Reduce HF calls during pytest execution

[jlarson4]

Description

The current TransformerBridge test suite runs hundreds of HF API calls. This PR is an effort to reduce the # of calls per-CI run, hopefully speeding up our overall CI feedback loop for v3.

The main focus of this PR is to scope model loading to a per-session operation, rather than a per test, so we don't have a multitude of tests reloading the same models. For operations that require modification of the model in place, I am deepcopying the original request model, and then making modifications to the copy. Any other operations run on the session model directly.

Type of change

• Bug fix (non-breaking change which fixes an issue)

Checklist:

• 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
• I have not rewritten tests relating to key interfaces which would affect backward compatibility