update transformers and vllm dependencies for transformers v5

[bghira]

This pull request updates dependencies and adjusts model loading logic to ensure compatibility with newer versions of core libraries. The most significant changes are the updates to the vllm and transformers packages, and a small refactor to model loading in the caption worker.

Dependency updates:

• Updated the vllm dependency to require version >=0.19.0,<0.20.0 and added a new requirement for transformers version >=5.0.0,<6.0.0 in pyproject.toml to ensure compatibility with the latest APIs.

Model loading logic:

• In src/caption_flow/workers/caption.py, removed the use_fast=True parameter from AutoTokenizer.from_pretrained and added trust_remote_code=True to both tokenizer and processor loading, ensuring both use trusted remote code and aligning with updated library requirements.

[codecov]

Codecov Report

❌ Patch coverage is 0% with 1 line in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/caption_flow/workers/caption.py	0.00%	1 Missing ⚠️

Files with missing lines	Coverage Δ
src/caption_flow/workers/caption.py	61.91% <0.00%> (ø)

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.