keras 怎么add图层_Keras-图层

客官°小女子只卖身不卖艺 2022-12-10 01:26 205阅读 0赞

![keras 怎么add图层][keras _add]

keras 怎么add图层

# Keras-图层 **(**Keras - Layers**)** #

As learned earlier, Keras layers are the primary building block of Keras models. Each layer receives input information, do some computation and finally output the transformed information. The output of one layer will flow into the next layer as its input. Let us learn complete details about layers in this chapter.

如前所述，Keras图层是Keras模型的主要构建块。 每层接收输入信息，进行一些计算，最后输出转换后的信息。 一层的输出将流入下一层作为其输入。 让我们在本章中了解有关图层的完整详细信息。

## 介绍 **(**Introduction**)** ##

A Keras layer requires ***shape of the input (input\_shape)*** to understand the structure of the input data, ***initializer*** to set the weight for each input and finally activators to transform the output to make it non-linear. In between, constraints restricts and specify the range in which the weight of input data to be generated and regularizer will try to optimize the layer (and the model) by dynamically applying the penalties on the weights during optimization process.

Keras层需要***输入的形状(input\_shape)***来了解输入数据的结构， ***初始化器***设置每个输入的权重，最后通过激活器将输出转换为非线性。 在这两者之间，约束条件限制并指定了要生成的输入数据权重的范围，并且正则化程序将通过在优化过程中动态对权重施加惩罚来尝试优化图层(和模型)。

To summarise, Keras layer requires below minimum details to create a complete layer.

总而言之，Keras图层需要以下最低限度的细节才能创建完整的图层。

*  Shape of the input data
    
    输入数据的形状
 *  Number of neurons / units in the layer
    
    层中神经元/单位的数量
 *  Initializers
    
    初始化器
 *  Regularizers
    
    正则化器
 *  Constraints
    
    约束条件
 *  Activations
    
    激活方式

Let us understand the basic concept in the next chapter. Before understanding the basic concept, let us create a simple Keras layer using Sequential model API to get the idea of how Keras model and layer works.

让我们在下一章中了解基本概念。 在理解基本概念之前，让我们使用顺序模型API创建一个简单的Keras层，以了解Keras模型和层如何工作。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    from keras import initializers 
    from keras import regularizers 
    from keras import constraints 
    
    model = Sequential() 
    
    model.add(Dense(32, input_shape=(16,), kernel_initializer = 'he_uniform', 
       kernel_regularizer = None, kernel_constraint = 'MaxNorm', activation = 'relu')) 
    model.add(Dense(16, activation = 'relu')) 
    model.add(Dense(8))

where,

哪里，

*  **Line 1-5** imports the necessary modules.
    
    **1-5行**导入了必要的模块。
 *  **Line 7** creates a new model using Sequential API.
    
    **第7行**使用Sequential API创建新模型。
 *  **Line 9** creates a new ***Dense*** layer and add it into the model. ***Dense*** is an entry level layer provided by Keras, which accepts the number of neurons or units (32) as its required parameter. If the layer is first layer, then we need to provide **Input Shape, (16,)** as well. Otherwise, the output of the previous layer will be used as input of the next layer. All other parameters are optional.
    
    **第9行**创建一个新的***Dense***层并将其添加到模型中。 ***密集***是Keras提供的入门级层，它接受神经元或单位数(32)作为其必需参数。 如果该层是第一层，那么我们还需要提供**输入形状(16)** 。 否则，上一层的输出将用作下一层的输入。 所有其他参数都是可选的。
    
     *  First parameter represents the number of units (neurons).
        
        第一个参数代表单位数(神经元)。
     *  ***input\_shape*** represent the shape of input data.
        
        ***input\_shape***代表输入数据的形状。
     *  ***kernel\_initializer*** represent initializer to be used. ***he\_uniform*** function is set as value.
        
        ***kernel\_initializer***表示要使用的初始化程序。 ***he\_uniform***函数设置为值。
     *  ***kernel\_regularizer*** represent **regularizer** to be used. None is set as value.
        
