Update app.py

app.py

@@ -18,9 +18,8 @@ llm = Llama(
     model_path=model_path,
     n_ctx=2048,
     n_threads=n_threads,
-    n_gpu_layers=0 
+    n_gpu_layers=0
 )
-# I see from your logs it found 16 threads! That's a powerful CPU Space.
 print(f"Model loaded for CPU execution with {n_threads} threads.")
 
 
@@ -37,23 +36,23 @@ You are Dolphin 3.0, a helpful and friendly AI assistant.
 {message}
 
 ### Assistant:"""
-    
+
     stream = llm(
         full_prompt,
         max_tokens=1024,
         stop=["</s>", "### User:", "### Assistant:"],
         stream=True
     )
-    
+
     partial_message = ""
     for output in stream:
         token = output['choices'][0]['text']
         partial_message += token
         yield partial_message
 
-# --- 4. The Enhanced Chatbot UI (Corrected!) ---
-# We've removed the 'retry_btn' and 'undo_btn' arguments to match the
-# version of Gradio running on Hugging Face Spaces.
+# --- 4. The Enhanced Chatbot UI (MAXIMUM COMPATIBILITY) ---
+# We are removing ALL custom button arguments to ensure this works on older Gradio versions.
+# Gradio will add the default 'Undo' and 'Clear' buttons for us.
 iface = gr.ChatInterface(
     fn=chat,
     title="🐬 Dolphin 3.0 on Hugging Face Spaces",
@@ -63,7 +62,6 @@ iface = gr.ChatInterface(
     theme="soft",
     examples=[["Hello!"], ["Write a short poem about the stars."], ["What is the capital of India?"]],
     cache_examples=False,
-    clear_btn="Clear Chat", # This one is universally loved and supported
 )