apunkt/plano
1
0
Fork 0
mirror of https://github.com/katanemo/plano.git synced 2026-06-17 15:25:17 +02:00
Code Issues Projects Releases Packages Wiki Activity

merge main

Browse source
This commit is contained in:
cotran 2024-10-31 09:53:15 -07:00
commit 2e71c326c8
43 changed files with 1508 additions and 607 deletions
Download patch file Download diff file Expand all files Collapse all files

16
model_server/app/function_calling/model_utils.py
View file

@ -75,12 +75,16 @@ async def chat_completion(req: ChatMessage, res: Response):
f"model_server => arch_function: {client_model_name}, messages: {json.dumps(messages)}"
)
resp = const.arch_function_client.chat.completions.create(
messages=messages,
model=client_model_name,
stream=True,
extra_body=const.arch_function_generation_params,
)
try:
resp = const.arch_function_client.chat.completions.create(
messages=messages,
model=client_model_name,
stream=False,
extra_body=const.arch_function_generation_params,
)
except Exception as e:
logger.error(f"model_server <= arch_function: error: {e}")
raise
# Retrieve the first token, handling the Stream object carefully
first_token_content = ""