        ***kernel\_regularizer***表示要使用的**正则化**器。 没有设置任何值。
     *  ***kernel\_constraint*** represent constraint to be used. ***MaxNorm*** function is set as value.
        
        ***kernel\_constraint***表示要使用的约束。 ***MaxNorm***函数设置为值。
     *  ***activation*** represent activation to be used. relu function is set as value.
        
        ***激活***表示要使用的激活。 将relu功能设置为值。
 *  **Line 10** creates second ***Dense*** layer with 16 units and set ***relu*** as the activation function.
    
    **第10行**使用16个单位创建第二个***Dense***层，并将***relu***设置为激活功能。
 *  **Line 11** creates final Dense layer with 8 units.
    
    **第11行**使用8个单位创建最终的密集层。

## 图层的基本概念 **(**Basic Concept of Layers**)** ##

Let us understand the basic concept of layer as well as how Keras supports each concept.

让我们了解层的基本概念以及Keras如何支持每个概念。

### 输入形状 **(**Input shape**)** ###

In machine learning, all type of input data like text, images or videos will be first converted into array of numbers and then feed into the algorithm. Input numbers may be single dimensional array, two dimensional array (matrix) or multi-dimensional array. We can specify the dimensional information using **shape**, a tuple of integers. For example, **(4,2)** represent matrix with four rows and two columns.

在机器学习中，所有类型的输入数据(例如文本，图像或视频)都将首先转换为数字数组，然后输入算法。 输入数字可以是一维数组，二维数组(矩阵)或多维数组。 我们可以使用**shape** (整数元组)指定尺寸信息。 例如， **(4,2)**表示具有四行两列的矩阵。

>>> import numpy as np 
    >>> shape = (4, 2) 
    >>> input = np.zeros(shape) 
    >>> print(input) 
    [
       [0. 0.] 
       [0. 0.] 
       [0. 0.] 
       [0. 0.]
    ] 
    >>>

Similarly, **(3,4,2)** three dimensional matrix having three collections of 4x2 matrix (two rows and four columns).

类似地， **(3,4,2)**具有三个4x2矩阵集合(两行四列)的三维矩阵。

>>> import numpy as np 
    >>> shape = (3, 4, 2) 
    >>> input = np.zeros(shape) 
    >>> print(input)
    [
       [[0. 0.] [0. 0.] [0. 0.] [0. 0.]] 
       [[0. 0.] [0. 0.] [0. 0.] [0. 0.]] 
       [[0. 0.] [0. 0.] [0. 0.] [0. 0.]]
    ]
    >>>

To create the first layer of the model (or input layer of the model), shape of the input data should be specified.

要创建模型的第一层(或模型的输入层)，应指定输入数据的形状。

## 初始化器 **(**Initializers**)** ##

In Machine Learning, weight will be assigned to all input data. ***Initializers*** module provides different functions to set these initial weight. Some of the ***Keras Initializer*** function are as follows −

在机器学习中，权重将分配给所有输入数据。 ***Initializers***模块提供了不同的功能来设置这些初始权重。 ***Keras初始化器的***一些功能如下-

### 零点 **(**Zeros**)** ###

Generates **0** for all input data.

为所有输入数据生成**0** 。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    from keras import initializers 
    
    my_init = initializers.Zeros() 
    model = Sequential() 
    model.add(Dense(512, activation = 'relu', input_shape = (784,), 
       kernel_initializer = my_init))

Where, ***kernel\_initializer*** represent the initializer for kernel of the model.

其中， ***kernel\_initializer***表示模型内核的初始化程序。

### 那些 **(**Ones**)** ###

Generates **1** for all input data.

为所有输入数据生成**1** 。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    from keras import initializers 
    
    my_init = initializers.Ones() 
    model.add(Dense(512, activation = 'relu', input_shape = (784,), 
       kernel_initializer = my_init))

### 不变 **(**Constant**)** ###

Generates a constant value (say, **5**) specified by the user for all input data.

