Update app.py

app.py

@@ -13,12 +13,14 @@ DEFAULT_MAX_NEW_TOKENS = 512
 MAX_INPUT_TOKEN_LENGTH = int(os.getenv("MAX_INPUT_TOKEN_LENGTH", "8192"))
 
 if not torch.cuda.is_available():
-    DESCRIPTION += "\n<p>Running on CPU 🥶 This demo does not work on CPU.</p>"
+    print("❌ CUDA GPU not available. This demo requires a GPU to function properly.")
+    print("Please run this application on a system with CUDA-compatible GPU.")
+    exit(1)
 
-if torch.cuda.is_available():
-    model_id = "utter-project/EuroLLM-9B-Instruct"
-    tokenizer = AutoTokenizer.from_pretrained(model_id)
-    model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto")
+# Load model and tokenizer
+model_id = "utter-project/EuroLLM-9B-Instruct"
+tokenizer = AutoTokenizer.from_pretrained(model_id)
+model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto")
 
 @spaces.GPU(duration=30)
 def generate(
@@ -121,4 +123,4 @@ with gr.Blocks(css="style.css") as demo:
     chat_interface.render()
 
 if __name__ == "__main__":
-    demo.queue(max_size=20).launch()
+    demo.queue(max_size=20).launch()