update

app.py

@@ -69,8 +69,8 @@ def respond(
                 if 'choices' in json_response and len(json_response['choices']) > 0:
                     content = json_response['choices'][0].get('message', {}).get('content', '')
                     if content:
-                        if '<think>' in content and '</think>' in content:
-                            content = content.split('</think>')[-1].strip()
+                        # if '<think>' in content and '</think>' in content:
+                        #     content = content.split('</think>')[-1].strip()
                         print(f"[INFO] response: {content}")
                         return content
             return "Service temporarily unavailable"
@@ -87,11 +87,11 @@ demo = gr.ChatInterface(
     additional_inputs=[
         gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
         gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
+        gr.Slider(minimum=0.1, maximum=4.0, value=0.6, step=0.1, label="Temperature"),
         gr.Slider(
             minimum=0.1,
             maximum=1.0,
-            value=0.96,
+            value=0.95,
             step=0.05,
             label="Top-p (nucleus sampling)",
         ),