Update eval_utils.py

eval_utils.py

@@ -423,6 +423,7 @@ def get_evaluation_scores(gold_data, submission_data):
                 gold_data[task], submission_data[task]
             )
         elif task == "cjpe":
+            nltk.download('punkt')
             evaluation_results.update(
                 evaluate_cjpe(gold_data[task], submission_data[task])
             )