[CICD]Enhance project's compatibility

[poinwater]

Description

Brief Description of the PR:
This pull request aims to enhance the project's compatibility by making it compatible with TensorFlow versions 2.6.0 to 2.11.0.

Type of change

1. Added compatible import statements for Adam, Version, and data_structures.
2. Updated the building scripts and Docker files.
3. Decoupled TensorFlow dependencies so that users don't need to reinstall TensorFlow with the TFRA wheels.
4. Added CI/CD tasks to build wheels with both the minimum and maximum TensorFlow versions.

• Bug fix
• CICD
• New Tutorial
• Updated or additional documentation
• Additional Testing
• New Feature

Checklist:

• I've properly formatted my code according to the guidelines
  • By running yapf
  • By running clang-format
• This PR addresses an already submitted issue for TensorFlow Recommenders-Addons
• I have made corresponding changes to the documentation
• I have added tests that prove my fix is effective or that my feature works

How Has This Been Tested?

If you're adding a bugfix or new feature, please describe the tests that you ran to verify your changes:

I tested the CI/CD with multiple TensorFlow versions in my repository. The code is working with TensorFlow versions between 2.6.0 to 2.11.0. However, for simplifying the pipeline, I only kept the minimum tasks.

[poinwater]

Not sure why it's closed

[MoFHeka]

LGTM

[rhdong]

Hi @poinwater, please use rebase on master instead of merge master, thank you~