fix(azure): return response headers for sync embedding calls

2025-04-26 11:14:04 +00:00 · 2024-09-28 17:03:13 -07:00 · 2024-09-28 17:03:13 -07:00 · d64e971d8c
commit d64e971d8c
parent 55d7bc7f32
3 changed files with 30 additions and 5 deletions
--- a/litellm/utils.py
+++ b/litellm/utils.py
@ -1422,9 +1422,11 @@ def client(original_function):
                or isinstance(result, EmbeddingResponse)
                or isinstance(result, TranscriptionResponse)
            ):
-                result._response_ms = (
-                    end_time - start_time
-                ).total_seconds() * 1000  # return response latency in ms like openai
+                setattr(
+                    result,
+                    "_response_ms",
+                    (end_time - start_time).total_seconds() * 1000,
+                )  # return response latency in ms like openai

            ### POST-CALL RULES ###
            post_call_processing(