The reason why arise this error is that my torch::Tensor is in c10::kCUDA, but i want copy its content to c10::CPU’s std::vector. So it cause SIGSEGV error.
The reason why arise this error is that my torch::Tensor is in c10::kCUDA, but i want copy its content to c10::CPU’s std::vector. So it cause SIGSEGV error.