Update app.py

app.py

@@ -17,8 +17,6 @@ processors = {
     "Qwen/Qwen2-VL-7B-Instruct": AutoProcessor.from_pretrained("Qwen/Qwen2-VL-7B-Instruct")
 }
 
-DESCRIPTION = "# Qwen2-VL Object Localization Demo"
-
 
 def image_to_base64(image):
     buffered = BytesIO()
@@ -101,10 +99,18 @@ css = """
     border: 1px solid #ccc; 
   }
 """
-default_system_prompt = "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], ...]."
+default_system_prompt = "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], ...]."
 
 with gr.Blocks(css=css) as demo:
     gr.Markdown(DESCRIPTION)
+    gr.Markdown(
+    """
+    # Qwen2-VL Object Detection Demo
+
+    Use the Qwen2-VL model to detect objects in an image.
+
+    Usage: Use the keyword "detect" and a description of the target (see examples below).
+    """)
     with gr.Tab(label="Qwen2-VL Input"):
         with gr.Row():
             with gr.Column():