kanshan/TensorRT-LLMs
1
0
Fork 0
mirror of https://github.com/NVIDIA/TensorRT-LLM.git synced 2026-01-14 06:27:45 +08:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity

fix: Align default setting & remove unnecessary check for chat and completion (#3888)

Browse Source 
Signed-off-by: Pengyun Lin <81065165+LinPoly@users.noreply.github.com>
This commit is contained in:
Pengyun Lin 2025-05-07 14:42:53 +08:00 committed by GitHub
parent 4d0e462723
commit 721f84a0ac
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 3 additions and 22 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

25
tensorrt_llm/serve/openai_protocol.py
View File

@ -234,9 +234,8 @@ class CompletionRequest(OpenAIBaseModel):
@model_validator(mode="before")
@classmethod
def check_logprobs(cls, data):
if ("top_logprobs" in data and data.get("top_logprobs")) or \
("logprobs" in data and data.get("logprobs")):
raise ValueError("returning log probs is not supported")
if data.get("logprobs"):
raise ValueError("logprobs is not supported")
return data
@model_validator(mode="before")
@ -270,15 +269,6 @@ class CompletionRequest(OpenAIBaseModel):
raise ValueError("suffix is not supported")
return data
@model_validator(mode="before")
@classmethod
def check_special_tokens(cls, data):
if data.get("skip_special_tokens") or data.get("add_special_tokens") or \
data.get("spaces_between_special_tokens"):
raise ValueError(
"special_tokens related settings are not supported")
return data
class FunctionCall(OpenAIBaseModel):
name: str
@ -425,7 +415,7 @@ class ChatCompletionRequest(OpenAIBaseModel):
stop: Optional[Union[str, List[str]]] = Field(default_factory=list)
stream: Optional[bool] = False
stream_options: Optional[StreamOptions] = None
temperature: Optional[float] = 0.7
temperature: Optional[float] = 1.0
top_p: Optional[float] = 1.0
tools: Optional[List[ChatCompletionToolsParam]] = None
tool_choice: Optional[Union[Literal["none"],
@ -606,15 +596,6 @@ class ChatCompletionRequest(OpenAIBaseModel):
raise ValueError("suffix is not supported")
return data
@model_validator(mode="before")
@classmethod
def check_special_tokens(cls, data):
if data.get("skip_special_tokens") or data.get("add_special_tokens") or \
data.get("spaces_between_special_tokens"):
raise ValueError(
"special_tokens related settings are not supported")
return data
def encode_opaque_state(opaque_state: Optional[bytes]) -> Optional[str]:
if opaque_state is None: