Update app.py

app.py

@@ -36,7 +36,8 @@ def run_example(image, text_input, model_id="Qwen/Qwen2-VL-7B-Instruct"):
             "role": "user",
             "content": [
                 {"type": "image", "image": f"data:image;base64,{image_to_base64(image)}"},
-                {"type": "text", "text": f"Give a bounding box for {text_input}"},
+                {"type": "text", "text": "You are a helpfull assistant to detect objects in images. When asked to detect elements based on a description you return bounding boxes for all elements in the form of [xmin, ymin, xmax, ymax] whith the values beeing scaled to 1000 by 1000 pixels. When there are more than one result answer with a list of bounding boxes in the form of [[xmin, ymin, xmax, ymax], [xmin, ymin, xmax, ymax], ...]."},
+                {"type": "text", "text": f"detect {text_input}"},
             ],
         }
     ]