[Bugfix][Core] Restore logging of stats in the async engine (vllm-pro…

…ject#4150)
sionic-ai · Apr 19, 2024 · 7be4f56 · 7be4f56
1 parent 8f20fc0
commit 7be4f56
Showing 1 changed file with 7 additions and 1 deletion.
diff --git a/vllm/engine/async_llm_engine.py b/vllm/engine/async_llm_engine.py
@@ -217,10 +217,16 @@ async def step_async(self) -> List[RequestOutput]:
         else:
             output = []
 
-        return self._process_model_outputs(
+        request_outputs = self._process_model_outputs(
             output, scheduler_outputs.scheduled_seq_groups,
             scheduler_outputs.ignored_seq_groups)
 
+        # Log stats.
+        if self.log_stats:
+            self.stat_logger.log(self._get_stats(scheduler_outputs))
+
+        return request_outputs
+
     async def encode_request_async(
         self,
         request_id: str,  # pylint: disable=unused-argument