phoenix/litellm
1
0
Fork 0
forked from phoenix/litellm-mirror
Code Pull requests Activity

fix(proxy_server.py): router model group alias routing

Browse source 
check model alias group routing before specific deployment routing, to deal with an alias being the same as a deployment name (e.g. gpt-3.5-turbo)

n
This commit is contained in:
Krrish Dholakia 2023-12-30 17:55:24 +05:30
parent 027218c3f0
commit 8ff3bbcfee
1 changed files with 14 additions and 14 deletions
Download patch file Download diff file Expand all files Collapse all files

28
litellm/proxy/proxy_server.py
View file

@ -1130,18 +1130,18 @@ async def completion(
llm_router is not None and data["model"] in router_model_names llm_router is not None and data["model"] in router_model_names
): # model in router model list ): # model in router model list
response = await llm_router.atext_completion(**data) response = await llm_router.atext_completion(**data)
elif (
llm_router is not None and data["model"] in llm_router.deployment_names
): # model in router deployments, calling a specific deployment on the router
response = await llm_router.atext_completion(
**data, specific_deployment=True
)
elif ( elif (
llm_router is not None llm_router is not None
and llm_router.model_group_alias is not None and llm_router.model_group_alias is not None
and data["model"] in llm_router.model_group_alias and data["model"] in llm_router.model_group_alias
): # model set in model_group_alias ): # model set in model_group_alias
response = await llm_router.atext_completion(**data) response = await llm_router.atext_completion(**data)
elif (
llm_router is not None and data["model"] in llm_router.deployment_names
): # model in router deployments, calling a specific deployment on the router
response = await llm_router.atext_completion(
**data, specific_deployment=True
)
else: # router is not set else: # router is not set
response = await litellm.atext_completion(**data) response = await litellm.atext_completion(**data)
@ -1278,16 +1278,16 @@ async def chat_completion(
llm_router is not None and data["model"] in router_model_names llm_router is not None and data["model"] in router_model_names
): # model in router model list ): # model in router model list
response = await llm_router.acompletion(**data) response = await llm_router.acompletion(**data)
elif (
llm_router is not None and data["model"] in llm_router.deployment_names
): # model in router deployments, calling a specific deployment on the router
response = await llm_router.acompletion(**data, specific_deployment=True)
elif ( elif (
llm_router is not None llm_router is not None
and llm_router.model_group_alias is not None and llm_router.model_group_alias is not None
and data["model"] in llm_router.model_group_alias and data["model"] in llm_router.model_group_alias
): # model set in model_group_alias ): # model set in model_group_alias
response = await llm_router.acompletion(**data) response = await llm_router.acompletion(**data)
elif (
llm_router is not None and data["model"] in llm_router.deployment_names
): # model in router deployments, calling a specific deployment on the router
response = await llm_router.acompletion(**data, specific_deployment=True)
else: # router is not set else: # router is not set
response = await litellm.acompletion(**data) response = await litellm.acompletion(**data)
@ -1438,10 +1438,6 @@ async def embeddings(
llm_router is not None and data["model"] in router_model_names llm_router is not None and data["model"] in router_model_names
): # model in router model list ): # model in router model list
response = await llm_router.aembedding(**data) response = await llm_router.aembedding(**data)
elif (
llm_router is not None and data["model"] in llm_router.deployment_names
): # model in router deployments, calling a specific deployment on the router
response = await llm_router.aembedding(**data, specific_deployment=True)
elif ( elif (
llm_router is not None llm_router is not None
and llm_router.model_group_alias is not None and llm_router.model_group_alias is not None
@ -1450,6 +1446,10 @@ async def embeddings(
response = await llm_router.aembedding( response = await llm_router.aembedding(
**data **data
) # ensure this goes the llm_router, router will do the correct alias mapping ) # ensure this goes the llm_router, router will do the correct alias mapping
elif (
llm_router is not None and data["model"] in llm_router.deployment_names
): # model in router deployments, calling a specific deployment on the router
response = await llm_router.aembedding(**data, specific_deployment=True)
else: else:
response = await litellm.aembedding(**data) response = await litellm.aembedding(**data)