Add flag to get attention from fusion

[ericharper]

What does this PR do ?

Currently GPT gets the attention_mask from the fused softmax kernel.

If, the kernel isn't used, then the fallback to pytorch softmax breaks.

This PR makes it configurable to get attention_mask from the kernel.

Collection: NLP

Changelog

• Update forward and loss func

Usage

• You can potentially add a usage example below

# Add a code snippet demonstrating how to use this 

Before your PR is "Ready for review"

Pre checks:

• Make sure you read and followed Contributor guidelines
• Did you write any new necessary tests?
• Did you add or update any necessary documentation?
• Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
  • Reviewer: Does the PR have correct import guards for all optional libraries?

PR Type:

• New Feature
• Bugfix
• Documentation

If you haven't finished some of the above items you can still open "Draft" PR.

Who can review?

Anyone in the community is free to review the PR once the checks have passed.
Contributor guidelines contains specific people who can review PRs to various areas.

Additional Information

• Related to # (issue)

[MaximumEntropy]

This looks good to me. @erhoo82 to confirm if you are okay with the default value of the new arg as False.

[github-actions]

This PR is stale because it has been open for 14 days with no activity. Remove stale label or comment or update or this will be closed in 7 days.