more bug fixes.

llm.py

@@ -218,18 +218,15 @@ Chat with **Llama 3.1 70B** (AWQ INT4) served via vLLM on AMD Strix Halo (ROCm).
         with gr.TabItem("💬 Chat"):
             chatbot = gr.ChatInterface(
                 fn=chat_stream,
-                type="messages",
                 additional_inputs=[system_prompt, temperature, max_tokens, top_p],
                 examples=[
-                    "Hello! What can you tell me about yourself?",
-                    "Explain how a GPU executes a matrix multiplication.",
-                    "Write a Python function to compute the Fibonacci sequence.",
-                    "What are the pros and cons of running LLMs on AMD GPUs?",
+                    ["Hello! What can you tell me about yourself?"],
+                    ["Explain how a GPU executes a matrix multiplication."],
+                    ["Write a Python function to compute the Fibonacci sequence."],
+                    ["What are the pros and cons of running LLMs on AMD GPUs?"],
                 ],
                 chatbot=gr.Chatbot(
                     height=520,
-                    type="messages",
-                    show_copy_button=True,
                     placeholder="Type a message to start chatting...",
                 ),
             )