Update app.py

app.py

@@ -23,7 +23,28 @@ examples = [
     ["What is the square root of 64?", "The square root of 64 is 8."]
 ]
 
+with gr.Blocks() as demo:
+    chatbot = gr.Chatbot()
+    msg = gr.Textbox()
+    clear = gr.Button("Clear")
+
+    def user(user_message, history):
+        return "", history + [[user_message, None]]
+
+    def bot(history):
+        bot_message = output_text
+        history[-1][1] = ""
+        for character in bot_message:
+            history[-1][1] += character
+            time.sleep(0.05)
+            yield history
+
+    msg.submit(user, [msg, chatbot], [msg, chatbot], queue=False).then(
+        bot, chatbot, chatbot
+    )
+    clear.click(lambda: None, None, chatbot, queue=False)
+    
+demo.queue()
+if __name__ == "__main__":
+    demo.launch()
 
-demo = gr.Interface(fn=generate_text, inputs=input_text, outputs=output_text, title="Llama Language Model", description=description, examples=examples)
-demo.queue(concurrency_count=3)
-demo.launch()