simplification of the sincos positional encoding in patchembedding.py

[Lucas-rbnt]

Fixes #7564 .

Description

As discussed, a small simplification for the creation of sincos positional encoding where we don't need to use the torch.no_grad() context or copy the tensor with copy_ from torch which doesn't preserve the requires_grad attribute here.

The changes are simple and are linked to the corresponding comment #7564, the output is already in float32 so it doesn't seem particularly necessary to apply the conversion previously done.

Types of changes

• Non-breaking change (fix or new feature that would not break existing functionality).
• Breaking change (fix or new feature that would cause existing functionality to change).
• New tests added to cover the changes.
• Integration tests passed locally by running ./runtests.sh -f -u --net --coverage.
• Quick tests passed locally by running ./runtests.sh --quick --unittests --disttests.
• In-line docstrings updated.
• Documentation updated, tested make html command in the docs/ folder.

[marksgraham]

LGTM

[KumoLiu]

/build

[KumoLiu]

/build