phoenix/litellm-mirror
0
0
Fork 1
mirror of https://github.com/BerriAI/litellm.git synced 2025-04-27 11:43:54 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

feat(proxy_server.py): working /audio/transcription endpoint

Browse source
This commit is contained in:
Krrish Dholakia 2024-03-08 18:20:27 -08:00
parent cc0294b2f2
commit 0fb7afe820
6 changed files with 95 additions and 54 deletions
Download patch file Download diff file Expand all files Collapse all files

27
litellm/llms/azure.py
View file

@ -770,9 +770,8 @@ class AzureChatCompletion(BaseLLM):
api_version: Optional[str] = None,
client=None,
azure_ad_token: Optional[str] = None,
max_retries=None,
logging_obj=None,
atranscriptions: bool = False,
atranscription: bool = False,
):
data = {"model": model, "file": audio_file, **optional_params}
@ -781,9 +780,11 @@ class AzureChatCompletion(BaseLLM):
"api_version": api_version,
"azure_endpoint": api_base,
"azure_deployment": model,
"max_retries": max_retries,
"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
)
@ -792,7 +793,10 @@ class AzureChatCompletion(BaseLLM):
elif azure_ad_token is not None:
azure_client_params["azure_ad_token"] = azure_ad_token
if atranscriptions == True:
if max_retries is not None:
azure_client_params["max_retries"] = max_retries
if atranscription == True:
return self.async_audio_transcriptions(
audio_file=audio_file,
data=data,
@ -845,18 +849,29 @@ class AzureChatCompletion(BaseLLM):
)
else:
async_azure_client = client
response = await async_azure_client.audio.transcriptions.create(
**data, timeout=timeout
) # type: ignore
stringified_response = response.model_dump()
## LOGGING
logging_obj.post_call(
input=audio_file.name,
api_key=api_key,
additional_args={"complete_input_dict": data},
additional_args={
"headers": {
"Authorization": f"Bearer {async_azure_client.api_key}"
},
"api_base": async_azure_client._base_url._uri_reference,
"atranscription": True,
"complete_input_dict": data,
},
original_response=stringified_response,
)
return convert_to_model_response_object(response_object=stringified_response, model_response_object=model_response, response_type="image_generation") # type: ignore
response = convert_to_model_response_object(response_object=stringified_response, model_response_object=model_response, response_type="audio_transcription") # type: ignore
return response
except Exception as e:
## LOGGING
logging_obj.post_call(