OSError pytorch

[yolichek]

When run via jupyter :

WARNING[XFORMERS]: xFormers can't load C++/CUDA extensions. xFormers was built for: PyTorch 2.0.1+cu118 with CUDA 1108 (you have 2.0.1+cpu) Python 3.10.11 (you have 3.10.14) Please reinstall xformers (see https://github.com/facebookresearch/xformers#installing-xformers) Memory-efficient attention, SwiGLU, sparse and more won't be available. Set XFORMERS_MORE_DETAILS=1 for more details

torch.cuda.is_available() False

OSError: Error no file named pytorch_model.bin, tf_model.h5, model.ckpt.index or flax_model.msgpack found in directory C:\Users...

[HowlingNitin]

1. Reinstall xFormers:

   • Follow the instructions provided in the warning message:

     pip uninstall xformers
     pip install xformers --extra-index-url https://download.pytorch.org/whl/cu118

   • This will reinstall xFormers with the correct CUDA support for your environment (assuming your GPU is compatible with CUDA 11.8).

   If your GPU doesn't support CUDA 11.8:

   • Check the xFormers documentation for compatible CUDA versions and reinstall with the appropriate version using the --extra-index-url option.
   • Consider upgrading your GPU drivers if necessary.

2. Locate or Provide the Model File:

   • Make sure the model file you're trying to load exists in the specified directory (C:\Users...). Double-check the path and filename.
   • If the model file isn't generated by your code, ensure you have downloaded it from the correct source.

Additional Tips:

• CUDA Compatibility: Check your NVIDIA GPU's compute capability to ensure it's compatible with the required CUDA version. You can find this information using the nvidia-smi command in your terminal or by searching for your GPU model online.
• CUDA Environment Setup: If you haven't already, ensure you have CUDA Toolkit and the necessary cuDNN libraries installed. Refer to the PyTorch documentation for instructions on setting up a CUDA environment: https://pytorch.org/
• Jupyter Kernel Restart: After reinstalling xFormers and (if applicable) updating your environment, restart your Jupyter kernel to ensure the changes take effect.