[PyTorch] Avoid garbage collection when capturing a CUDA Graph

[timmoon10]

Description

This PR avoids a situation where automatic garbage collection destroys a graph while another graph is being captured, which results in a CUDA error.

The bug was introduced with pytorch/pytorch#158193 and pytorch/pytorch#158649. See pytorch/pytorch#161037 for a bugfix and more details.

Type of change

• Documentation change (change only to the documentation, either a fix or a new content)
• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Infra/Build change
• Code refactoring

Changes

• Avoid garbage collection when capturing a CUDA Graph

Checklist:

• I have read and followed the contributing guidelines
• The functionality is complete
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes

[timmoon10]

/te-ci pytorch

[pggPL]

LGTM

[ksivaman]

LGTM