-
Notifications
You must be signed in to change notification settings - Fork 6.3k
[Data][LLM] Bump vLLM version to support new models #51726
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Can we also run release test for Ray Serve LLM? just to ensure that's also compatible with the new version.
Can we run serve llm release tests and see if they fail with the upgrade? This release tests only covered the batch ones. |
seems serve tests are failing. I'm turning it to draft for now. |
Note that 0.8.2 will use v1 by default, so you may need to combine with #51490, or explicitly add VLLM_USE_V1=0 in this PR first. |
f2810c7
to
f743e6b
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
let me know when this is ready for review (or if you have any questions)
python/ray/llm/_internal/serve/deployments/llm/vllm/vllm_engine.py
Outdated
Show resolved
Hide resolved
8691592
to
ebc4b15
Compare
Signed-off-by: Linkun Chen <github@lkchen.net>
115f395
to
4b97d68
Compare
Signed-off-by: Linkun Chen <github@lkchen.net>
Signed-off-by: Linkun Chen <github@lkchen.net>
281b4ac
to
5a9f079
Compare
Signed-off-by: Linkun Chen <github@lkchen.net>
Signed-off-by: Linkun Chen <github@lkchen.net>
Signed-off-by: Linkun Chen <github@lkchen.net>
Signed-off-by: Linkun Chen <github@lkchen.net>
0f87863
to
a5d74ea
Compare
Signed-off-by: Linkun Chen <github@lkchen.net>
Signed-off-by: Linkun Chen <github@lkchen.net>
a5d74ea
to
152cb54
Compare
Signed-off-by: Linkun Chen <github@lkchen.net>
kicked off the release tests: https://buildkite.com/ray-project/release/builds/37586 |
d70a554
to
04592da
Compare
Signed-off-by: Linkun Chen <github@lkchen.net>
04592da
to
a9351b3
Compare
Signed-off-by: Linkun Chen <github@lkchen.net>
Signed-off-by: Linkun Chen <github@lkchen.net>
Why are these changes needed?
Vision models (e.g. pixtral) are only supported in vLLM release (0.8.2), this PR bumps vLLM requirements and adds release tests for new models.
Related issue number
#51313
Checks
git commit -s
) in this PR.scripts/format.sh
to lint the changes in this PR.method in Tune, I've added it in
doc/source/tune/api/
under thecorresponding
.rst
file.