Update app.py

app.py

@@ -108,9 +108,9 @@ with gr.Blocks(css=css) as demo:
     """
     # Qwen2-VL Object Detection Demo
 
-    Use the Qwen2-VL model to detect objects in an image.
+    Use the Qwen2-VL models to detect objects in an image. The 7B variant seems to work much better.
 
-    Usage: Use the keyword "detect" and a description of the target (see examples below).
+    **Usage**: Use the keyword "detect" and a description of the target (see examples below).
     """)
     with gr.Tab(label="Qwen2-VL Input"):
         with gr.Row():