Update app.py

app.py

@@ -19,8 +19,7 @@ from exception import CustomExceptionHandling
 
 
 # Download gguf model files
-llm = None
-llm_model = None
+huggingface_token = os.getenv("HUGGINGFACE_TOKEN")
 
 hf_hub_download(
     repo_id="bartowski/google_gemma-3-1b-it-GGUF",
@@ -38,6 +37,10 @@ title = "Gemma3 Llama.cpp"
 description = """SmolLM2, a family of three small language models, performs well in instruction following and reasoning. The largest model significantly improves over its predecessor through advanced training techniques."""
 
 
+# Download gguf model files
+llm = None
+llm_model = None
+
 def respond(
     message: str,
     history: List[Tuple[str, str]],
@@ -88,7 +91,7 @@ def respond(
         # Create the agent
         agent = LlamaCppAgent(
             provider,
-            # system_prompt=f"{system_message}",
+            system_prompt=f"{system_message}",
             predefined_messages_formatter_type=MessagesFormatterType.GEMMA_2,
             debug_output=True,
         )
@@ -148,16 +151,16 @@ demo = gr.ChatInterface(
                 "google_gemma-3-1b-it-Q6_K.gguf",
                 "google_gemma-3-1b-it-Q5_K_M.gguf",
             ],
-            value="google_gemma-3-1b-it-Q6_K.gguf",
+            value="google_gemma-3-1b-it-Q5_K_M.gguf",
             label="Model",
             info="Select the AI model to use for chat",
         ),
-        # gr.Textbox(
-        #     value="You are a helpful AI assistant focused on accurate and ethical responses.",
-        #     label="System Prompt",
-        #     info="Define the AI assistant's personality and behavior",
-        #     lines=2,
-        # ),
+        gr.Textbox(
+            value="You are a helpful AI assistant focused on accurate and ethical responses.",
+            label="System Prompt",
+            info="Define the AI assistant's personality and behavior",
+            lines=2,
+        ),
         gr.Slider(
             minimum=512,
             maximum=2048,