fix(main.py): support max retries for transcription calls

litellm/llms/azure.py

@@ -799,6 +799,7 @@ class AzureChatCompletion(BaseLLM):
         optional_params: dict,
         model_response: TranscriptionResponse,
         timeout: float,
+        max_retries: int,
         api_key: Optional[str] = None,
         api_base: Optional[str] = None,
         api_version: Optional[str] = None,
@@ -817,8 +818,6 @@ class AzureChatCompletion(BaseLLM):
             "timeout": timeout,
         }
 
-        max_retries = optional_params.pop("max_retries", None)
-
         azure_client_params = select_azure_base_url_or_endpoint(
             azure_client_params=azure_client_params
         )

litellm/llms/openai.py

@@ -785,10 +785,10 @@ class OpenAIChatCompletion(BaseLLM):
         optional_params: dict,
         model_response: TranscriptionResponse,
         timeout: float,
+        max_retries: int,
         api_key: Optional[str] = None,
         api_base: Optional[str] = None,
         client=None,
-        max_retries=None,
         logging_obj=None,
         atranscription: bool = False,
     ):

litellm/main.py

@@ -3553,6 +3553,7 @@ def transcription(
     api_key: Optional[str] = None,
     api_base: Optional[str] = None,
     api_version: Optional[str] = None,
+    max_retries: Optional[int] = None,
     litellm_logging_obj=None,
     custom_llm_provider=None,
     **kwargs,
@@ -3568,6 +3569,8 @@ def transcription(
     proxy_server_request = kwargs.get("proxy_server_request", None)
     model_info = kwargs.get("model_info", None)
     metadata = kwargs.get("metadata", {})
+    if max_retries is None:
+        max_retries = openai.DEFAULT_MAX_RETRIES
 
     model_response = litellm.utils.TranscriptionResponse()
 
@@ -3611,6 +3614,7 @@ def transcription(
             api_key=api_key,
             api_version=api_version,
             azure_ad_token=azure_ad_token,
+            max_retries=max_retries,
         )
     elif custom_llm_provider == "openai":
         response = openai_chat_completions.audio_transcriptions(
@@ -3621,6 +3625,7 @@ def transcription(
             atranscription=atranscription,
             timeout=timeout,
             logging_obj=litellm_logging_obj,
+            max_retries=max_retries,
         )
     return response