Move from torch.cuda.amp to torch.autocast; Add tests for amp

[misko]

This PR updates torch.cuda.amp.autocast(args...) and torch.cuda.amp.GradScaler(args...) (deprecated) and also adds CPU AMP tests. ( https://pytorch.org/docs/stable/amp.html )
Recently there was an eSCN model that did not run on GPU AMP and the current set of tests did not catch it. This PR adds those equivalent tests on CPU AMP, which will test this going forward.
Additional fixes haven been made to eSCN / SCN / Gemnet OC in this PR.

[lbluque]

Thanks @misko, mostly style suggestions

[codecov]

Codecov Report

Attention: Patch coverage is 92.85714% with 2 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
...e/models/equiformer_v2/equiformer_v2_deprecated.py	0.00%	1 Missing ⚠️
src/fairchem/core/trainers/base_trainer.py	83.33%	1 Missing ⚠️

Files with missing lines	Coverage Δ
...rc/fairchem/applications/cattsunami/core/ocpneb.py	87.12% <100.00%> (+0.09%)	⬆️
src/fairchem/core/common/relaxation/ase_utils.py	63.15% <100.00%> (+1.20%)	⬆️
...c/fairchem/core/models/equiformer_v2/layer_norm.py	58.62% <100.00%> (+0.72%)	⬆️
src/fairchem/core/models/escn/escn.py	95.65% <100.00%> (ø)
src/fairchem/core/models/gemnet_oc/gemnet_oc.py	89.68% <100.00%> (-0.23%)	⬇️
src/fairchem/core/models/scn/scn.py	93.56% <100.00%> (ø)
src/fairchem/core/modules/scaling/fit.py	73.80% <100.00%> (ø)
src/fairchem/core/trainers/ocp_trainer.py	69.12% <100.00%> (+0.10%)	⬆️
...e/models/equiformer_v2/equiformer_v2_deprecated.py	91.15% <0.00%> (ø)
src/fairchem/core/trainers/base_trainer.py	87.52% <83.33%> (-0.40%)	⬇️

... and 2 files with indirect coverage changes

[lbluque]

lg, just minor suggestions!

[lbluque]

lgtm, thanks @misko !