Custom kernel like matmul can not be dispatched

Some ops are implemented as a composition of other ATen functions like matmul, it consists of expand , view, and bmm in the at::native::matmul. If I implement the kernel of matmul and register it in the following way:


I wonder how to skip the stitching under at::native::matmul to reach my custom kernel directly without modification of the at::native::matmul source code ?