Update app.py

app.py

@@ -28,7 +28,17 @@ model = AutoModelForCausalLM.from_pretrained(
         model_id, token=huggingface_token ,torch_dtype=dtype,device_map=device
     )
 text_generator = pipeline("text-generation", model=model, tokenizer=tokenizer,torch_dtype=dtype,device_map=device) #pipeline has not to(device)
-    
+
+if next(model.parameters()).is_cuda:
+    print("The model is on a GPU")
+else:
+    print("The model is on a CPU")
+
+# Assuming 'text_generator' is your initialized pipeline
+if text_generator.model.is_cuda:
+    print("The pipeline is using a GPU")
+else:
+    print("The pipeline is using a CPU")
 
 @spaces.GPU(duration=120)
 def generate_text(messages):