[Bugfix][FE]: Always include usage with --enable-force-include-usage (#20983)

Signed-off-by: Max Wittig <max.wittig@siemens.com>
Signed-off-by: Antoine Auger <antoineauger@users.noreply.github.com>
Co-authored-by: Antoine Auger <antoineauger@users.noreply.github.com>
This commit is contained in:
Max Wittig
2025-10-14 09:17:39 +02:00
committed by GitHub
co-authored by Antoine Auger
parent d32c611f45
commit fd85c9f426
11 changed files with 172 additions and 30 deletions
+18 -1
View File
@@ -14,7 +14,11 @@ from starlette.background import BackgroundTask, BackgroundTasks
from vllm.engine.arg_utils import EngineArgs
from vllm.entrypoints.openai.cli_args import make_arg_parser
from vllm.entrypoints.openai.protocol import ChatCompletionRequest, CompletionRequest
from vllm.entrypoints.openai.protocol import (
ChatCompletionRequest,
CompletionRequest,
StreamOptions,
)
from vllm.logger import init_logger
from vllm.platforms import current_platform
from vllm.utils import FlexibleArgumentParser
@@ -237,3 +241,16 @@ def log_non_default_args(args: Namespace | EngineArgs):
)
logger.info("non-default args: %s", non_default_args)
def should_include_usage(
stream_options: StreamOptions | None, enable_force_include_usage: bool
) -> tuple[bool, bool]:
if stream_options:
include_usage = stream_options.include_usage or enable_force_include_usage
include_continuous_usage = include_usage and bool(
stream_options.continuous_usage_stats
)
else:
include_usage, include_continuous_usage = enable_force_include_usage, False
return include_usage, include_continuous_usage