@spaces.GPU decorator added

app.py

@@ -2,6 +2,7 @@ import gradio as gr
 import torch
 import numpy
 from huggingface_hub import login
+import spaces
 import os
 login(os.getenv('LOGIN_TOKEN'))
 
@@ -15,6 +16,7 @@ gemma_model = AutoModelForCausalLM.from_pretrained(
     torch_dtype=torch.bfloat16
 )
 
+@spaces.GPU
 def promptGemma2(textprompt, max_tokens=20):
   input_ids = gemma_tokenizer(textprompt, return_tensors="pt")#.to("cuda")
   outputs = gemma_model.generate(**input_ids, max_new_tokens=max_tokens)