frozenleaves/vllm
1
0
Fork 0
mirror of https://github.com/vllm-project/vllm.git synced 2025-10-20 14:53:52 +08:00
Code Issues Packages Projects Releases Wiki Activity

[Bugfix] Disable prefix caching by default for benchmark (#18771)

Browse Source 
Signed-off-by: cascade812 <cascade812@outlook.com>
This commit is contained in:
cascade
2025-05-27 17:18:09 -07:00
committed by GitHub
parent e56f44d9ec
commit 51e98e4ffd
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
vllm/benchmarks/latency.py
View File

@ -82,7 +82,7 @@ def add_cli_args(parser: argparse.ArgumentParser):
parser = EngineArgs.add_cli_args(parser)
# V1 enables prefix caching by default which skews the latency
# numbers. We need to disable prefix caching by default.
parser.set_defaults(enable_prefix_caching=True)
parser.set_defaults(enable_prefix_caching=False)
def main(args: argparse.Namespace):