updated spaces.GPU

app.py

@@ -197,7 +197,6 @@ class InferenceRunner:
         return graph_list, input_text
 
     @torch.no_grad()
-    @spaces.GPU
     def predict(self, rxn_dict, temperature=1):
         graphs, prompt_tokens = self.tokenize(rxn_dict)
         result_dict = rxn_dict
@@ -236,6 +235,7 @@ class InferenceRunner:
         input_prompt_tokens['is_mol_token'] = is_mol_token
         return graphs, input_prompt_tokens
 
+@spaces.GPU
 def main(args):
     device = torch.device('cuda')
     # model