biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[BugFix] get_and_reset only when scheduler outputs are not empty (#6266)

Browse Source
This commit is contained in:
Mor Zusman
2024-07-11 17:40:20 +03:00
committed by GitHub
parent 8a1415cf77
commit 55f692b46e
2 changed files with 4 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
vllm/engine/async_llm_engine.py
View File

@@ -225,11 +225,11 @@ class _AsyncLLMEngine(LLMEngine):
"""
seq_group_metadata_list, scheduler_outputs = self.scheduler[
virtual_engine].schedule()
finished_requests_ids = self.scheduler[
virtual_engine].get_and_reset_finished_requests_ids()
if not scheduler_outputs.is_empty():
# Execute the model.
finished_requests_ids = self.scheduler[
virtual_engine].get_and_reset_finished_requests_ids()
execute_model_req = ExecuteModelRequest(
seq_group_metadata_list=seq_group_metadata_list,
blocks_to_swap_in=scheduler_outputs.blocks_to_swap_in,