Update eval_utils.py

eval_utils.py

@@ -450,6 +450,7 @@ def get_evaluation_scores(gold_data, submission_data):
                 gold_data[task], submission_data[task]
             )
         elif task == "summ":
+            nltk.download('punkt')
             evaluation_results[task] = evaluate_summ(
                 gold_data[task], submission_data[task]
             )