prepare for zerogpu

app.py

@@ -20,6 +20,7 @@ from transformers import (
 from peft import AutoPeftModelForCausalLM
 import torch
 import os
+import spaces
 
 # Maximum execution time
 thread_timeout = 600
@@ -32,7 +33,7 @@ if os.environ.get("HF_TOKEN") is None:
 # sets the main paremeters
 hugging_face_model_id = "eltorio/Llama-3.2-3B-appreciation"
 base_model_path = "meta-llama/Llama-3.2-3B-Instruct"
-device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu")
+device = 'cuda' if torch.cuda.is_available() else 'cpu'
 
 # Define the title, description, and device description for the Gradio interface
 device_desc = (
@@ -89,6 +90,7 @@ def get_conversation(
 
 
 # Define a function to infer a evaluation from the incoming parameters
+@spaces.GPU
 def infere(
     trimestre: str,
     moyenne_1: float,