生成用户为所有输入数据指定的常数值(例如**5** )。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    from keras import initializers 
    
    my_init = initializers.Constant(value = 0) model.add(
       Dense(512, activation = 'relu', input_shape = (784,), kernel_initializer = my_init)
    )

where, **value** represent the constant value

其中， **值**代表常数

### 随机法线 **(**RandomNormal**)** ###

Generates value using normal distribution of input data.

使用输入数据的正态分布生成值。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    from keras import initializers 
    
    my_init = initializers.RandomNormal(mean=0.0, 
    stddev = 0.05, seed = None) 
    model.add(Dense(512, activation = 'relu', input_shape = (784,), 
       kernel_initializer = my_init))

where,

哪里，

*  ***mean*** represent the mean of the random values to generate
    
    ***平均值***表示要生成的随机值的平均值
 *  ***stddev*** represent the standard deviation of the random values to generate
    
    ***stddev***表示要生成的随机值的标准偏差
 *  ***seed*** represent the values to generate random number
    
    ***种子***代表生成随机数的值

### 随机统一 **(**RandomUniform**)** ###

Generates value using uniform distribution of input data.

使用输入数据的均匀分布生成值。

from keras import initializers 
    
    my_init = initializers.RandomUniform(minval = -0.05, maxval = 0.05, seed = None) 
    model.add(Dense(512, activation = 'relu', input_shape = (784,), 
       kernel_initializer = my_init))

where,

哪里，

*  ***minval*** represent the lower bound of the random values to generate
    
    ***minval***表示要生成的随机值的下限
 *  ***maxval*** represent the upper bound of the random values to generate
    
    ***maxval***表示要生成的随机值的上限

### 截断法线 **(**TruncatedNormal**)** ###

Generates value using truncated normal distribution of input data.

使用输入数据的截断正态分布生成值。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    from keras import initializers 
    
    my_init = initializers.TruncatedNormal(mean = 0.0, stddev = 0.05, seed = None
    model.add(Dense(512, activation = 'relu', input_shape = (784,), 
       kernel_initializer = my_init))

### 方差缩放 **(**VarianceScaling**)** ###

Generates value based on the input shape and output shape of the layer along with the specified scale.

根据图层的输入形状和输出形状以及指定的比例尺生成值。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    from keras import initializers 
    
    my_init = initializers.VarianceScaling(
       scale = 1.0, mode = 'fan_in', distribution = 'normal', seed = None) 
    model.add(Dense(512, activation = 'relu', input_shape = (784,), 
       skernel_initializer = my_init))

where,

哪里，

*  **scale** represent the scaling factor
    
    **比例**代表比例因子
 *  **mode** represent any one of **fan\_in, fan\_out** and **fan\_avg** values
    
    **模式**代表**fan\_in，fan\_out**和**fan\_avg**值中的任何一个
 *  **distribution** represent either of **normal** or **uniform**
    
    **分布**代表**正态分布**或**均匀** **分布**

### 方差缩放 **(**VarianceScaling**)** ###

It finds the ***stddev*** value for normal distribution using below formula and then find the weights using normal distribution,

它使用以下公式找到正态分布的***stddev***值，然后使用正态分布找到权重，

stddev = sqrt(scale / n)

where **n** represent,

**n**代表

*  number of input units for mode = fan\_in
    
    模式的输入单元数= fan\_in
 *  number of out units for mode = fan\_out
    
    模式的输出单元数= fan\_out
 *  average number of input and output units for mode = fan\_avg
    
    模式的平均输入和输出单元数= fan\_avg

Similarly, it finds the *limit* for uniform distribution using below formula and then find the weights using uniform distribution,

同样，它使用以下公式找到均匀分布的*极限* ，然后使用均匀分布找到权重，

limit = sqrt(3 * scale / n)

### lecun\_normal **(**lecun\_normal**)** ###

Generates value using lecun normal distribution of input data.

使用lecun输入数据的正态分布生成值。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    from keras import initializers 
    
    my_init = initializers.RandomUniform(minval = -0.05, maxval = 0.05, seed = None)
    model.add(Dense(512, activation = 'relu', input_shape = (784,), 
       kernel_initializer = my_init))

It finds the ***stddev*** using the below formula and then apply normal distribution

它使用以下公式找到***stddev*** ，然后应用正态分布

stddev = sqrt(1 / fan_in)

where, ***fan\_in*** represent the number of input units.

