update deep_attention_matching_net _ce.py (#1587)

fluid/PaddleNLP/deep_attention_matching_net/_ce.py

@@ -7,8 +7,8 @@ from kpi import CostKpi, DurationKpi, AccKpi
 
 #### NOTE kpi.py should shared in models in some way!!!!
 
-train_cost_kpi = CostKpi('train_cost', 0.02, actived=True)
-train_duration_kpi = DurationKpi('train_duration', 0.05, actived=True)
+train_cost_kpi = CostKpi('train_cost', 0.02, 0, actived=True)
+train_duration_kpi = DurationKpi('train_duration', 0.05, 0, actived=True)
 
 tracking_kpis = [
     train_cost_kpi,