[CUDA] windows 11 설치 : torch.cuda.is_available()

대상 : NVIDIA GeForce GTX 1650

 

python & conda : https://www.anaconda.com/download

# python 실행시 MS WindowsApps 으로 실행되어 환경변수 WindosApps 의 순서를 맨 아래로 변경.

 

CUDA : https://developer.nvidia.com/cuda-downloads

cuDNN : https://developer.nvidia.com/rdp/cudnn-archive

 

CUDA install 가이드 : https://docs.nvidia.com/cuda/cuda-installation-guide-microsoft-windows/index.html

이해도가 낮아 'CUDA', 'cuDNN' 설치 후 가이드로 진행함.

# https://docs.nvidia.com/cuda/cuda-installation-guide-microsoft-windows/index.html#using-conda-to-install-the-cuda-software
conda install cuda -c nvidia

# https://docs.nvidia.com/cuda/cuda-installation-guide-microsoft-windows/index.html#pip-wheels
pip install --upgrade setuptools pip wheel
pip install nvidia-pyindex

 

pytorch : https://pytorch.org/get-started/locally/

conda install pytorch torchvision torchaudio pytorch-cuda=12.1 -c pytorch -c nvidia

 

 

확인

def get_device():
    if torch.cuda.is_available(): device = 'cuda'
    elif torch.backends.mps.is_available(): device = 'mps'
    else: device = 'cpu'
    return device

print(get_device())

 

 

 

---