其中， ***fan\_in***代表输入单元的数量。

### lecun\_uniform **(**lecun\_uniform**)** ###

Generates value using lecun uniform distribution of input data.

使用lecun输入数据的均匀分布来生成值。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    from keras import initializers 
    
    my_init = initializers.lecun_uniform(seed = None) 
    model.add(Dense(512, activation = 'relu', input_shape = (784,), 
       kernel_initializer = my_init))

It finds the ***limit*** using the below formula and then apply uniform distribution

它使用以下公式找到***极限*** ，然后应用均匀分布

limit = sqrt(3 / fan_in)

where,

哪里，

*  ***fan\_in*** represents the number of input units
    
    ***fan\_in***代表输入单元的数量
 *  ***fan\_out*** represents the number of output units
    
    ***fan\_out***代表输出单元的数量

### glorot\_normal **(**glorot\_normal**)** ###

Generates value using glorot normal distribution of input data.

使用输入数据的正态分布生成值。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    from keras import initializers 
    
    my_init = initializers.glorot_normal(seed=None) model.add(
       Dense(512, activation = 'relu', input_shape = (784,), kernel_initializer = my_init)
    )

It finds the ***stddev*** using the below formula and then apply normal distribution

它使用以下公式找到***stddev*** ，然后应用正态分布

stddev = sqrt(2 / (fan_in + fan_out))

where,

哪里，

### glorot\_uniform **(**glorot\_uniform**)** ###

Generates value using glorot uniform distribution of input data.

使用输入数据的全局均匀分布来生成值。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    from keras import initializers 
    
    my_init = initializers.glorot_uniform(seed = None) 
    model.add(Dense(512, activation = 'relu', input_shape = (784,), 
       kernel_initializer = my_init))

It finds the ***limit*** using the below formula and then apply uniform distribution

它使用以下公式找到***极限*** ，然后应用均匀分布

limit = sqrt(6 / (fan_in + fan_out))

where,

哪里，

*  ***fan\_in*** represent the number of input units.
    
    ***fan\_in***代表输入单元的数量。
 *  ***fan\_out*** represents the number of output units
    
    ***fan\_out***代表输出单元的数量

### 他正常 **(**he\_normal**)** ###

Generates value using he normal distribution of input data.

使用输入数据的正态分布生成值。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    from keras import initializers 
    
    my_init = initializers.RandomUniform(minval = -0.05, maxval = 0.05, seed = None) 
    model.add(Dense(512, activation = 'relu', input_shape = (784,), 
       kernel_initializer = my_init))

It finds the *stddev* using the below formula and then apply normal distribution.

它使用以下公式找到*stddev* ，然后应用正态分布。

stddev = sqrt(2 / fan_in)

where, ***fan\_in*** represent the number of input units.

其中， ***fan\_in***代表输入单元的数量。

### he\_uniform **(**he\_uniform**)** ###

Generates value using he uniform distribution of input data.

使用输入数据的均匀分布来产生价值。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    from keras import initializers 
    
    my_init = initializers.he_normal(seed = None) 
    model.add(Dense(512, activation = 'relu', input_shape = (784,), 
       kernel_initializer = my_init))

It finds the ***limit*** using the below formula and then apply uniform distribution.

它使用以下公式找到***极限*** ，然后应用均匀分布。

limit = sqrt(6 / fan_in)

where, ***fan\_in*** represent the number of input units.

其中， ***fan\_in***代表输入单元的数量。

### 正交的 **(**Orthogonal**)** ###

Generates a random orthogonal matrix.

生成随机正交矩阵。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    from keras import initializers 
    
    my_init = initializers.Orthogonal(gain = 1.0, seed = None) 
    model.add(Dense(512, activation = 'relu', input_shape = (784,), 
       kernel_initializer = my_init))

where, ***gain*** represent the multiplication factor of the matrix.

