biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[VLM] Add max-count checking in data parser for single image models (#11661)

Browse Source 
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
Signed-off-by: Roger Wang <ywang@roblox.com>
Co-authored-by: Roger Wang <ywang@roblox.com>
This commit is contained in:
Cyrus Leung
2025-01-01 14:15:21 +08:00
committed by GitHub
parent 4db72e57f6
commit 365801fedd
6 changed files with 48 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
vllm/model_executor/models/chameleon.py
View File

@@ -31,6 +31,7 @@ from vllm.multimodal import MULTIMODAL_REGISTRY
from vllm.multimodal.inputs import (MultiModalDataDict, MultiModalFieldConfig,
MultiModalInputsV2, MultiModalKwargs,
NestedTensors, PlaceholderRange)
from vllm.multimodal.parse import MultiModalDataParser
from vllm.multimodal.processing import (BaseMultiModalProcessor,
MultiModalDataItems, ProcessorInputs,
PromptReplacement)
@@ -60,6 +61,9 @@ def get_max_chameleon_image_tokens(ctx: InputContext):
class ChameleonMultiModalProcessor(BaseMultiModalProcessor):
def _get_data_parser(self) -> MultiModalDataParser:
return MultiModalDataParser(max_mm_counts={"image": 1})
def _get_hf_processor(self) -> ChameleonProcessor:
return self.ctx.get_hf_processor(ChameleonProcessor)