[Refactor] Define MultiModalKwargsItems separate from MultiModalKwargs (#23053)

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
This commit is contained in:
Cyrus Leung
2025-08-18 09:52:00 +00:00
committed by GitHub
parent 5c79b0d648
commit 27e8d1ea3e
77 changed files with 431 additions and 383 deletions
@@ -30,7 +30,7 @@ from vllm.model_executor.models.module_mapping import MultiModelKeys
from vllm.model_executor.sampling_metadata import SamplingMetadata
from vllm.multimodal import MULTIMODAL_REGISTRY
from vllm.multimodal.inputs import (MultiModalDataDict, MultiModalFieldConfig,
MultiModalKwargs, NestedTensors)
MultiModalKwargsItems, NestedTensors)
from vllm.multimodal.parse import (AudioProcessorItems, ImageEmbeddingItems,
ImageProcessorItems, ImageSize,
MultiModalDataItems, MultiModalDataParser)
@@ -1029,7 +1029,7 @@ class Phi4MMMultiModalProcessor(BaseMultiModalProcessor[Phi4MMProcessingInfo]):
self,
mm_items: MultiModalDataItems,
hf_processor_mm_kwargs: Mapping[str, Any],
out_mm_kwargs: MultiModalKwargs,
out_mm_kwargs: MultiModalKwargsItems,
) -> Sequence[PromptUpdate]:
tokenizer = self.info.get_tokenizer()
image_token_id = tokenizer.vocab[tokenizer.image_token]