use cuda if available

app.py

@@ -1,7 +1,8 @@
 import gradio as gr
 from transformers import GPT2LMHeadModel, pipeline
+import torch
 
-device = 'cuda' if args.cuda else 'cpu'
+device = 'cuda' if torch.cuda.is_available() else 'cpu'
 
 # load pretrained + finetuned GPT2
 model = GPT2LMHeadModel.from_pretrained("./model/pytorch_model.bin", from_pt=True)