Repository navigation
Conversation
3 of 4 tasks
Contributor
|
This pull request has merge conflicts that must be resolved before it can be |
Signed-off-by: Yifan Zong <yzong@redhat.com>
yzong-rh
force-pushed
the
yzong-rh/preempt-stop-resumable
branch
from
October 5, 2026 21:43
983f6db to
938db41
Compare
Contributor
|
This pull request has merge conflicts that must be resolved before it can be |
sawsa307
added a commit
to sawsa307/vllm
that referenced
this pull request
Oct 6, 2026
Adopts the approach from vllm-project#59501: - Don't enqueue in _handle_stopped_request; update_from_output removes stopped requests from their queues and enqueues unfinished ones exactly once, so a request is never in two queues at once. - Fold at num_tokens - 1 directly instead of max(num_computed_tokens, num_tokens - 1); after vllm-project#58668 the two coincide in the normal autoregressive case, and the new assert guards that invariant. - Reject resumable requests for encoder-only and diffusion models, which do not satisfy the invariant. Existing fold tests now append a sampled token so the session state satisfies the fold invariant, and the upstream empty-stream test sets the model-type flags exercised by the new validation. Co-authored-by: Yifan Zong <yzong@redhat.com> Signed-off-by: David Cheung <d7cheung@gmail.com>
Contributor
Author
|
Superceded by #58778 |
sawsa307
added a commit
to sawsa307/vllm
that referenced
this pull request
Oct 15, 2026
Adopts the approach from vllm-project#59501: - Don't enqueue in _handle_stopped_request; update_from_output removes stopped requests from their queues and enqueues unfinished ones exactly once, so a request is never in two queues at once. - Fold at num_tokens - 1 directly instead of max(num_computed_tokens, num_tokens - 1); after vllm-project#58668 the two coincide in the normal autoregressive case, and the new assert guards that invariant. - Reject resumable requests for encoder-only and diffusion models, which do not satisfy the invariant. Existing fold tests now append a sampled token so the session state satisfies the fold invariant, and the upstream empty-stream test sets the model-type flags exercised by the new validation. Co-authored-by: Yifan Zong <yzong@redhat.com> Signed-off-by: David Cheung <d7cheung@gmail.com>
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Overview
My attempt to address the issues #58778 with the following modifications per #58778 (review)
_enqueue_waiting_requestcalls. Remove it from_handle_stopped_requestnum_tokens - 1instead ofmax(num_computed_tokens, num_tokens - 1)Requires #58668 to land first.
Claims
Validation
Details
Pull Request Checklist
I used vLLM's
/pr-checklistskill. (Mandatory for agents, optional for humans).AI assistance was used during the creation of this PR.
Design Fit: Minimizes impact on core components, reuses existing functionality, and justifies added complexity.
Testing and Validation: Validates the change and ensures any added tests are meaningful and reliable, with CI coverage or documented CI resource constraints and validation performed outside CI.
Code Quality and Style: Keeps code and comments clear and concise, and updates relevant documentation and examples.
Pull Request Contents: Includes a brief summary and relevant links, supports claims with evidence, explains root causes and implementation trade-offs, and follows the contributing guide.
BEFORE SUBMITTING, PLEASE READ https://docs.vllm.ai/en/latest/contributing (anything written below this line will be removed by GitHub Actions)