A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 15, 04:45 UTC
[Bugfix] Fix the linter (#19826)
Signed-off-by: Lu Fang <lufang@fb.com>
Lu Fang committed
Jun 19, 2025 at 12:44 UTC
b1098b40723fd3cd02a6e30ca766bde083f61552
3 files changed
+15
-15
vllm/model_executor/models/qwen2_5_omni_thinker.py
+5
-5
@@ -146,11 +146,11 @@ class Qwen2_5OmniThinkerProcessingInfo(Qwen2AudioProcessingInfo,
146
kwargs["fps"] = fps
147
processor = self.ctx.get_hf_processor(
148
Qwen2_5OmniProcessor,
149
- image_processor=self.get_image_processor(
150
- min_pixels=min_pixels,
151
- max_pixels=max_pixels,
152
- size=size,
153
- use_fast=kwargs.get("use_fast", True)),
149
+ image_processor=self.get_image_processor(min_pixels=min_pixels,
150
+ max_pixels=max_pixels,
151
+ size=size,
152
+ use_fast=kwargs.get(
153
+ "use_fast", True)),
154
**kwargs,
155
)
156
if not hasattr(processor, "audio_token"):
vllm/model_executor/models/qwen2_5_vl.py
+5
-5
@@ -794,11 +794,11 @@ class Qwen2_5_VLProcessingInfo(Qwen2VLProcessingInfo):
794
795
return self.ctx.get_hf_processor(
796
Qwen2_5_VLProcessor,
797
- image_processor=self.get_image_processor(
798
- min_pixels=min_pixels,
799
- max_pixels=max_pixels,
800
- size=size,
801
- use_fast=kwargs.get("use_fast", True)),
797
+ image_processor=self.get_image_processor(min_pixels=min_pixels,
798
+ max_pixels=max_pixels,
799
+ size=size,
800
+ use_fast=kwargs.get(
801
+ "use_fast", True)),
802
**kwargs,
803
)
804
vllm/model_executor/models/qwen2_vl.py
+5
-5
@@ -759,11 +759,11 @@ class Qwen2VLProcessingInfo(BaseProcessingInfo):
759
) -> Qwen2VLProcessor:
760
return self.ctx.get_hf_processor(
761
Qwen2VLProcessor,
762
- image_processor=self.get_image_processor(
763
- min_pixels=min_pixels,
764
- max_pixels=max_pixels,
765
- size=size,
766
- use_fast=kwargs.get("use_fast", True)),
762
+ image_processor=self.get_image_processor(min_pixels=min_pixels,
763
+ max_pixels=max_pixels,
764
+ size=size,
765
+ use_fast=kwargs.get(
766
+ "use_fast", True)),
767
**kwargs,
768
)
769