes\ComfyUI-Easy-Use
6.7 seconds: C:\ComfyUI\ComfyUI\custom_nodes\ComfyUI-Manager

Warning, you are using an old pytorch version and some ckpt/pt files might be loaded unsafely. Upgrading to 2.4 or above is recommended.
Total VRAM 24576 MB, total RAM 32690 MB
pytorch version: 2.3.1+cu121
Set vram state to: NORMAL_VRAM
Device: cuda:0 NVIDIA GeForce RTX 3090 : cudaMallocAsync
Please update pytorch to use native RMSNorm
Torch version too old to set sdpa backend priority.
Using pytorch attention
Python version: 3.11.8 (tags/v3.11.8:db85d51, Feb 6 2024, 22:03:32) [MSC v.1937 64 bit (AMD64)]
ComfyUI version: 0.3.63
ComfyUI frontend version: 1.28.2
[Prompt Server] web root: C:\ComfyUI\python_embeded\Lib\site-packages\comfyui_frontend_package\static
C:\ComfyUI\python_embeded\Lib\site-packages\pydantic_internal_fields.py:132: UserWarning: Field "model_url" in Model has conflict with protected namespace "model_".`

I tried to update pytorch using following command in my C:\ComfyUI\python_embeded\python.exe directory: pip install --upgrade torch torchvision torchaudio and it also updates to version 2.8.0 installed, but somehow, ComfyUI does not recognize this. Is there a way to change this?