CUDA GPU is not available. Using CPU.

[Sevallis]

Hi, I'm on Windows 11 Pro 23H2 , and I installed eSpeak via MSI file and Abogen via Windows script, and when I run a book it tells me that CUDA is not available even though I have a 5070ti. I didn't notice any errors about CUDA during the install terminal text. Any idea how to get it to accelerate creation with my GPU?

[denizsafak]

Please go to abogen's folder, open terminal and type:

python_embedded\python.exe -m pip install gpustat && python_embedded\python.exe main\is_nvidia.py

and show me the output at the end

[Sevallis]

At line:1 char:51
+ python_embedded\python.exe -m pip install gpustat && python_embedded\ ...
+                                                   ~~
The token '&&' is not a valid statement separator in this version.
    + CategoryInfo          : ParserError: (:) [], ParentContainsErrorRecordException
    + FullyQualifiedErrorId : InvalidEndOfLine

[denizsafak]

What about this

python_embedded\python.exe -m pip install gpustat; python_embedded\python.exe main\is_nvidia.py

[Sevallis]

python_embedded\python.exe : The module 'python_embedded' could not be loaded. For more information, run
'Import-Module python_embedded'.
At line:1 char:1
+ python_embedded\python.exe -m pip install gpustat; python_embedded\py ...
+ ~~~~~~~~~~~~~~~~~~~~~~~~~~
    + CategoryInfo          : ObjectNotFound: (python_embedded\python.exe:String) [], CommandNotFoundException
    + FullyQualifiedErrorId : CouldNotAutoLoadModule

python_embedded\python.exe : The module 'python_embedded' could not be loaded. For more information, run
'Import-Module python_embedded'.
At line:1 char:52
+ ... python.exe -m pip install gpustat; python_embedded\python.exe main\is ...
+                                        ~~~~~~~~~~~~~~~~~~~~~~~~~~
    + CategoryInfo          : ObjectNotFound: (python_embedded\python.exe:String) [], CommandNotFoundException
    + FullyQualifiedErrorId : CouldNotAutoLoadModule

[Sevallis]

oops, wrong directory. Ran it in abogen-main and heres the output:

Requirement already satisfied: gpustat in c:\users\username\downloads\abogen-main\abogen-main\python_embedded\lib\site-packages (1.1.1)
Requirement already satisfied: nvidia-ml-py>=11.450.129 in c:\users\username\downloads\abogen-main\abogen-main\python_embedded\lib\site-packages (from gpustat) (12.575.51)
Requirement already satisfied: psutil>=5.6.0 in c:\users\username\downloads\abogen-main\abogen-main\python_embedded\lib\site-packages (from gpustat) (7.0.0)
Requirement already satisfied: blessed>=1.17.1 in c:\users\username\downloads\abogen-main\abogen-main\python_embedded\lib\site-packages (from gpustat) (1.21.0)
Requirement already satisfied: wcwidth>=0.1.4 in c:\users\username\downloads\abogen-main\abogen-main\python_embedded\lib\site-packages (from blessed>=1.17.1->gpustat) (0.2.13)
Requirement already satisfied: jinxed>=1.1.0 in c:\users\username\downloads\abogen-main\abogen-main\python_embedded\lib\site-packages (from blessed>=1.17.1->gpustat) (1.3.0)
Requirement already satisfied: ansicon in c:\users\username\downloads\abogen-main\abogen-main\python_embedded\lib\site-packages (from jinxed>=1.1.0->blessed>=1.17.1->gpustat) (1.89.0)
C:\Users\username\Downloads\abogen-main\abogen-main\python_embedded\python.exe: can't open file 'C:\\Users\\username\\Downloads\\abogen-main\\abogen-main\\main\\is_nvidia.py': [Errno 2] No such file or directory
PS C:\Users\username\Downloads\abogen-main\abogen-main>

[denizsafak]

Sorry please change "main" with "abogen"

[Sevallis]

New text says

Requirement already satisfied: ansicon in c:\users\username\downloads\abogen-main\abogen-main\python_embedded\lib\site-packages (from jinxed>=1.1.0->blessed>=1.17.1->gpustat) (1.89.0)
NVIDIA GeForce RTX 5070 Ti
True

tested an epub file again and it still claims "Checking GPU acceleration...
CUDA GPU is not available. Using CPU." at the beginning though. Task Manager confirms 50% CPU i7-12700k. None on 5070ti. Takes 1:31 to generate a 8 min 35 sec audio of an introduction section.

