I reinserted my GPU. but then I came into a very strange problem. the code stucks at test.(training successfully for 1 epoch but then it got stuck at testing)
Alexnet is with the error:
Couldn’t open shared file mapping: <torch_14808_1591070686>, error code: <1455> at D:\Downloads\pytorch-master-1\torch\lib\TH\THAllocator.c:154
resnet is with the error of “driver shut down” beside the above.
but the gpu worked well for keras and matlab.
so, what happened?
this is a driver / hardware issue, looks like it. Maybe while using keras/matlab the GPU did not overheat, but using pytorch put it in overheating…