Update app.py

app.py

@@ -6,11 +6,12 @@ os.environ["CUDA_VISIBLE_DEVICES"] = ""  # Prevents CUDA initialization
 import spaces
 import gradio as gr
 from huggingface_hub import InferenceClient, login
-
 import time
 
 
-
+@spaces.GPU  # Forces GPU allocation before execution
+def force_gpu_allocation():
+    pass  # Dummy function to trigger GPU setup
 
 # Base model (LLaMA 3.1 8B) from Meta
 base_model_name = "meta-llama/Llama-3.1-8B"
@@ -35,9 +36,7 @@ def chatbot_response(user_input):
     if torch.cuda.is_initialized():
         print("CUDA was already initialized before Accelerator!")
 
-    @spaces.GPU  # Forces GPU allocation before execution
-    def force_gpu_allocation():
-        pass  # Dummy function to trigger GPU setup
+
 
     accelerator = Accelerator()