ValueError: X.dtype should be np.float32, got float64

背景

在用GBDT系列训练时，报错ValueError: X.dtype should be np.float32, got float64，如下所示。

ValueError                                Traceback (most recent call last)
<ipython-input-14-aa936862d7d7> in <module>()
----> 1 abc.apply(X_train)

~/tmp/dataset/Augboost+FM/AugBoost.py in apply(self, X)
    461             for j in range(n_classes):
    462                 estimator = self.estimators_[i, j]
--> 463                 leaves[:, i, j] = estimator.apply(np.concatenate([X_original, X], axis=1), check_input=False)
    464 
    465         return leaves

~/anaconda3/lib/python3.7/site-packages/sklearn/tree/tree.py in apply(self, X, check_input)
    464         check_is_fitted(self, 'tree_')
    465         X = self._validate_X_predict(X, check_input)
--> 466         return self.tree_.apply(X)
    467 
    468     def decision_path(self, X, check_input=True):

sklearn/tree/_tree.pyx in sklearn.tree._tree.Tree.apply()

sklearn/tree/_tree.pyx in sklearn.tree._tree.Tree.apply()

sklearn/tree/_tree.pyx in sklearn.tree._tree.Tree._apply_dense()

ValueError: X.dtype should be np.float32, got float64

解决方法

很显然，就是字面上的意思，只能是np.float32，但是给出的是float64
我看了tree.py

sklearn内置的代码一路走下来就应该是32位的。所以怀疑是自己前面int类型的输入训练集在转化是转化为了64位

看了一下，前面有这样的代码

X_original = X
X_normed = self.normalizer.transform(X)

X是我输入的X_train ，dataframe格式，int类型的数据

看下X_normed

ok，发现问题了，经过normalizer.transform()我的数据变成了numpy.ndarray类型，float64。那么咱们把类型转化过来就行了

ndarray的数据类型：

https://blog.csdn.net/weixin_43181110/article/details/83996915?spm=1001.2101.3001.6650.1&utm_medium=distribute.pc_relevant.none-task-blog-2%7Edefault%7ECTRLIST%7Edefault-1.no_search_link&depth_1-utm_source=distribute.pc_relevant.none-task-blog-2%7Edefault%7ECTRLIST%7Edefault-1.no_search_link

X_normed = X_normed.astype(np.float32)

这样就可以啦