[PyTorch] Linear op avoids saving input tensor if weight grad is not needed

[timmoon10]

Description

The linear op caches its input tensor after the forward pass so that it can be used in the backward pass to compute the grad weight tensor. However, if the grad weight tensor is not needed, we can save some memory by not caching.

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

• Linear op avoids saving input tensor if weight grad is not needed

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 L1

[timmoon10]

/te-ci pytorch L1