作用:类型转换函数
tf.cast(x, dtype, name=None)
参数:
- x:输入
- dtype:转换目标类型
- name:名称
返回:Tensor,跟输入同shape
这里补充下tf的数据类型,作为tf.cast()的参数时,需要写为tf.float32.
uint8, uint16, uint32, uint64, int8, int16, int32, int64, float16, float32, float64, complex64, complex128, bfloat16
例子:
import tensorflow as tf
sess = tf.Session()
a = tf.constant(8)
print a
print sess.run(a)
b = tf.cast(a, tf.float32)
print b
print sess.run(b)
输出:
Tensor("Const:0", shape=(), dtype=int32)
8
Tensor("Cast:0", shape=(), dtype=float32)
8.0
该函数常用于解决:TypeError: Expected int32, got 512.0 of type 'float' instead.
举个例子,会报上述错误,也就是说进行相除的两个tensor,得是同类型的,此时tf.cast()来救场:
import tensorflow as tf
a = tf.constant(50)
f = tf.div(a, 512.0)
sess = tf.Session()
print sess.run(f)
转换类型后:
import tensorflow as tf
a = tf.constant(50)
f = tf.div(tf.cast(a, tf.float32), 512.0)
sess = tf.Session()
print sess.run(f)
输出:
0.09765625