More batched functions such as batched_svd, batched_diagm

In order to write a multiple view geometry and make it easy to used in deep learning, I think the input and output tensor should have a batch dimension. And I need a few batched versions of the linear algebra functions, such as torch.bmm, torch.svd, torch.diag_embd. I traced into the NNlib module and noticed it has batched_mul, batched_transpose/adjoint, but not svd, diagm.

Is this the correct place to add these batched version functions?