tensorflow-条件循环控制(1)

TensorFlow 提供了一些操作和类，您可以使用它们来控制操作的执行，并向图表添加条件依赖项。

tf.identity
tf.tuple
tf.group
tf.no_op
tf.count_up_to
tf.cond
tf.case
tf.while_loop

tf.identity

 
tf.identity(
    input,
    name=None
)

返回和输入大小与内容一致的tensor

参数:

input: 一个Tensor.
name: 操作名字（可选）
这个方法的源码：

def identity(input, name=None):  # pylint: disable=redefined-builtin
  r"""Return a tensor with the same shape and contents as input.
  Args:
    input: A `Tensor`.
    name: A name for the operation (optional).
  Returns:
    A `Tensor`. Has the same type as `input`.
  """
  if context.executing_eagerly():
    input = ops.convert_to_tensor(input)
    in_device = input.device

    # TODO(ashankar): Does 'identity' need to invoke execution callbacks?
    context_device = context.context().device_name
    if not context_device:
      context_device = "/job:localhost/replica:0/task:0/device:CPU:0"
    if context_device != in_device:
      return input._copy()  # pylint: disable=protected-access
    return input
  else:
    return gen_array_ops.identity(input, name=name)

从源码可看出，input的device会与context.context()相比较，如果相同，则直接返回input，如果不相同则返回disable=protected-access权限的input的copy。

#!/usr/bin/env python2
# -*- coding: utf-8 -*-
"""
Created on Mon Aug 27 11:16:32 2018
@author: myhaspl
"""

import tensorflow as tf
x1 = tf.constant(2.)
x2 = [1,2,3]
y1=tf.identity(x1,name="my_y1")
y2=tf.identity(x2,name="my_y")
sess=tf.Session()
with sess: 
    print sess.run(y1)
    print sess.run(y2)
2.0
[1 2 3]