add large model stable training fix and contrastive loss update for variable seq

[nithinraok]

What does this PR do ?

This PR brings two updates:

• Fix SSL contrastive loss to support loss computation for variable input length rather than computation based on subsampling factor
• Add an option to remove bias from Linear and Conv layers in Conformer layers to support scaling for multi billion parameter training [tested currently for 1B parameter using bf16]

Collection: ASR

Changelog

• Updated ssl_loss
• Added argument for conformerlayer to optionally remove bias, default is False

GitHub Actions CI

The Jenkins CI system has been replaced by GitHub Actions self-hosted runners.

The GitHub Actions CI will run automatically when the "Run CICD" label is added to the PR. To re-run CI remove and add the label again. To run CI on an untrusted fork, a NeMo user with write access must first click "Approve and run".

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:

• [x] New Feature
• [x] Bugfix
• [ ] Documentation