fix policy evaluation

convlab2/policy/evaluate.py

@@ -222,7 +222,6 @@ def evaluate(dataset_name, model_name, load_path, calculate_reward=True):
             for key in sess.evaluator.goal: 
                 if key not in task_success: 
                     task_success[key] = []
-                else: 
                 task_success[key].append(task_succ)
             task_success['All'].append(task_succ)