TensorFlow 中的变量

TensorFlow 是一个用于高效数值计算的Python库。它是一个基础库，可用于开发机器学习和深度学习模型。 TensorFlow 是一个高级库。变量是可以通过对其执行操作来修改的状态或值。在 TensorFlow 中，变量是使用 Variable() 构造函数创建的。

Variable() 构造函数需要一个变量的初始值，它可以是任何种类或形状的张量。变量的类型和形式由其初始值定义。形状和变量一旦创建就固定了。让我们看几个如何在 TensorFlow 中创建变量的示例。

Syntax: tf.Variable(initial_value=None, trainable=None, validate_shape=True, caching_device=None, name=None, variable_def=None, dtype=None, import_scope=None, constraint=None,synchronization=tf.VariableSynchronization.AUTO, aggregation=tf.compat.v1.VariableAggregation.NONE, shape=None)

Parameters:

• initial_value: by default None. The initial value for the Variable is a Tensor, or a Python object convertible to a Tensor.
• trainable: by default None.  If True, GradientTapes will keep an eye on this variable’s usage.
• validate_shape: by default True. Allows the variable to be initialised with an unknown shape value if False. The shape of initial value must be known if True, which is the default.
• name:by default None. The variable’s optional name. Defaults to ‘Variable’ and is automatically uniquified.
• variable_def: by default None.
•  dtype: by default None. If set, initial_value will be converted to the given type. If None, either the datatype will be kept (if initial_value is a Tensor), or convert_to_tensor will decide.
•  shape: by default None. if None the shape of initial_value will be used. if any shape is specified, the variable will be assigned with that particular shape.

创建变量

tf.Variable() 构造函数用于在 TensorFlow 中创建变量。

Python3

tensor = tf.Variable([3,4])

输出：

TensorFlow 变量的维度、大小、形状和数据类型

Python3

# import packages
import tensorflow as tf
 
# create variable
tensor1 = tf.Variable([3, 4])
 
# The shape of the variable
print("The shape of the variable: ",
      tensor1.shape)
 
# The number of dimensions in the variable
print("The number of dimensions in the variable:",
      tf.rank(tensor1).numpy())
 
# The size of the variable
print("The size of the tensorflow variable:",
      tf.size(tensor1).numpy())
 
# checking the datatype of the variable
print("The datatype of the tensorflow variable is:",
      tensor1.dtype)

输出：

The shape of the variable:  (2,)
The number of dimensions in the variable: 1
The size of the tensorflow variable: 2
The datatype of the tensorflow variable is: 

分配或修改变量中的元素

我们使用 assign() 方法来修改变量。它更像是索引，然后使用 assign() 方法。还有更多方法可以分配或修改变量，例如 Variable.assign_add() 和 Variable.assign_sub())。

示例 1：

assign()：用于更新或添加新值。

Syntax: assign(value, use_locking=False, name=None, read_value=True)

parameters:

• value: The new value for this variable.
• use_locking:  locking during assignment if “true”.

Python3

import tensorflow as tf
 
tensor1 = tf.Variable([3, 4])
tensor1[1].assign(5)
tensor1

输出：

示例 2：

Syntax: assign_add(delta, use_locking=False, name=None, read_value=True)

parameters:

• delta: The value to be added to the variable(Tensor).
• use_locking: During the operation, if True, utilise locking.
• name:  name of the operation.
• read_value: If True, anything that evaluates to the modified value of the variable will be returned; if False, the assign op will be returned.

Python3

# import packages
import tensorflow as tf
 
# create variable
tensor1 = tf.Variable([3, 4])
 
# using assign_add() function
tensor1.assign_add([1, 1])
tensor1

输出：

示例 3：

Syntax: assign_sub(  delta, use_locking=False, name=None, read_value=True)

parameters:

• delta: The value to be subtracted from the variable
• use_locking: During the operation, if True, utilise locking.
• name: name of the operation.
• read_value: If True, anything that evaluates to the modified value of the variable will be returned; if False, the assign op will be returned.

Python3

# import packages
import tensorflow as tf
 
# create variable
tensor1 = tf.Variable([3, 4])
 
# using assign_sub() function
tensor1.assign_sub([1, 1])
tensor1

输出：

改变变量的形状

tf.reshape() 方法用于改变变量的形状。必须传递变量和形状。

Python3

import tensorflow as tf
 
tensor1 = tf.Variable([[1, 2, 3, 4]])
tf.reshape(tensor1, shape=(2, 2))
tensor1

输出：

更改张量的数据类型

如果我们希望变量具有特定的数据类型，我们必须在创建变量时指定 dtype。在此示例中，我们将 dtype 指定为 float。

Python3

import tensorflow as tf
 
tensor1 = tf.Variable([[1, 2, 3, 4]], dtype=float)
tensor1

输出：

变量操作

我们可以使用 TensorFlow 变量执行加法、减法、乘法、除法和更多操作。

Python3

# import packages
import tensorflow as tf
 
# create two variables
tensor1 = tf.Variable([3, 4])
tensor2 = tf.Variable([5, 6])
print("Addition of tensors", tensor1+tensor2)
 
print("Subtraction of tensors", tensor1-tensor2)
 
print("Multiplication of tensors", tensor1*tensor2)
 
print("division of tensors", tensor1/tensor2)

输出：

Addition of tensors tf.Tensor([ 8 10], shape=(2,), dtype=int32)

Subtraction of tensors tf.Tensor([-2 -2], shape=(2,), dtype=int32)

Multiplication of tensors tf.Tensor([15 24], shape=(2,), dtype=int32)

division of tensors tf.Tensor([0.6        0.66666667], shape=(2,), dtype=float64)

广播

当我们尝试对多个变量对象执行组合操作时，就像使用张量对象一样，较小的变量可以立即扩展以适应较大的变量，就像 NumPy 数组一样。当您尝试将标量变量与变量相乘时，标量会被拉伸以乘以变量的每个元素。

Python3

# import packages
import tensorflow as tf
 
# create two variables
tensor1 = tf.Variable([3, 4])
tensor2 = tf.Variable([2])
 
# broadcasting
output = tensor1*tensor2
print(output)

输出：

tf.Tensor([6 8], shape=(2,), dtype=int32)

变量的硬件选择

我们可以利用它来查看使用什么类型的设备（即处理器）来处理我们的变量。 .device 属性被使用。

Python3

# import packages
import tensorflow as tf
 
# create a variable
tensor1 = tf.Variable([3, 4])
 
print('The type of hardware variable used : '+tensor1.device)

输出：

The type of hardware variable used : /job:localhost/replica:0/task:0/device:CPU:0