Update app.py

app.py

@@ -2,6 +2,7 @@ import gradio as gr
 import torch
 from transformers import AutoModelForCausalLM, AutoTokenizer
 import os
+import spaces
 
 MODEL_NAME = "swiss-ai/Apertus-8B-Instruct-2509"
 device = "cuda" if torch.cuda.is_available() else "cpu"
@@ -11,6 +12,7 @@ HF_TOKEN = os.getenv("HF_TOKEN")
 tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME, token=HF_TOKEN)
 model = AutoModelForCausalLM.from_pretrained(MODEL_NAME, token=HF_TOKEN).to(device)
 
+@spaces.GPU
 def predict(message, history):
     messages = []
     
@@ -53,4 +55,4 @@ chatbot = gr.ChatInterface(
 )
 
 # Launch the app
-chatbot.launch()
+chatbot.launch(share=True, server_port=80, enable_queue=True)