Pin PyTorch and supporting package versions

[bandish-shah]

What does this PR do?

User's reported an issue where manually updating packages in our Docker images could lead to unintended PyTorch upgrades if any supporting PyTorch packages (e.g., torchaudo, torchtext, torchvision) are installed without a version specification.

This PR implements a workaround in our Composer PyPi package and Docker images that should pin the Torch version to what we've tested and the Docker images to the specific version we built them for.

What issue(s) does this change relate to?

Resolves CO-1367

Before submitting

• Have you read the contributor guidelines?
• Is this change a documentation change or typo fix? If so, skip the rest of this checklist.
• Was this change discussed/approved in a GitHub issue first? It is much more likely to be merged if so.
• Did you update any related docs and document your change?
• Did you update any related tests and add any new tests related to your change? (see testing)
• Did you run the tests locally to make sure they pass?
• Did you run pre-commit on your change? (see the pre-commit section of prerequisites)

[mvpatel2000]

Assuming you've tested. Can we open a jira for 1.13? Has some important bugfixes