xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-13 21:45:25 +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
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 1 additions and 1 deletions
Show Stats 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) parser = EngineArgs.add_cli_args(parser)
# V1 enables prefix caching by default which skews the latency # V1 enables prefix caching by default which skews the latency
# numbers. We need to disable prefix caching by default. # 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): def main(args: argparse.Namespace):