[fix] fix completion stream api output_tokens not in usage (#3247)

2025-10-05 16:48:03 +08:00 · 2025-08-07 10:36:00 +08:00
parent d9e3f88f9e
commit 09cc4e2802
1 changed files with 1 additions and 0 deletions
--- a/fastdeploy/entrypoints/openai/serving_completion.py
+++ b/fastdeploy/entrypoints/openai/serving_completion.py
@@ -358,6 +358,7 @@ class OpenAIServingCompletion:
                                usage=UsageInfo(
                                    prompt_tokens=len(prompt_batched_token_ids[idx]),
                                    completion_tokens=output_tokens[idx],
+                                    total_tokens=len(prompt_batched_token_ids[idx]) + output_tokens[idx],
                                ),
                            )
                            yield f"data: {usage_chunk.model_dump_json(exclude_unset=True)}\n\n"