[Benchmark] Parameterization of streaming loading of multimodal datasets (#20528)

Signed-off-by: wangli <wangli858794774@gmail.com>

Author: Potabk

benchmarks/benchmark_dataset.py

@@ -701,13 +701,15 @@ def __init__(
         self,
         dataset_path: str,
         dataset_split: str,
+        no_stream: bool = False,
         dataset_subset: Optional[str] = None,
         **kwargs,
     ) -> None:
         super().__init__(dataset_path=dataset_path, **kwargs)
 
         self.dataset_split = dataset_split
         self.dataset_subset = dataset_subset
+        self.load_stream = not no_stream
         self.load_data()
 
     def load_data(self) -> None:
@@ -716,7 +718,7 @@ def load_data(self) -> None:
             self.dataset_path,
             name=self.dataset_subset,
             split=self.dataset_split,
-            streaming=True,
+            streaming=self.load_stream,
         )
         self.data = self.data.shuffle(seed=self.random_seed)
 

benchmarks/benchmark_serving.py

@@ -825,6 +825,7 @@ def main(args: argparse.Namespace):
             dataset_subset=args.hf_subset,
             dataset_split=args.hf_split,
             random_seed=args.seed,
+            no_stream=args.no_stream,
         ).sample(
             num_requests=args.num_prompts,
             tokenizer=tokenizer,
@@ -1033,6 +1034,11 @@ def create_argument_parser():
         help="Path to the sharegpt/sonnet dataset. "
         "Or the huggingface dataset ID if using HF dataset.",
     )
+    parser.add_argument(
+        "--no-stream",
+        action="store_true",
+        help="Do not load the dataset in streaming mode.",
+    )
     parser.add_argument(
         "--max-concurrency",
         type=int,

benchmarks/benchmark_throughput.py

@@ -356,6 +356,7 @@ def get_requests(args, tokenizer):
     elif args.dataset_name == "burstgpt":
         dataset_cls = BurstGPTDataset
     elif args.dataset_name == "hf":
+        common_kwargs["no_stream"] = args.no_stream
         if args.dataset_path in VisionArenaDataset.SUPPORTED_DATASET_PATHS:
             dataset_cls = VisionArenaDataset
             common_kwargs["dataset_subset"] = None
@@ -610,6 +611,11 @@ def create_argument_parser():
         help="Name of the dataset to benchmark on.",
         default="sharegpt",
     )
+    parser.add_argument(
+        "--no-stream",
+        action="store_true",
+        help="Do not load the dataset in streaming mode.",
+    )
     parser.add_argument(
         "--dataset",
         type=str,

vllm/benchmarks/datasets.py

@@ -481,6 +481,11 @@ def add_dataset_parser(parser: FlexibleArgumentParser):
         choices=["sharegpt", "burstgpt", "sonnet", "random", "hf", "custom"],
         help="Name of the dataset to benchmark on.",
     )
+    parser.add_argument(
+        "--no-stream",
+        action="store_true",
+        help="Do not load the dataset in streaming mode.",
+    )
     parser.add_argument(
         "--dataset-path",
         type=str,
@@ -674,6 +679,7 @@ def get_samples(args, tokenizer) -> list[SampleRequest]:
             dataset_subset=args.hf_subset,
             dataset_split=args.hf_split,
             random_seed=args.seed,
+            no_stream=args.no_stream,
         ).sample(
             num_requests=args.num_prompts,
             tokenizer=tokenizer,
@@ -971,13 +977,15 @@ def __init__(
         self,
         dataset_path: str,
         dataset_split: str,
+        no_stream: bool = False,
         dataset_subset: Optional[str] = None,
         **kwargs,
     ) -> None:
         super().__init__(dataset_path=dataset_path, **kwargs)
 
         self.dataset_split = dataset_split
         self.dataset_subset = dataset_subset
+        self.load_stream = not no_stream
         self.load_data()
 
     def load_data(self) -> None:
@@ -986,7 +994,7 @@ def load_data(self) -> None:
             self.dataset_path,
             name=self.dataset_subset,
             split=self.dataset_split,
-            streaming=True,
+            streaming=self.load_stream,
         )
         self.data = self.data.shuffle(seed=self.random_seed)