Update app.py

app.py

@@ -10,10 +10,14 @@ args = dict(
     template="llama3",
     finetuning_type="lora",
     quantization_bit=4,
-    device="cpu",  # Forces CPU usage
 )
+
 chat_model = ChatModel(args)
 
+# Force the model to CPU
+device = torch.device("cpu")
+chat_model.model.to(device)
+
 # Step 2: Create a function for chatting
 def chat(user_input, history):
     messages = history + [{"role": "user", "content": user_input}]