Make T5 models interchangeable between FLUX/SD3 workflows

[RyanJDick]

Summary

Prior to this change, attempting to use a T5 model that was bundled with an SD3 base model in a FLUX workflow resulted in the following exception.

[2025-01-09 22:58:53,711]::[InvokeAI]::ERROR --> Error while invoking session 3f4f9287-bb23-44f7-8e52-bbf6c4b114e8, invocation c3bd3f6d-71cd-4b86-a4c2-8e43758e4185 (flux_text_encoder): 
[2025-01-09 22:58:53,711]::[InvokeAI]::ERROR --> Traceback (most recent call last):
  File "/home/ryan/src/InvokeAI/invokeai/app/services/session_processor/session_processor_default.py", line 129, in run_node
    output = invocation.invoke_internal(context=context, services=self._services)
  File "/home/ryan/src/InvokeAI/invokeai/app/invocations/baseinvocation.py", line 300, in invoke_internal
    output = self.invoke(context)
  File "/home/ryan/.pyenv/versions/3.10.14/envs/InvokeAI_3.10.14/lib/python3.10/site-packages/torch/utils/_contextlib.py", line 116, in decorate_context
    return func(*args, **kwargs)
  File "/home/ryan/src/InvokeAI/invokeai/app/invocations/flux_text_encoder.py", line 60, in invoke
    t5_embeddings = self._t5_encode(context)
  File "/home/ryan/src/InvokeAI/invokeai/app/invocations/flux_text_encoder.py", line 78, in _t5_encode
    assert isinstance(t5_text_encoder, T5EncoderModel)
AssertionError

After this change, all T5 models can be used interchangeably for both FLUX and SD3 workflows.

Related Issues / Discussions

• Closes [bug]: Assertion error related to text_encoder while trying to generate with Flux Dev Quantized  #7370

QA Instructions

I tested the following combinations:

• FLUX text-to-image, standalone T5
• FLUX text-to-image, SD3-bundled T5
• SD3 text-to-image, SD3-bundled T5
• SD3 text-to-image, standalone T5
• SD3 text-to-image, BnB-quantized T5

Merge Plan

No special instructions.

Checklist

• The PR has a short but descriptive title, suitable for a changelog
• Tests added / updated (if applicable)
• Documentation added / updated (if applicable)
• Updated What's New copy (if doing a release after this PR)