Hello, I’d like to use torch to run a package with a neural network on my notebook but meet a problem: no kernel image is available for execution on the device. Information about the system and GPU is described below.
System: window 10
python version: 3.11.5
GPU: NVIDIA GeForce MX250
architecture: PASCAL GPU Architecture Compatibility
compute capability: 6.1
CUDA toolkit: 11.8
torch version: 2.0.0+cu118
torch.cuda.get_arch_list() :
[‘sm_37’, ‘sm_50’, ‘sm_60’, ‘sm_61’, ‘sm_70’, ‘sm_75’, ‘sm_80’, ‘sm_86’, ‘sm_90’, ‘compute_37’]
when I run print(torch.zeros(1).cuda()), get the output: tensor([0.], device=‘cuda:0’)
error when I run the package of PyroNN:
GPUassert: no kernel image is available for execution on the device C:\Users\sun\AppData\Local\Temp\build-via-sdist-igv1vm0r\pyronn-0.3.1\pyronn\ct_reconstruction\cores\kernels\cone_projector_3D_CudaKernel.cu 252
what should I do to address this problem?
thanks very much!