其中， ***增益***代表矩阵的乘法因子。

### 身分识别 **(**Identity**)** ###

Generates identity matrix.

生成单位矩阵。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    from keras import initializers 
    
    my_init = initializers.Identity(gain = 1.0) model.add(
       Dense(512, activation = 'relu', input_shape = (784,), kernel_initializer = my_init)
    )

## 约束条件 **(**Constraints**)** ##

In machine learning, a constraint will be set on the parameter (weight) during optimization phase. <>Constraints module provides different functions to set the constraint on the layer. Some of the constraint functions are as follows.

在机器学习中，将在优化阶段对参数(权重)设置约束。 <>约束模块提供了不同的功能来设置图层上的约束。 一些约束函数如下。

### 非负 **(**NonNeg**)** ###

Constrains weights to be non-negative.

将权重限制为非负数。

where, ***kernel\_constraint*** represent the constraint to be used in the layer.

其中， ***kernel\_constraint***表示要在层中使用的约束。

### 单位范数 **(**UnitNorm**)** ###

Constrains weights to be unit norm.

将权重约束为单位范数。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    from keras import constraints 
    
    my_constrain = constraints.UnitNorm(axis = 0) 
    model = Sequential() 
    model.add(Dense(512, activation = 'relu', input_shape = (784,), 
       kernel_constraint = my_constrain))

### 最大规范 **(**MaxNorm**)** ###

Constrains weight to norm less than or equals to the given value.

将权重限制为小于或等于给定值的范数。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    from keras import constraints 
    
    my_constrain = constraints.MaxNorm(max_value = 2, axis = 0) 
    model = Sequential() 
    model.add(Dense(512, activation = 'relu', input_shape = (784,), 
       kernel_constraint = my_constrain))

where,

哪里，

*  ***max\_value*** represent the upper bound
    
    ***max\_value***代表上限
 *  *axis* represent the dimension in which the constraint to be applied. e.g. in Shape (2,3,4) axis 0 denotes first dimension, 1 denotes second dimension and 2 denotes third dimension
    
    *轴*表示要应用约束的尺寸。 例如在形状(2,3,4)轴中0表示第一尺寸，1表示第二尺寸，2表示第三尺寸

### 最小最大范数 **(**MinMaxNorm**)** ###

Constrains weights to be norm between specified minimum and maximum values.

将权重限制在指定的最小值和最大值之间。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    from keras import constraints 
    
    my_constrain = constraints.MinMaxNorm(min_value = 0.0, max_value = 1.0, rate = 1.0, axis = 0) 
    model = Sequential() 
    model.add(Dense(512, activation = 'relu', input_shape = (784,), 
       kernel_constraint = my_constrain))

where, ***rate*** represent the rate at which the weight constrain is applied.

其中， ***rate***表示施加重量约束的速率。

## 正则化器 **(**Regularizers**)** ##

In machine learning, regularizers are used in the optimization phase. It applies some penalties on the layer parameter during optimization. Keras regularization module provides below functions to set penalties on the layer. Regularization applies per-layer basis only.

在机器学习中，正则化器用于优化阶段。 在优化过程中会对层参数施加一些惩罚。 Keras正则化模块提供以下功能来设置层上的惩罚。 正则化仅适用于每层。

### L1正则器 **(**L1 Regularizer**)** ###

It provides L1 based regularization.

它提供基于L1的正则化。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    from keras import regularizers 
    
    my_regularizer = regularizers.l1(0.) 
    model = Sequential() 
    model.add(Dense(512, activation = 'relu', input_shape = (784,), 
       kernel_regularizer = my_regularizer))

where, ***kernel\_regularizer*** represent the rate at which the weight constrain is applied.

其中， ***kernel\_regularizer***表示施加权重约束的速率。

### L2调节器 **(**L2 Regularizer**)** ###

It provides L2 based regularization.

它提供基于L2的正则化。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    from keras import regularizers 
    
    my_regularizer = regularizers.l2(0.) 
    model = Sequential() 
    model.add(Dense(512, activation = 'relu', input_shape = (784,), 
       kernel_regularizer = my_regularizer))

### L1和L2正则器 **(**L1 and L2 Regularizer**)** ###

It provides both L1 and L2 based regularization.

