[Model] support MiniMax-VL-01 model

[qscqesze]

CLOSES #12073

This PR is intended to support the MiniMaxVL01 model inference.
It adopts the “ViT-MLP-LLM” framework, which is a commonly used technique in the field of multimodal large language models. Add the MiniMaxVL01 model inference implementation, and a image processer.

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.

💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels.

Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your fastcheck build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping simon-mo or khluu to add you in our Buildkite org.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can either: Add ready label to the PR or enable auto-merge.

🚀

[mergify]

This pull request has merge conflicts that must be resolved before it can be
merged. Please rebase the PR, @qscqesze.

https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/working-with-forks/syncing-a-fork

[DarkLight1337]

Why can't we test processing correctness for this model?

[DarkLight1337]

Can you also add a get_language_model method to match the new interface since #16007?

[qscqesze]

Can you also add a get_language_model method to match the new interface since #16007?

Done.

[qscqesze]

@DarkLight1337 Hi. If you have time, could you help us move this forward?

[DarkLight1337]

Nice, it passes now. I'll take care of the processor test in another PR