Update app.py

app.py

@@ -5,9 +5,18 @@ from assistant import get_assistant_response
 # Load model pipeline
 pipe = load_model()
 
+# Wrapper to handle dynamic output length
+def smart_response(command, execute):
+    # Dynamically adjust max_new_tokens based on complexity
+    simple_keywords = ["open", "close", "shutdown", "restart", "lock", "screenshot"]
+    is_simple = any(keyword in command.lower() for keyword in simple_keywords)
+    max_tokens = 128 if is_simple else 1024  # fast for simple, detailed for complex
+
+    return get_assistant_response(pipe, command, execute, max_new_tokens=max_tokens)
+
 # Gradio UI
 demo = gr.Interface(
-    fn=lambda command, execute: get_assistant_response(pipe, command, execute),
+    fn=smart_response,
     inputs=[
         gr.Textbox(lines=2, placeholder="e.g. Open Chrome or Take a screenshot", label="🧾 Command"),
         gr.Checkbox(label="🛠️ Execute command (if possible)")
@@ -15,9 +24,10 @@ demo = gr.Interface(
     outputs="text",
     title="🧠 Smart AI Assistant",
     description=(
-        "AI assistant powered by Qwen2.5-VL.\n"
+        "AI assistant powered by google/medgemma-4b-it\n\n"
         "Try commands like: 'Open Chrome', 'Take a screenshot', 'Create a file', or 'What is AI?'\n"
-        "Check the box to allow real OS command execution."
+        "✔️ Short commands = fast responses | 🧠 Big questions = detailed answers\n"
+        "✅ Enable the checkbox to let it run real OS commands (where allowed)"
     ),
     flagging_mode="never"
 )