xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-10 19:04:55 +08:00
Code Issues Packages Projects Releases Wiki Activity

[BugFix][Core] Multistep Fix Crash on Request Cancellation (#8059)

Browse Source
This commit is contained in:
Robert Shaw 2024-08-31 15:44:03 -04:00 committed by GitHub
parent 4f5d8446ed
commit 8423aef4c8
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 7 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
vllm/engine/output_processor/multi_step.py
View File

@ -88,9 +88,15 @@ class MultiStepOutputProcessor(SequenceGroupOutputProcessor):
# TODO: Add support for async if necessary # TODO: Add support for async if necessary
assert not is_async assert not is_async
# Sequences can be in RUNNING or FINISHED_ABORTED state
# once scheduled, as a sequence is moved to FINSIHED_ABORTED
# if a client disconnects from the api server.
seqs = sequence_group.get_seqs(status=SequenceStatus.RUNNING) seqs = sequence_group.get_seqs(status=SequenceStatus.RUNNING)
if seqs is None:
seqs = sequence_group.get_seqs(
status=SequenceStatus.FINISHED_ABORTED)
assert seqs, "expected running sequences" assert seqs, "Expected RUNNING or FINISHED_ABORTED sequences"
assert len(seqs) == 1, ( assert len(seqs) == 1, (
"Beam search not supported in multi-step decoding.") "Beam search not supported in multi-step decoding.")
seq = seqs[0] seq = seqs[0]