refactor: Turn GPUModelRunner.inputs_embeds to a CpuGpuBuffer #24345
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: Andrew Sansom <[email protected]>
qthequartermasterman
requested review from
WoosukKwon,
alexm-redhat,
comaniac,
njhill,
robertgshaw2-redhat and
ywang96
as code owners
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Code Review
This pull request refactors GPUModelRunner.inputs_embeds to use the CpuGpuBuffer class. This is achieved by extending CpuGpuBuffer to optionally disable the creation of a NumPy view, which is necessary for bfloat16 tensors. The changes are consistent and correctly adapt the usage of inputs_embeds throughout GPUModelRunner. My main feedback is on improving the robustness of the CpuGpuBuffer class to prevent potential runtime errors.
5 tasks
Contributor
Author
DarkLight1337
approved these changes
Sep 6, 2025
github-actions
bot
added
the
ready
eicherseiji
pushed a commit
to eicherseiji/vllm
that referenced
this pull request
Sep 9, 2025
…roject#24345) Signed-off-by: Andrew Sansom <[email protected]>
skyloevil
pushed a commit
to skyloevil/vllm
that referenced
this pull request
Sep 13, 2025
…roject#24345) Signed-off-by: Andrew Sansom <[email protected]>
FeiDaLI
pushed a commit
to FeiDaLI/vllm
that referenced
this pull request
Sep 25, 2025
…roject#24345) Signed-off-by: Andrew Sansom <[email protected]>
xuebwang-amd
pushed a commit
to xuebwang-amd/vllm
that referenced
this pull request
Oct 10, 2025
…roject#24345) Signed-off-by: Andrew Sansom <[email protected]> Signed-off-by: xuebwang-amd <[email protected]>
xuebwang-amd
pushed a commit
to xuebwang-amd/vllm
that referenced
this pull request
Oct 24, 2025
…roject#24345) Signed-off-by: Andrew Sansom <[email protected]> Signed-off-by: xuebwang-amd <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Purpose
As requested by @DarkLight1337 in #24278 (comment), this PR refactors GPUModelRunner.inputs_embeds to use CpuGpuBuffer to reduce the total diff in #24278.
Test Plan
All relevant unit tests are passing locally for me. Pending CI, it should be fine. This change works in the full context of #24278.
Test Result
Pending CI.
Essential Elements of an Effective PR Description Checklist
supported_models.mdandexamplesfor a new model.