update example to run on click

app.py

@@ -157,7 +157,7 @@ with gr.Blocks() as iface:
                     maximum=1.0,
                     value=0.8,
                     interactive=True,
-                    label="Temperature (set to 0 for greedy decoding under nucleus sampling)",
+                    label="Temperature (set to 0 for greedy decoding with nucleus sampling)",
                 )
 
                 len_penalty = gr.Slider(
@@ -166,7 +166,7 @@ with gr.Blocks() as iface:
                     value=1.0,
                     step=0.5,
                     interactive=True,
-                    label="Length Penalty (larger value encourages longer sequence under beam search)",
+                    label="Length Penalty (larger value encourages longer sequence with beam search)",
                 )
 
                 rep_penalty = gr.Slider(
@@ -231,9 +231,8 @@ with gr.Blocks() as iface:
     examples = gr.Examples(
         examples=examples,
         inputs=[image_input, chat_input],
-        cache_examples=True,
+        run_on_click=True,
         fn = inference_chat,
-        outputs = [chatbot, state],
     )
 
 iface.queue(concurrency_count=1, api_open=False, max_size=10)