Back out "Back out "[const_fold] Set requires_grad based on the folded tensor; add device_for_folding option"" (#79696)
Summary:
This is an un-backout but with a small change to set the default device `device_for_folded_attrs="cuda"` instead of `"cpu"`, which should avoid BC issues for TRT lowering.
Original commit changeset: 4ae1863e28ff
Original Phabricator Diff: D37192230 (https://github.com/pytorch/pytorch/commit/24c2aff1b23729b5ded089ffbc81ac3239fc47bd)
Test Plan: Added unit test
Differential Revision: D37205432
Pull Request resolved: https://github.com/pytorch/pytorch/pull/79696
Approved by: https://github.com/dborkovic
