Tensorflow float64 转 float32

如下:


X = tf.dtypes.cast(X, tf.float32)