[Model] H2O Danube3-4b #6451
[Model] H2O Danube3-4b #6451
Conversation
|
👋 Hi! Thank you for contributing to the vLLM project. Full CI run is still required to merge this PR so once the PR is ready to go, please make sure to run it. If you need all test signals in between PR commits, you can trigger full CI as well. To run full CI, you can do one of these:
🚀 |
|
/ready |
github-actions
bot
added
the
ready
|
If ready should you mark as not draft? |
|
@pseudotensor it is not ready. That comment triggers the full test suite. |
fp8 kv cache values are encoded as uint8. The element size of uint8 is 1. 16 divided by any int over 1 is going to be 8 or less which is compatible with head size 120. But that doesn't happen with fp8, which leads to test failures for head size 120.
|
The |
|
@comaniac thank you for the quick response. All tests are passing now. |
[Model] H2O Danube3-4b (vllm-project#6451)
Signed-off-by: Alvant <[email protected]>
This PR mainly focuses on adding a head size of 120 for GPU inference, to support h2oai/h2o-danube3-4b-base.
Head sizes that are not a multiple of 16 aren't compatible with
fp8kv cache, so those tests are skipped.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,LLMclass, 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:
format.shto format your code.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-requiredand 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:
action-requiredlabel on the PR if there are changes required. The contributor should address the comments and ping the reviewer to re-review the PR.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!