phoenix/litellm-mirror
0
0
Fork 1
mirror of https://github.com/BerriAI/litellm.git synced 2025-04-26 11:14:04 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Litellm dev 12 25 2024 p1 (#7411)

Browse source 
* test(test_watsonx.py): e2e unit test for watsonx custom header

covers https://github.com/BerriAI/litellm/issues/7408

* fix(common_utils.py): handle auth token already present in headers (watsonx + openai-like base handler)

Fixes https://github.com/BerriAI/litellm/issues/7408

* fix(watsonx/chat): fix chat route

Fixes https://github.com/BerriAI/litellm/issues/7408

* fix(huggingface/chat/handler.py): fix huggingface async completion calls

* Correct handling of max_retries=0 to disable AzureOpenAI retries (#7379)

* test: fix test

---------

Co-authored-by: Minh Duc <phamminhduc0711@gmail.com>
This commit is contained in:
Krish Dholakia 2024-12-25 17:36:30 -08:00 • committed by GitHub
parent 55b5a1221c
commit b961f96b35
9 changed files with 299 additions and 14 deletions
Download patch file Download diff file Expand all files Collapse all files

3
litellm/llms/azure/azure.py
View file

@ -855,7 +855,8 @@ class AzureChatCompletion(BaseLLM):
self._client_session = self.create_client_session()
try:
data = {"model": model, "input": input, **optional_params}
max_retries = max_retries or litellm.DEFAULT_MAX_RETRIES
if max_retries is None:
max_retries = litellm.DEFAULT_MAX_RETRIES
if not isinstance(max_retries, int):
raise AzureOpenAIError(
status_code=422, message="max retries must be an int"