Ways to contribute
There are several ways you can contribute to the project:- Identify and report any issues or bugs
- Request or add support for a new model
- Suggest or implement new features
- Improve documentation or contribute a how-to guide
- Answer queries, offer PR reviews, and assist others
Getting started
Development setup
Set up your local development environment
Testing guide
Learn how to write and run tests
Adding models
Implement new model architectures
Model registration
Register your model with vLLM
Finding tasks to work on
Unsure where to start? Check out these resources:Reporting issues
If you encounter a bug or have a feature request:- Search existing issues first to see if it has already been reported
- If not, file a new issue with as much relevant information as possible
Pull request guidelines
Before submitting a pull request, ensure it meets the following criteria:DCO and signed-off-by
When contributing changes to this project, you must agree to the DCO. Commits must include aSigned-off-by: header which certifies agreement with the terms of the DCO.
Using -s with git commit will automatically add this header:
PR title and classification
The PR title must be prefixed appropriately to indicate the type of change:[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,LLMclass)[Kernel]for changes affecting CUDA kernels or other compute kernels[Core]for changes in the core vLLM logic (e.g.,LLMEngine,AsyncLLMEngine,Scheduler)[Hardware][Vendor]for hardware-specific changes (vendor name should appear in the prefix)[Misc]for PRs that do not fit the above categories (use sparingly)
If the PR spans more than one category, include all relevant prefixes.
Code quality standards
Your PR must meet these code quality standards:- Adhere to Google Python style guide and Google C++ style guide
- Pass all linter checks
- Include well-documented code for future contributors
- Include sufficient tests (both unit tests and integration tests)
- Add documentation to
docs/if the PR modifies user-facing behaviors
Notes for large changes
For major architectural changes (>500 LOC excluding kernel/data/config/test), please create a GitHub issue (RFC) discussing the technical design and justification. PRs without an RFC may be tagged withrfc-required.
Review process
The vLLM team aims to be a transparent reviewing machine. Here’s what you can expect:- After the PR is submitted, it will be assigned to a reviewer based on their expertise and availability
- The reviewer will provide status updates every 2-3 days
- If the PR is not reviewed within 7 days, feel free to ping the reviewer or the vLLM team
- The reviewer will add an
action-requiredlabel if changes are needed - Not all CI checks will be executed due to limited computational resources. The reviewer will add the
readylabel when the PR is ready to merge or a full CI run is needed
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.