“be8b12684508396987a972eb1188e38a9253b10f”上不存在“paddle/legacy/cuda/include/hl_cpu_scalar.cuh”
* Chage `IndicateDataType` to `GetKernelType`. Make it easier to understand. * Change `OpKernelKey` to `OpKernelType` * Make operator developers can customize which kernel the operator will use in runtime.