A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 10, 01:30 UTC
[responsesAPI] allow tuning include_stop_str_in_output (#32383)
Signed-off-by: Andrew Xia <axia@fb.com> Co-authored-by: Andrew Xia <axia@fb.com>
Andrew Xia committed
Jan 16, 2026 at 16:14 UTC
4c82b6fac7700ef544e748aa4c6d1774cebc6ab2
1 file changed
+2
vllm/entrypoints/openai/responses/protocol.py
+2
@@ -172,6 +172,7 @@ class ResponsesRequest(OpenAIBaseModel):
172
truncation: Literal["auto", "disabled"] | None = "disabled"
173
user: str | None = None
174
skip_special_tokens: bool = True
175
+ include_stop_str_in_output: bool = False
176
177
# --8<-- [start:responses-extra-params]
178
request_id: str = Field(
@@ -281,6 +282,7 @@ class ResponsesRequest(OpenAIBaseModel):
282
logit_bias=self.logit_bias,
283
skip_clone=True, # Created fresh per request, safe to skip clone
284
skip_special_tokens=self.skip_special_tokens,
285
+ include_stop_str_in_output=self.include_stop_str_in_output,
286
)
287
288
def is_include_output_logprobs(self) -> bool: