I am having some issues with getting CUDA to work with pytorch on my laptop. I have a gtx 1660ti laptop and I know that all GPUs from NVIDIA that are newer than 8 years are CUDA compatible.
I have created a new conda environment with python 3.9 and ran the command: conda install pytorch torchvision torchaudio pytorch-cuda=11.8 -c pytorch -c nvidia
As per the pytorch website to install it with CUDA so that my GPU can be used. However if I make a jupyter notebook and run this:
import torch
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
print(f"Using device: {device}")
I get this output:
/home/joash/anaconda3/envs/pytorchCUDA/lib/python3.9/site-packages/torch/cuda/init.py:141: UserWarning: CUDA initialization: Unexpected error from cudaGetDeviceCount(). Did you run some cuda functions before calling NumCudaDevices() that might have already set an error? Error 804: forward compatibility was attempted on non supported HW (Triggered internally at /opt/conda/conda-bld/pytorch_1711403392949/work/c10/cuda/CUDAFunctions.cpp:108.) return torch._C._cuda_getDeviceCount() > 0
As you can see it is running in that conda environment but its not picking up the GPU, the only thing I can think of is that the GPU is not CUDA compatible. But if I run: nvcc --version
This is the output:
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2024 NVIDIA Corporation
Built on Tue_Feb_27_16:19:38_PST_2024
Cuda compilation tools, release 12.4, V12.4.99
Build cuda_12.4.r12.4/compiler.33961263_0
Any help on this matter will be greatly appreciated.