In order to move these functions to torch.linalg we will need to:

• Update the docs: When are this function preferable over a full SVD? Some estimations on the size of the input matrix and its rank would be very helpful to make them more usable.
• Format the docs for them to be in line with the rest of torch.linalg
• Solve Batched SVD_LOWRANK being much slower than loop implementation (both CPU and GPU)  #56891

cc @jianyuh @nikitaved @pearu @mruberry @heitorschueroff @walterddr @IvanYashchuk @xwang233 @lezcano