pytorch:0.4.1
cuda:9.2
解决方案:
添加这行就行:
torch.backends.cudnn.benchmark = True
引用自:
https://stackoverflow.com/questions/62752522/runtimeerror-cudnn-error-cudnn-status-success
pytorch:0.4.1
cuda:9.2
解决方案:
添加这行就行:
torch.backends.cudnn.benchmark = True
引用自:
https://stackoverflow.com/questions/62752522/runtimeerror-cudnn-error-cudnn-status-success