Update README.md

README.md

@@ -61,7 +61,8 @@ def generate(question_list,model_path):
     sampling_params = SamplingParams(
         max_tokens=4096,
         temperature=0.0,
-        n=1
+        n=1,
+        skip_special_tokens=False
     )
     outputs = llm.generate(question_list, sampling_params, use_tqdm=True)
     completions = [[output.text for output in output_item.outputs] for output_item in outputs]