mirror of
https://github.com/hiyouga/LLaMA-Factory.git
synced 2025-08-05 21:22:50 +08:00
parent
927ff702ff
commit
772a1b16a1
@ -68,7 +68,11 @@ def create_app(chat_model: ChatModel) -> FastAPI:
|
|||||||
return EventSourceResponse(generate, media_type="text/event-stream")
|
return EventSourceResponse(generate, media_type="text/event-stream")
|
||||||
|
|
||||||
response, (prompt_length, response_length) = chat_model.chat(
|
response, (prompt_length, response_length) = chat_model.chat(
|
||||||
query, history, system, temperature=request.temperature, top_p=request.top_p, max_new_tokens=request.max_tokens
|
query, history, system,
|
||||||
|
do_sample=request.do_sample,
|
||||||
|
temperature=request.temperature,
|
||||||
|
top_p=request.top_p,
|
||||||
|
max_new_tokens=request.max_tokens
|
||||||
)
|
)
|
||||||
|
|
||||||
usage = ChatCompletionResponseUsage(
|
usage = ChatCompletionResponseUsage(
|
||||||
@ -95,7 +99,11 @@ def create_app(chat_model: ChatModel) -> FastAPI:
|
|||||||
yield chunk.json(exclude_unset=True, ensure_ascii=False)
|
yield chunk.json(exclude_unset=True, ensure_ascii=False)
|
||||||
|
|
||||||
for new_text in chat_model.stream_chat(
|
for new_text in chat_model.stream_chat(
|
||||||
query, history, system, temperature=request.temperature, top_p=request.top_p, max_new_tokens=request.max_tokens
|
query, history, system,
|
||||||
|
do_sample=request.do_sample,
|
||||||
|
temperature=request.temperature,
|
||||||
|
top_p=request.top_p,
|
||||||
|
max_new_tokens=request.max_tokens
|
||||||
):
|
):
|
||||||
if len(new_text) == 0:
|
if len(new_text) == 0:
|
||||||
continue
|
continue
|
||||||
|
@ -43,6 +43,7 @@ class DeltaMessage(BaseModel):
|
|||||||
class ChatCompletionRequest(BaseModel):
|
class ChatCompletionRequest(BaseModel):
|
||||||
model: str
|
model: str
|
||||||
messages: List[ChatMessage]
|
messages: List[ChatMessage]
|
||||||
|
do_sample: Optional[bool] = True
|
||||||
temperature: Optional[float] = None
|
temperature: Optional[float] = None
|
||||||
top_p: Optional[float] = None
|
top_p: Optional[float] = None
|
||||||
n: Optional[int] = 1
|
n: Optional[int] = 1
|
||||||
|
Loading…
x
Reference in New Issue
Block a user