Update evaluation/evaluator.py

evaluation/evaluator.py

@@ -77,6 +77,11 @@ def fix_prediction(prediction, gold_answer, evaluator):
 
 def question_scorer(prediction, gold_answer):
 
+    try:
+        prediction = json.loads(prediction)
+    except:
+        prediction = prediction
+
     answer_list = [x for x in gold_answer.split("\n") if len(x.strip()) > 0] if type(gold_answer) != list else gold_answer
     gold_answer, evaluator = parse_answer(answer_list)
     prediction, run_eval = fix_prediction(prediction, gold_answer, evaluator)