predict_label()一堆错误啊。。。

[1980695671]

Traceback (most recent call last):
File "test.py", line 14, in
model.predict_label(dateset)
File "/Users/yuan.jin/Desktop/gbdt_test/GBDT/gbdt/model.py", line 313, in predict_label
probs = self.predict_prob(instance)
File "/Users/yuan.jin/Desktop/gbdt_test/GBDT/gbdt/model.py", line 293, in predict_prob
f_value = self.compute_instance_f_value(instance)
File "/Users/yuan.jin/Desktop/gbdt_test/GBDT/gbdt/model.py", line 269, in compute_instance_f_value
f_value += self.learn_rate * iter.get_predict_value(instance)
File "/Users/yuan.jin/Desktop/gbdt_test/GBDT/gbdt/tree.py", line 24, in get_predict_value
elif not self.real_value_feature and instance[self.split_feature] == self.conditionValue:
TypeError: 'DataSet' object is not subscriptable