Update app.py

app.py

@@ -31,7 +31,7 @@ from torchvision.ops import nms
 
 
 from transformers import (AutoTokenizer, CLIPTextModelWithProjection)
-from transformers import (AutoProcessor, CLIPVisionModelWithProjection)
+from transformers import (AutoProcessor, CLIPImageProcessor, CLIPVisionModelWithProjection)
 
 BOUNDING_BOX_ANNOTATOR = sv.BoundingBoxAnnotator(thickness=2)
 MASK_ANNOTATOR = sv.MaskAnnotator()
@@ -298,7 +298,7 @@ if __name__ == '__main__':
     # init vision encoder
     clip_model = "openai/clip-vit-base-patch32"
     vision_model = CLIPVisionModelWithProjection.from_pretrained(clip_model)
-    processor = AutoProcessor.from_pretrained(clip_model)
+    processor = CLIPImageProcessor.from_pretrained(clip_model)
     device = 'cuda'
     vision_model.to(device)