Merge pull request #1935 from stephenleo/main

fix(lowest_latency.py): fix latency calc (lower better)
2026-08-21 02:23:59 +00:00 · 2024-02-14 14:59:11 -08:00
parent ed8f507536 37c83e0023
commit a6d5d0f714
1 changed files with 1 additions and 1 deletions
@@ -86,7 +86,7 @@ class LowestLatencyLoggingHandler(CustomLogger):
                if isinstance(response_obj, ModelResponse):
                    completion_tokens = response_obj.usage.completion_tokens
                    total_tokens = response_obj.usage.total_tokens
-                    final_value = float(completion_tokens / response_ms.total_seconds())
+                    final_value = float(response_ms.total_seconds() / completion_tokens)

                # ------------
                # Update usage