CUDA not detected on Windows

Apologies for a simple question. I have installed Torch on Windows 10.

Torch has Cuda .
print( torch.version)
2.7.0+cu118

Cuda also seems to be available on my computer.
nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2022 NVIDIA Corporation
Built on Tue_May__3_19:00:59_Pacific_Daylight_Time_2022
Cuda compilation tools, release 11.7, V11.7.64
Build cuda_11.7.r11.7/compiler.31294372_0

But:
print(torch.cuda.is_available() )
False

This is a new install of both Conda and Torch.

Any help would be appreciated.

Hello @Hector

This error could be due to your NVIDIA GPU not being able to support the CUDA 11.8 version, since you installed PyTorch with CUDA 11.8 support (2.7.0+cu118).

Perhaps you can check which CUDA version does your GPU support, by matching your GPU’s compute capability, microarchitecture and supported CUDA SDK using these tables.


Machine Learning Engineer at RidgeRun.ai
Contact us: support@ridgerun.ai

Thank you, Daniela. This helps.