Update README.md

README.md

@@ -89,6 +89,19 @@ Traditional approaches typically rely on a simple binary classification — "fin
 
 <br/>
 
+<div align="center">
+  <video src="https://huggingface.co/brgroup/TurnSense/resolve/main/image/PR_new.mp4" 
+         width="820" 
+         height="400" 
+         controls 
+         muted 
+         loop 
+         autoplay 
+         playsinline>
+  </video>
+</div>
+
+
 TurnSense classifies user input into three semantic states:
 
 | State | Description | Example |