update app

app.py

@@ -25,7 +25,7 @@ def generate(gen_prompt, max_tokens, temperature=0.6, seed=42):
     output = model.generate(input_ids=input_ids, do_sample=True, top_p=0.95, temperature=temperature, max_length=max_length)
     generated_text = tokenizer.decode(output.flatten())
     if generated_text.startswith(BOS):
-        generated_text = detok_hypo_str[len(BOS):]
+        generated_text = generated_text[len(BOS):]
     return generated_text
 
 iface = gr.Interface(