下面是重现以下错误的示例！
import tensorflow.keras.layers as tks
input_ = tks.Input(shape=[None, None, 3], dtype='float64')
conv = tks.Conv2D(filters=3, kernel_size=3)(input_)
batch = tks.BatchNormalization()(conv)
我正在跑步tensorflow 1.14 and keras 2.4.0
TypeError：传递给参数'x'的值的数据类型float64不在允许的值列表中：float32