Update app.py

app.py

@@ -8,10 +8,6 @@ tokenizer = AutoTokenizer.from_pretrained("gpt2")
 model = AutoModelForCausalLM.from_pretrained("gpt2")
 
 print("Loading finished.")
-global n_steps=12
-global n_beams=1
-global length_penalty=1
-global num_return_sequences=3
 
 print(f"Is CUDA available: {torch.cuda.is_available()}")
 # True
@@ -485,7 +481,10 @@ with gr.Blocks(
         value="Hugging Face is",
     )
 
-   
+    n_steps=12
+    n_beams=1
+    length_penalty=1
+    num_return_sequences=3
     button = gr.Button()
     out_html = gr.Markdown()
     out_markdown = gr.Markdown()