[VLM] Refactor MultiModalConfig initialization and profiling

[ywang96]

Previously we initialize MultiModalConfig for all models regardless of whether or not the model itself is multi-modal. This can cause some developer confusion, especially since we're going to use this class for multimodal models starting #7126.

This PR refactors the logic and make MultiModalConfig an attribute of ModelConfig to make the separation cleaner.

Also, the common call of MultiModalRegistry.init_mm_limits_per_prompt has been moved from profile_run to ModelRunner.__init__.

---

PR Checklist (Click to Expand)

Thank you for your contribution to vLLM! Before submitting the pull request, please ensure the PR meets the following criteria. This helps vLLM maintain the code quality and improve the efficiency of the review process.

PR Title and Classification

Only specific types of PRs will be reviewed. The PR title is prefixed appropriately to indicate the type of change. Please use one of the following:

• [Bugfix] for bug fixes.
• [CI/Build] for build or continuous integration improvements.
• [Doc] for documentation fixes and improvements.
• [Model] for adding a new model or improving an existing model. Model name should appear in the title.
• [Frontend] For changes on the vLLM frontend (e.g., OpenAI API server, LLM class, etc.)
• [Kernel] for changes affecting CUDA kernels or other compute kernels.
• [Core] for changes in the core vLLM logic (e.g., LLMEngine, AsyncLLMEngine, Scheduler, etc.)
• [Hardware][Vendor] for hardware-specific changes. Vendor name should appear in the prefix (e.g., [Hardware][AMD]).
• [Misc] for PRs that do not fit the above categories. Please use this sparingly.

Note: If the PR spans more than one category, please include all relevant prefixes.

Code Quality

The PR need to meet the following code quality standards:

• We adhere to Google Python style guide and Google C++ style guide.
• Pass all linter checks. Please use format.sh to format your code.
• The code need to be well-documented to ensure future contributors can easily understand the code.
• Include sufficient tests to ensure the project to stay correct and robust. This includes both unit tests and integration tests.
• Please add documentation to docs/source/ if the PR modifies the user-facing behaviors of vLLM. It helps vLLM user understand and utilize the new features or changes.

Notes for Large Changes

Please keep the changes as concise as possible. For major architectural changes (>500 LOC excluding kernel/data/config/test), we would expect a GitHub issue (RFC) discussing the technical design and justification. Otherwise, we will tag it with rfc-required and might not go through the PR.

What to Expect for the Reviews

The goal of the vLLM team is to be a transparent reviewing machine. We would like to make the review process transparent and efficient and make sure no contributor feel confused or frustrated. However, the vLLM team is small, so we need to prioritize some PRs over others. Here is what you can expect from the review process:

• After the PR is submitted, the PR will be assigned to a reviewer. Every reviewer will pick up the PRs based on their expertise and availability.
• After the PR is assigned, the reviewer will provide status update every 2-3 days. If the PR is not reviewed within 7 days, please feel free to ping the reviewer or the vLLM team.
• After the review, the reviewer will put an action-required label on the PR if there are changes required. The contributor should address the comments and ping the reviewer to re-review the PR.
• Please respond to all comments within a reasonable time frame. If a comment isn't clear or you disagree with a suggestion, feel free to ask for clarification or discuss the suggestion.

Thank You

Finally, thank you for taking the time to read these guidelines and for your interest in contributing to vLLM. Your contributions make vLLM a great tool for everyone!

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.
Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which consists a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of default ones by unblocking the steps in your fast-check build on Buildkite UI.

Once the PR is approved and ready to go, please make sure to run full CI as it is required to merge (or just use auto-merge).

To run full CI, you can do one of these:

• Comment /ready on the PR
• Add ready label to the PR
• Enable auto-merge.

🚀

[ywang96]

@DarkLight1337 I realized this won't work in the way I wanted since currently there isn't a clean way to check at engine init time whether a model supports multi-modal. Going to revert my changes from commit 814c2bc and just change raise to assert.

[DarkLight1337]

@DarkLight1337 I realized this won't work in the way I wanted since currently there isn't a clean way to check at engine init time whether a model supports multi-modal. Going to revert my changes from commit 814c2bc and just change raise to assert.

I think it should be possible to add a new function ModelRegistry.is_multimodal_model, similar to ModelRegistry.is_embedding_model. Then we can initialize MultiModalConfig inside ModelConfig only when the model is indeed multi-modal.

[ywang96]

@DarkLight1337 I realized this won't work in the way I wanted since currently there isn't a clean way to check at engine init time whether a model supports multi-modal. Going to revert my changes from commit 814c2bc and just change raise to assert.

I think it should be possible to add a new function ModelRegistry.is_multimodal_model, similar to ModelRegistry.is_embedding_model. Then we can initialize MultiModalConfig inside ModelConfig only when the model is indeed multi-modal.

Yea the more I think about it, the more I feel like it makes more sense to have a ModelRegistry.is_multimodal_model function for us to easily differentiate the models. Will update my code accordingly.

[DarkLight1337]

Since MultiModalConfig is contained in ModelConfig, it is redundant to pass MultiModalConfig separately from ModelConfig to the various classes (executor, model runner, model loader and model classes).

[DarkLight1337]

LGTM, can merge after you have finished manual testing.

[DarkLight1337]

@youkaichao is there a good way to avoid initializing CUDA prematurely here? We shouldn't rely on not importing modules that initialize CUDA inside the model files.

[DarkLight1337]

I can't think of anything outside of performing this check in a separate process.

[ywang96]

Will use a hardcoded class of multimodal models as a workaround for now