#6182 No CUDA detected. Check NVIDIA GPU/Drivers. See Help -> CUDA.

This error means pyVideoTrans did not detect a usable NVIDIA CUDA environment on your machine. The usual causes are: your GPU is not NVIDIA, the NVIDIA driver is outdated, CUDA Toolkit 12.8+ / cuDNN 9.11+ is not installed, or the CUDA bin/lib paths are missing from your system PATH.

Follow these checks:

1. Confirm your GPU is NVIDIA
   Open CMD and run nvidia-smi.
   If it shows GPU/driver info, the driver is fine.
   If it says "NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver", update your GPU driver to the latest version from NVIDIA's website.
2. Verify CUDA Toolkit version
   Run nvcc -V in CMD.
   pyVideoTrans requires CUDA Toolkit ≥ 12.8.
   If the command is not found, install CUDA 12.8+ and ensure its bin and lib directories are added to the system PATH environment variable.
3. Verify cuDNN version
   Install cuDNN ≥ 9.11 and add its bin/lib directories to PATH as well.
   After changing PATH/environment variables, restart the application and the terminal.

4. Check your PyTorch build (important for source deployment)
   You are running from source code (frozen:False). Verify that your Python environment has a CUDA-enabled PyTorch build. Run:

   import torch
   print(torch.cuda.is_available())

   If it prints False, the installed torch is a CPU-only version and cannot use CUDA. You must reinstall a CUDA-enabled torch build compatible with CUDA 12.8.

5. CPU fallback
   Even without CUDA, the software can still run on CPU, but local ASR/TTS will be extremely slow. For normal use, it is recommended to fix the CUDA environment.

If the issue persists after these steps, please provide the latest log file from the logs/ folder (the last ~30 lines) so we can analyze further.

Please consult related documents:

• Configure CUDA environment on Windows
• Install CUDA and cuDNN on Windows