Skip to content

[BugFix] Fix PP/async scheduling with pooling models #28899

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.

Sign up for GitHub

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

Merged
merged 2 commits into from
Nov 18, 2025
Merged

[BugFix] Fix PP/async scheduling with pooling models #28899

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
fcfcf6f
[BugFix] Fix PP/async scheduling with pooling models
njhill Nov 18, 2025
3b55e5c
Merge branch 'main' into fix-embedding-pp
DarkLight1337 Nov 18, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion vllm/v1/engine/core.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -184,6 +184,7 @@ def __init__(
vllm_config.ec_transfer_config is not None
and vllm_config.ec_transfer_config.is_ec_producer
)
self.is_pooling_model = vllm_config.model_config.runner_type == "pooling"

self.request_block_hasher: Callable[[Request], list[BlockHash]] | None = None
if vllm_config.cache_config.enable_prefix_caching or kv_connector is not None:
Expand Down Expand Up @@ -392,7 +393,7 @@ def step_with_batch_queue(
if not self.ec_producer:
model_executed = scheduler_output.total_num_scheduled_tokens > 0

if not model_executed:
if self.is_pooling_model or not model_executed:
# No sampling required (no requests scheduled).
future = cast(Future[ModelRunnerOutput], exec_future)
else:
Expand Down
8 changes: 4 additions & 4 deletions vllm/v1/executor/ray_executor.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -99,9 +99,9 @@ def _init_executor(self) -> None:
# KV connector setup
self.has_connector = self.vllm_config.kv_transfer_config is not None

self.ec_producer = (
self.vllm_config.ec_transfer_config is not None
and self.vllm_config.ec_transfer_config.is_ec_producer
self.uses_sampler = self.vllm_config.model_config.runner_type != "pooling" and (
self.vllm_config.ec_transfer_config is None
or not self.vllm_config.ec_transfer_config.is_ec_producer
)

self.scheduler_output: SchedulerOutput | None = None
Expand Down Expand Up @@ -401,7 +401,7 @@ def execute_model( # type: ignore[override]
"after execute_model() returns None."
)

if self.ec_producer or not scheduler_output.total_num_scheduled_tokens:
if not self.uses_sampler or not scheduler_output.total_num_scheduled_tokens:
# Model will not execute, call model runner immediately.
return self._execute_dag(scheduler_output, None, non_block)

Expand Down