adapt to transformers==4.57.6

image_processing_qwen2_vl_fast.py

@@ -43,6 +43,8 @@ class Qwen2VLFastImageProcessorKwargs(DefaultFastImageProcessorKwargs):
         The temporal patch size of the vision encoder.
     merge_size (`int`, *optional*, defaults to 2):
         The merge size of the vision encoder to llm encoder.
+    focus_size (`int`, *optional*, defaults to 2):
+        The focus size of the ZoomFocus.
     """
 
     min_pixels: Optional[int]