except custom openai proxy

litellm/main.py

@@ -162,6 +162,7 @@ def completion(
         ):  # allow custom provider to be passed in via the model name "azure/chatgpt-test"
             custom_llm_provider = model.split("/", 1)[0]
             model = model.split("/", 1)[1]
+        model, custom_llm_provider = get_llm_provider(model=model, custom_llm_provider=custom_llm_provider)
         # check if user passed in any of the OpenAI optional params
         optional_params = get_optional_params(
             functions=functions,
@@ -199,7 +200,6 @@ def completion(
             completion_call_id=id
         )
         logging.update_environment_variables(model=model, user=user, optional_params=optional_params, litellm_params=litellm_params)
-        model, custom_llm_provider = get_llm_provider(model=model, custom_llm_provider=custom_llm_provider)
         if custom_llm_provider == "azure":
             # azure configs
             api_type = get_secret("AZURE_API_TYPE") or "azure"
@@ -281,7 +281,6 @@ def completion(
                 litellm.openai_key or
                 get_secret("OPENAI_API_KEY")
             )
-
             ## LOGGING
             logging.pre_call(
                 input=messages,
@@ -375,7 +374,7 @@ def completion(
                 **optional_params
             )
             if "stream" in optional_params and optional_params["stream"] == True:
-                response = CustomStreamWrapper(response, model, logging_obj=logging)
+                response = CustomStreamWrapper(response, model, custom_llm_provider="openai", logging_obj=logging)
                 return response
             ## LOGGING
             logging.post_call(

litellm/tests/test_streaming.py

@@ -698,4 +698,4 @@ def test_openai_streaming_and_function_calling():
         pytest.fail(f"Error occurred: {e}")
         raise e 
 
-test_openai_streaming_and_function_calling()
+# test_openai_streaming_and_function_calling()

litellm/utils.py

@@ -889,7 +889,7 @@ def get_optional_params(  # use the openai defaults
         optional_params["return_full_text"] = return_full_text
         optional_params["details"] = True
         optional_params["task"] = task
-    elif custom_llm_provider == "together_ai" or ("togethercomputer" in model):
+    elif custom_llm_provider == "together_ai":
         if stream:
             optional_params["stream_tokens"] = stream
         if temperature != 1:
@@ -2520,8 +2520,7 @@ class CustomStreamWrapper:
                 chunk = next(self.completion_stream)
                 completion_obj["content"] = chunk
             elif (
-                self.custom_llm_provider and self.custom_llm_provider == "together_ai"
-            ) or ("togethercomputer" in self.model):
+                self.custom_llm_provider and self.custom_llm_provider == "together_ai"):
                 chunk = next(self.completion_stream)
                 text_data = self.handle_together_ai_chunk(chunk)
                 if text_data == "":

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "litellm"
-version = "0.1.681"
+version = "0.1.682"
 description = "Library to easily interface with LLM API providers"
 authors = ["BerriAI"]
 license = "MIT License"