TensorFlow 中的占位符

占位符是 Tensorflow 中的一个变量，稍后将为其分配数据。它使我们能够在不需要数据的情况下创建流程或操作。会话开始时，数据被输入占位符，会话开始运行。我们可以使用占位符将数据输入到张量流图中。

Syntax: tf.compat.v1.placeholder(dtype, shape=None, name=None)

Parameters:

dtype: the datatype of the elements in the tensor that will be fed.
shape : by default None. The tensor’s shape that will be fed , it is an optional parameter. One can feed a tensor of any shape if the shape isn’t specified.
name: by default None. The operation’s name , optional parameter.

Returns:

A Tensor that can be used to feed a value but cannot be evaluated directly.

编程需要懂一点英语

示例 1：

Python3

# importing packages
import tensorflow.compat.v1 as tf
  
# disabling eager mode
tf.compat.v1.disable_eager_execution()
  
# creating a placeholder
a = tf.placeholder(tf.float32, None)
  
# creating an operation
b = a + 10
  
# creating a session
with tf.Session() as session:
    
    # feeding data in the placeholder
    operation_res = session.run(b, feed_dict={a: [10, 20, 30, 40]})
    print("after executing the operation: " + str(operation_res))

Python3

# importing packages
import tensorflow.compat.v1 as tf
  
# disabling eager mode
tf.compat.v1.disable_eager_execution()
  
# creating a tensorflow graph
graph = tf.Graph()
with graph.as_default():
    
    # creating a placeholder
    a = tf.placeholder(tf.float64, shape=(3, 3), name='tensor1')
      
    # creating an operation
    b = a ** 2
  
# array1 will be fed into 'a'
array1 = [[1, 2, 3], [4, 5, 6], [7, 8, 9]]
  
# Creating a session, and running the graph
with tf.Session(graph=graph) as session:
      
    # run the session until it reaches node b,
    # then input an array of values into a
    operation_res = session.run(b, feed_dict={a: array1})
    print("after executing the operation: ")
    print(operation_res)

输出：

after executing the operation: [20. 30. 40. 50.]

解释：

万一出现任何错误，急切模式将被禁用。
占位符是使用 tf.placeholder() 方法创建的，该方法具有 dtype 'tf.float32'，None 表示我们没有指定任何大小。
在输入数据之前创建操作。
该操作将 10 添加到张量。
使用 tf.Session() 创建和启动会话。
Session.run 将我们创建的操作和要输入的数据作为参数并返回结果。

示例 2：

Python3

# importing packages
import tensorflow.compat.v1 as tf
  
# disabling eager mode
tf.compat.v1.disable_eager_execution()
  
# creating a tensorflow graph
graph = tf.Graph()
with graph.as_default():
    
    # creating a placeholder
    a = tf.placeholder(tf.float64, shape=(3, 3), name='tensor1')
      
    # creating an operation
    b = a ** 2
  
# array1 will be fed into 'a'
array1 = [[1, 2, 3], [4, 5, 6], [7, 8, 9]]
  
# Creating a session, and running the graph
with tf.Session(graph=graph) as session:
      
    # run the session until it reaches node b,
    # then input an array of values into a
    operation_res = session.run(b, feed_dict={a: array1})
    print("after executing the operation: ")
    print(operation_res)

输出：

after executing the operation: 
[[ 1.  4.  9.]
 [16. 25. 36.]
 [49. 64. 81.]]