Update app.py

app.py

@@ -3,9 +3,8 @@ from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
 import torch
 import spaces
 
-# Load tokenizer and model once at startup
+# Load model and tokenizer
 model_name = "Qwen/Qwen2.5-0.5B-Instruct"
-
 tokenizer = AutoTokenizer.from_pretrained(model_name)
 model = AutoModelForCausalLM.from_pretrained(
     model_name,
@@ -23,7 +22,7 @@ pipe = pipeline(
     do_sample=True
 )
 
-@spaces.GPU(duration=60)  # Use shared GPU for faster inference
+@spaces.GPU(duration=60)
 def sales_agent(message):
     prompt = f"""
 You are a helpful pre-sales agent at Cotubex, a tech store in Brussels. Answer in the same language as the question.
@@ -42,14 +41,20 @@ Answer:
     response = pipe(prompt)
     return response[0]['generated_text'].replace(prompt, "").strip()
 
-# Gradio Chat Interface
-demo = gr.ChatInterface(
-    fn=sales_agent,
-    chatbot=gr.Chatbot(height=400),
-    additional_inputs=[],
-    title="Cotubex Pre-Sales Assistant 🇫🇷🇧🇪🇳🇱",
-    description="Ask about product details, prices, availability — in French, Dutch, or English!"
-)
+# ChatInterface with default examples
+examples = [
+    ["Quel est le prix de la carte graphique Zotac 5060 TI ?"],
+    ["Wat kost de Zotac 5060 TI videokaart?"],
+    ["Is the Be Quiet water cooler available?"]
+]
+
+with gr.Blocks(css="style.css") as demo:
+    gr.Markdown("### Cotubex Pre-Sales Assistant 🇫🇷🇧🇪🇳🇱\nAsk us anything about products, pricing, availability.")
+    gr.ChatInterface(
+        fn=sales_agent,
+        chatbot=gr.Chatbot(height=400, type="messages"),
+        examples=examples
+    )
 
 if __name__ == "__main__":
-    demo.launch(css="style.css")
+    demo.launch()