xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-06-02 18:31:19 +08:00
Code Issues Packages Projects Releases Wiki Activity

Code quality improvements: version update, type annotation enhancement, and enum usage simplification (#27581)

Browse Source 
Signed-off-by: Bradley <bradley.b.pitt@gmail.com>
This commit is contained in:
usberkeley 2025-10-28 01:50:22 +08:00 committed by GitHub
parent 921e78f4bb
commit 69f064062b
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
3 changed files with 7 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
docs/deployment/docker.md
View File

@ -41,11 +41,11 @@ You can add any other [engine-args](../configuration/engine_args.md) you need af
create a custom Dockerfile on top of the base image with an extra layer that installs them: create a custom Dockerfile on top of the base image with an extra layer that installs them:
```Dockerfile ```Dockerfile
FROM vllm/vllm-openai:v0.9.0 FROM vllm/vllm-openai:v0.11.0
# e.g. install the `audio` optional dependencies # e.g. install the `audio` optional dependencies
# NOTE: Make sure the version of vLLM matches the base image! # NOTE: Make sure the version of vLLM matches the base image!
RUN uv pip install --system vllm[audio]==0.9.0 RUN uv pip install --system vllm[audio]==0.11.0
``` ```
!!! tip !!! tip

2
vllm/multimodal/profiling.py
View File

@ -368,7 +368,7 @@ class MultiModalProfiler(Generic[_I]):
self, self,
seq_len: int, seq_len: int,
mm_counts: Mapping[str, int] | None = None, mm_counts: Mapping[str, int] | None = None,
): ) -> Mapping[str, int]:
""" """
Returns the maximum length of the multimodal (image placeholders+text) Returns the maximum length of the multimodal (image placeholders+text)
tokens, including any break/text tokens in-between image embeddings. tokens, including any break/text tokens in-between image embeddings.

10
vllm/v1/core/sched/scheduler.py
View File

@ -113,14 +113,12 @@ class Scheduler(SchedulerInterface):
# req_id -> Request # req_id -> Request
self.requests: dict[str, Request] = {} self.requests: dict[str, Request] = {}
# Scheduling policy # Scheduling policy
if self.scheduler_config.policy == "priority": try:
self.policy = SchedulingPolicy.PRIORITY self.policy = SchedulingPolicy(self.scheduler_config.policy)
elif self.scheduler_config.policy == "fcfs": except ValueError as e:
self.policy = SchedulingPolicy.FCFS
else:
raise ValueError( raise ValueError(
f"Unknown scheduling policy: {self.scheduler_config.policy}" f"Unknown scheduling policy: {self.scheduler_config.policy}"
) ) from e
# Priority queues for requests. # Priority queues for requests.
self.waiting = create_request_queue(self.policy) self.waiting = create_request_queue(self.policy)
self.running: list[Request] = [] self.running: list[Request] = []