180412 tf.placehoder()调用任意尺寸的数据以及tf.cast()的使用方法

① tf.placeholder()

极客学院-数据读取

• 代码

import tensorflow as tf
a = tf.placeholder(tf.float32) # 不指定tf.placeholder(tf.float32,shape)中的shape
v1 = [11.,12.]
v2 = [[1.,2.],[3.,4.]]
with tf.Session() as sess:
    print("Values of v1 is:\n",sess.run(a,{a:v1}))
    print("Values of v2 is:\n",sess.run(a,{a:v2}))

• 结果

Shape of v1 is:
 [ 11.  12.]
Shape of v2 is:
 [[ 1.  2.]
 [ 3.  4.]]

---

② tf.cast()

tf.cast(data,dtype,name=None)
作用：将data类型的数据转换成指定dtype类型的数据

tf.cast

代码

import tensorflow as tf
a = tf.placeholder(tf.bool)
v1 = [[True,False],[False,True]]
v2 = tf.cast(v1,dtype=tf.float32) # 将tf.bool型数据v1转换成tf.float32类型v2
with tf.Session() as sess:
    init = tf.global_variables_initializer()
    print(sess.run(a,{a:v1}))
    print(sess.run(v2))

结果

[[ True False]
 [False  True]]
[[ 1.  0.]
 [ 0.  1.]]