small fix

Author: wangshangsam

multimodal/vl2l/src/mlperf_inference_multimodal_vl2l/task.py

@@ -84,7 +84,7 @@ async def _cancel_all_tasks() -> None:
             tasks = [
                 task
                 for task in asyncio.all_tasks(self.event_loop)
-                if not task.done()
+                if task is not asyncio.current_task() and not task.done()
             ]
             for task in tasks:
                 task.cancel()
@@ -183,9 +183,7 @@ def estimated_num_performance_samples(self) -> int:
         """
         estimation_indices = random.sample(
             range(self.total_num_samples),
-            k=min(
-                MAX_NUM_ESTIMATION_PERFORMANCE_SAMPLES,
-                self.total_num_samples),
+            k=min(MAX_NUM_ESTIMATION_PERFORMANCE_SAMPLES, self.total_num_samples),
         )
         estimation_samples = [
             self.formulate_loaded_sample(
@@ -250,8 +248,7 @@ def _unload_samples_from_ram(query_sample_indices: list[int]) -> None:
             _unload_samples_from_ram,
         )
 
-    async def _query_endpoint_async_batch(
-            self, query_sample: lg.QuerySample) -> None:
+    async def _query_endpoint_async_batch(self, query_sample: lg.QuerySample) -> None:
         """Query the endpoint through the async OpenAI API client."""
         try:
             sample = self.loaded_samples[query_sample.index]
@@ -328,8 +325,7 @@ async def _query_endpoint_async_batch(
                 ],
             )
 
-    async def _query_endpoint_async_stream(
-            self, query_sample: lg.QuerySample) -> None:
+    async def _query_endpoint_async_stream(self, query_sample: lg.QuerySample) -> None:
         """Query the endpoint through the async OpenAI API client."""
         ttft_set = False
         try: