Conv3d, gpu,cuda

saran · July 23, 2021, 2:31am

i am facing this error when i am runing my infer.py

RuntimeError: Could not run ‘aten::slow_conv3d_forward’ with arguments from the ‘CUDA’ backend. ‘aten::slow_conv3d_forward’ is only available for these backends: [CPU, Autograd, Profiler, Tracer].

ptrblck · July 23, 2021, 5:36am

I guess this error is raised, if you’ve installed a CPU-only libtorch library, so you would need to use the one with the CUDA bindings.