Trying PIL and Image

utils/caption_utils.py

@@ -1,3 +1,4 @@
+from PIL import Image
 import torch
 from transformers import BlipProcessor, BlipForConditionalGeneration
 from utils.image_utils import load_image
@@ -14,7 +15,8 @@ class ImageCaptioning:
     def get_caption(self, image):
 
         # Preprocessing the Image
-        img = self.processor(image, return_tensors="pt").to(device)
+        img = Image.open(image)
+        img = self.processor(img, return_tensors="pt").to(device)
 
         # Generating captions
         output = self.model.generate(**img)