Update vlm.py

vlm.py

@@ -62,7 +62,12 @@ def build_messages(message: dict, history: list[tuple]):
     combined_user_input = [] #Combine images and text if found in same turn.
     for user_turn, bot_turn in history:
         if isinstance(user_turn, tuple):  # Image input
-            image_content = [{"type": "image_url", "image_url": image_url} for image_url in user_turn]
+            image_content = [
+                {
+                    "type": "image_url",
+                    "image_url": f"data:image/jpeg;base64,{encode_image(image)}"
+                } for image_url in user_turn
+            ]
             combined_user_input.extend(image_content)
         elif isinstance(user_turn, str): #Text input
             combined_user_input.append({"type":"text", "text": user_turn})