[Misc]feat: adapt to vLLM main (5af684c3...66d1cc0c) - #235
Draft
Meihan-chen wants to merge 25 commits into
Draft
Conversation
Signed-off-by: Meihan-chen <jcccx.cmh@gmail.com>
Signed-off-by: Meihan-chen <jcccx.cmh@gmail.com>
Signed-off-by: Meihan-chen <jcccx.cmh@gmail.com>
Signed-off-by: Meihan-chen <jcccx.cmh@gmail.com>
Signed-off-by: Meihan-chen <jcccx.cmh@gmail.com>
Signed-off-by: Meihan-chen <jcccx.cmh@gmail.com>
Signed-off-by: Meihan-chen <jcccx.cmh@gmail.com>
Signed-off-by: Meihan-chen <jcccx.cmh@gmail.com>
Signed-off-by: Meihan-chen <jcccx.cmh@gmail.com>
Signed-off-by: Meihan-chen <239345480+Meihan-chen@users.noreply.github.com>
Upstream vLLM merged the kv_offload module structure: - vllm/v1/kv_offload/abstract.py (deleted) -> vllm/v1/kv_offload/base.py - vllm/v1/kv_offload/mediums.py (deleted) -> vllm/v1/kv_offload/base.py - vllm/v1/kv_offload/spec.py (deleted) -> vllm/v1/kv_offload/base.py This commit adapts vllm-ascend to use the new unified base.py module and updates the API to work with the new CanonicalKVCaches format. Key changes: - Update imports from deleted modules to use vllm.v1.kv_offload.base - Update get_handlers() signature to accept CanonicalKVCaches - Update CpuNpuOffloadingHandler to work with CanonicalKVCaches - Remove unused attn_backends parameter Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> Signed-off-by: main2main-bot <main2main-bot@users.noreply.github.com>
Upstream vLLM renamed and reorganized the MoE runner classes: - vllm.model_executor.layers.fused_moe.runner.default_moe_runner (deleted) - DefaultMoERunner class renamed to MoERunner - MoERunnerBase renamed to MoERunnerInterface This commit updates vllm-ascend to use the new module path and class name: - Change import from default_moe_runner to moe_runner - Change DefaultMoERunner to MoERunner Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> Signed-off-by: main2main-bot <main2main-bot@users.noreply.github.com>
Upstream vLLM commit c08f3b2a6 introduced a change to measure encoder compile time separately from the LLM backbone. This changed the return type of compile_or_warm_up_model() from float to CompilationTimes, a NamedTuple containing (language_model, encoder) compilation times. This commit adapts vllm-ascend's NPUWorker to match the new upstream API: - Import CompilationTimes from vllm.v1.worker.worker_base - Change return type annotation from float to CompilationTimes - Return CompilationTimes with both language_model and encoder times Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> Signed-off-by: main2main-bot <main2main-bot@users.noreply.github.com>
Meihan-chen
force-pushed
the
main
branch
9 times, most recently
from
May 20, 2026 03:18
41aae44 to
9af7666
Compare
zhangxinyuehfad
force-pushed
the
main
branch
3 times, most recently
from
June 9, 2026 06:11
e4d6423 to
9f88e4b
Compare
MrZ20
force-pushed
the
main
branch
2 times, most recently
from
June 10, 2026 03:25
677396a to
9554149
Compare
zhangxinyuehfad
force-pushed
the
main
branch
5 times, most recently
from
June 16, 2026 04:48
c14c88e to
9768184
Compare
zhangxinyuehfad
force-pushed
the
main
branch
3 times, most recently
from
June 17, 2026 12:22
24629f9 to
b5ed5cc
Compare
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
2 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.
Automated adaptation to upstream vLLM main branch changes.
Commit range: 5af684c31912232e5c89484c2e8259e0fac6c55b...66d1cc0c77628e36df8e5e245c116e6aac3045fb
phase: detect
round: 0
git_commits:
sha:
0204395cf25b0fa0c4fc7802d6dd6d8e240306besubject: fix(kv_offload): adapt to upstream kv_offload module restructuring
body:
Upstream vLLM merged the kv_offload module structure:
This commit adapts vllm-ascend to use the new unified base.py module
and updates the API to work with the new CanonicalKVCaches format.
Key changes:
Co-Authored-By: Claude Opus 4.7 noreply@anthropic.com
Signed-off-by: main2main-bot main2main-bot@users.noreply.github.com
vLLM version: v0.19.0
vLLM main: vllm-project/vllm@5af684c