Hi,
when addmv_impl_cpu is being used in cpu and is it faster than addmm_impl_cpu interms of performance if the addmm shapes are small?
The impl is available here (pytorch/aten/src/ATen/native/Blas.cpp at 03a05e791aabe4f2e3938626bff849641dc101ea · pytorch/pytorch · GitHub)
but I am not able to find its usage.
Thanks