Edit: got to get some sleep for now, thanks for the help. Will check back after work tomorrow when I can get to my PC again, most likely.

[denizsafak]

Hmm, okay. Now type:

python_embedded\python.exe -m pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu128

And try again

[Sevallis]

Edit: Ran again, took 24 seconds with GPU acceleration!

Seemed fine, but this appears relevant:

Downloading https://download.pytorch.org/whl/cu128/torchvision-0.22.1%2Bcu128-cp312-cp312-win_amd64.whl (7.6 MB)
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 7.6/7.6 MB 52.7 MB/s eta 0:00:00
Downloading https://download.pytorch.org/whl/cu128/torch-2.7.1%2Bcu128-cp312-cp312-win_amd64.whl (3273.0 MB)
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 3.3/3.3 GB 32.0 MB/s eta 0:00:00
Downloading https://download.pytorch.org/whl/cu128/torchaudio-2.7.1%2Bcu128-cp312-cp312-win_amd64.whl (4.7 MB)
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 4.7/4.7 MB 70.9 MB/s eta 0:00:00
Downloading https://download.pytorch.org/whl/pillow-11.0.0-cp312-cp312-win_amd64.whl (2.6 MB)
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 2.6/2.6 MB 49.1 MB/s eta 0:00:00
Installing collected packages: pillow, torch, torchvision, torchaudio
  Attempting uninstall: torch
    Found existing installation: torch 2.7.1
    Uninstalling torch-2.7.1:
      Successfully uninstalled torch-2.7.1
   ━━━━━━━━━━╺━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 1/4 [torch]  WARNING: The scripts torchfrtrace.exe and torchrun.exe are installed in 'C:\Users\username\Downloads\abogen-main\abogen-main\python_embedded\Scripts' which is not on PATH.
  Consider adding this directory to PATH or, if you prefer to suppress this warning, use --no-warn-script-location.
Successfully installed pillow-11.0.0 torch-2.7.1+cu128 torchaudio-2.7.1+cu128 torchvision-0.22.1+cu128

[Kinasa0096]

Hmm, okay. Now type:

python_embedded\python.exe -m pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu128

And try again

Had the same issue with my NVIDIA RTX 3070 Ti. It got fixed after running this command. Thanks.

[jefro108]

Wondering if it's possible to add support for CUDA in M1 Mac using PyTorch as shown in https://youtu.be/53PjsHUd46E 🙏🏻

[jefro108]

Was able to implement a version of the above command via https://stackoverflow.com/a/77646995

pip3 install --pre torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/nightly/cpu
Only thing is it doesn't seem to work. Abogen doesn't recognise a GPU - it says

CUDA GPU is not available. Using CPU. (No devices found)CUDA GPU is not available. Using CPU. (No devices found)

[jefro108]

I see now that in the code

# Enable MPS GPU acceleration on Mac Apple Silicon
if platform.system() == "Darwin" and platform.processor() == "arm":
    os.environ["PYTORCH_ENABLE_MPS_FALLBACK"] = "1"

it should work for Mac Apple Silicon

I've installed torch torchvision torchaudio so what else do I need to do to make it work on a Mac?

[denizsafak]

@jefro108 The “CUDA available” message only appears when the system detects a compatible NVIDIA GPU with CUDA support (or in some cases, an AMD GPU with ROCm).

Since macOS doesn't support either CUDA or ROCm, you won’t see that message on a Mac. While your Mac will still be able to run the program, it’ll do so using the CPU or Apple's own GPU architecture, which works, but won’t offer the same performance as a dedicated CUDA-capable GPU.

So everything should still function, just with reduced acceleration compared to systems with NVIDIA or AMD GPU support.

[jefro108]

Thanks for explaining all that.

The confusion comes from abogen telling me

Checking GPU acceleration

And then I’m told CUDA GPU is not available.

If I was told the Apple Silicon GPU is being used I wouldn’t have had a question about this.

I request that you add an appropriate message when Apple Silicon acceleration is happening.