Output sample: 0.7010 0.8819 0.8081 0.3136 0.8220 0.6841 0.4798 0.9717 0.2883 [ CUDAFloatType{3,3} ] 0.7010 0.8819 0.8081 0.3136 0.8220 0.6841 0.4798 0.9717 0.2883 [ CUDAFloatType{3,3} ]

References: https://radicalrafi.github.io/posts/pytorch-cpp-intro/ https://discuss.pytorch.org/t/error-pytorch-is-not-linked-with-support-for-cuda-devices/103807

No success w/PRE_TARGETDEPS, the real catch was QMAKE_LFLAGS += -Wl,--no-as-needed Otherwise gets the runtime error PyTorch is not linked with support for cuda devices as bellow when compiling using eigther opts or nograds:

terminate called after throwing an instance of 'c10::Error' what(): PyTorch is not linked with support for cuda devices Exception raised from getDeviceGuardImpl at ../c10/core/impl/DeviceGuardImplInterface.h:224 (most recent call first):