[FixBug]Aeala/ShareGPT_Vicuna_unfiltered marked as multimodal benchmark (#28265)

Signed-off-by: princepride <wangzhipeng628@gmail.com>
This commit is contained in:
汪志鹏
2025-11-07 09:35:22 +00:00
committed by GitHub
parent ccd98b59c1
commit 315068eb4a
2 changed files with 73 additions and 2 deletions
+7
View File
@@ -21,6 +21,7 @@ from vllm.benchmarks.datasets import (
BurstGPTDataset,
ConversationDataset,
InstructCoderDataset,
MultiModalConversationDataset,
PrefixRepetitionRandomDataset,
RandomDataset,
SampleRequest,
@@ -367,6 +368,11 @@ def get_requests(args, tokenizer):
elif args.dataset_path in InstructCoderDataset.SUPPORTED_DATASET_PATHS:
dataset_cls = InstructCoderDataset
common_kwargs["dataset_split"] = "train"
elif args.dataset_path in MultiModalConversationDataset.SUPPORTED_DATASET_PATHS:
dataset_cls = MultiModalConversationDataset
common_kwargs["dataset_subset"] = args.hf_subset
common_kwargs["dataset_split"] = args.hf_split
sample_kwargs["enable_multimodal_chat"] = True
elif args.dataset_path in ConversationDataset.SUPPORTED_DATASET_PATHS:
dataset_cls = ConversationDataset
common_kwargs["dataset_subset"] = args.hf_subset
@@ -456,6 +462,7 @@ def validate_args(args):
elif args.dataset_name == "hf":
if args.dataset_path in (
VisionArenaDataset.SUPPORTED_DATASET_PATHS.keys()
| MultiModalConversationDataset.SUPPORTED_DATASET_PATHS
| ConversationDataset.SUPPORTED_DATASET_PATHS
):
assert args.backend == "vllm-chat", (