[PyTorch] Add ops for dropout and constant scale

[timmoon10]

Description

This PR adds fusible ops for dropout (heavily based on torch.nn.Dropout) and multiplying by a constant scalar. These are not performant, but they provide an API for future custom implementations or fusions.

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

• Add fusible ops for dropout and constant scale

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

[negvet]

LGTM

[negvet]

Handing this case similar to torch.nn.Dropout:

Suggested change

	keep_prob = 1 - self.dropout_probability
	if self.dropout_probability == 1:
	mask = torch.zeros_like(input_)
	out = mask
	else:
	keep_prob = 1 - self.dropout_probability

[timmoon10]

The existing impl should handle this case correctly. We will also replace this mask-based impl soon, so no need to optimize aggressively.