search

Vision-CAIR / MiniGPT-4

Open-sourced codes for MiniGPT-4 and MiniGPT-v2 (https://minigpt-4.github.io, https://minigpt-v2.github.io/)
https://minigpt-4.github.io
BSD 3-Clause "New" or "Revised" License
25.33k stars 2.91k forks source link

Bitsandbytes cant find CUDA #63

Open dangelo352 opened 1 year ago

dangelo352 commented 1 year ago

keep getting CUDA Setup failed despite GPU being available. Please run the following command to get more information:

    python -m bitsandbytes

    Inspect the output of the command and see if you can locate CUDA libraries. You might need to add them
    to your LD_LIBRARY_PATH. If you suspect a bug, please take the information from python -m bitsandbytes
    and open an issue at: https://github.com/TimDettmers/bitsandbytes/issues

    I've done everything I could think of I'm growing very confused I have reinstalled tried adding the correct paths to your LD_LIBRARY_PATH variable: installed pytorch 15 different times, installed from source, got CUDA manually from nvida. Idk what to do please help
FrancescoSaverioZuppichini commented 1 year ago

please refer to the bitsansbytes repo -> you need to have the correct version of cuda

https://github.com/TimDettmers/bitsandbytes/issues/112

kelheor commented 1 year ago

Fix your bitsadnbytes package (C:\Users\yourusername\anaconda3\envs\minigpt4\Lib\site-packages\bitsandbytes): https://github.com/DeXtmL/bitsandbytes-win-prebuilt Download dlls and read this: https://github.com/Vision-CAIR/MiniGPT-4/issues/28#issuecomment-1514181638 Then execute in case if you have error that torch is not compiled with CUDA support: pip uninstall torch pip cache purge pip install torch -f https://download.pytorch.org/whl/torch_stable.html pip uninstall torchaudio pip install torchaudio