它提供了基于L1和L2的正则化。

## 激活方式 **(**Activations**)** ##

In machine learning, activation function is a special function used to find whether a specific neuron is activated or not. Basically, the activation function does a nonlinear transformation of the input data and thus enable the neurons to learn better. Output of a neuron depends on the activation function.

在机器学习中，激活功能是一种特殊功能，用于查找特定神经元是否被激活。 基本上，激活函数对输入数据进行非线性转换，从而使神经元学习得更好。 神经元的输出取决于激活功能。

As you recall the concept of single perception, the output of a perceptron (neuron) is simply the result of the activation function, which accepts the summation of all input multiplied with its corresponding weight plus overall bias, if any available.

当您回想起单感知的概念时，感知器(神经元)的输出只是激活函数的结果，该函数接受所有输入的总和乘以其相应的权重加上总体偏差(如果有)。

result = Activation(SUMOF(input * weight) + bias)

So, activation function plays an important role in the successful learning of the model. Keras provides a lot of activation function in the activations module. Let us learn all the activations available in the module.

因此，激活函数在模型的成功学习中起着重要作用。 Keras在激活模块中提供了很多激活功能。 让我们学习该模块中所有可用的激活。

### 线性的 **(**linear**)** ###

Applies Linear function. Does nothing.

应用线性函数。 什么也没做。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    
    model = Sequential() 
    model.add(Dense(512, activation = 'linear', input_shape = (784,)))

Where, ***activation*** refers the activation function of the layer. It can be specified simply by the name of the function and the layer will use corresponding activators.

其中， ***激活***是指该层的激活功能。 可以简单地通过函数名称指定它，并且该层将使用相应的激活器。

### lu **(**elu**)** ###

Applies Exponential linear unit.

应用指数线性单位。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    
    model = Sequential() 
    model.add(Dense(512, activation = 'elu', input_shape = (784,)))

### selu **(**selu**)** ###

Applies Scaled exponential linear unit.

应用比例指数线性单位。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    
    model = Sequential() 
    model.add(Dense(512, activation = 'selu', input_shape = (784,)))

### relu **(**relu**)** ###

Applies Rectified Linear Unit.

应用整流线性单位。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    
    model = Sequential() 
    model.add(Dense(512, activation = 'relu', input_shape = (784,)))

### softmax **(**softmax**)** ###

Applies Softmax function.

应用Softmax功能。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    
    model = Sequential() 
    model.add(Dense(512, activation = 'softmax', input_shape = (784,)))

### softplus **(**softplus**)** ###

Applies Softplus function.

应用Softplus功能。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    
    model = Sequential() 
    model.add(Dense(512, activation = 'softplus', input_shape = (784,)))

### softsign **(**softsign**)** ###

Applies Softsign function.

应用软件签名功能。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    
    model = Sequential() 
    model.add(Dense(512, activation = 'softsign', input_shape = (784,)))

### tanh **(**tanh**)** ###

Applies Hyperbolic tangent function.

应用双曲正切函数。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    model = Sequential() 
    model.add(Dense(512, activation = 'tanh', input_shape = (784,)))

### sigmod **(**sigmoid**)** ###

Applies Sigmoid function.

应用Sigmoid函数。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    
    model = Sequential() 
    model.add(Dense(512, activation = 'sigmoid', input_shape = (784,)))

### 硬sigmod **(**hard\_sigmoid**)** ###

Applies Hard Sigmoid function.

应用硬Sigmoid函数。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    
    model = Sequential() 
    model.add(Dense(512, activation = 'hard_sigmoid', input_shape = (784,)))

### 指数的 **(**exponential**)** ###

Applies exponential function.

应用指数函数。

from keras.models import Sequential 
    from keras.layers import Activation, Dense 
    
    model = Sequential() 
    model.add(Dense(512, activation = 'exponential', input_shape = (784,)))

<table> 
 <tbody> 
 <tr> 
 <th>Sr.No</th> 
 <th>Layers &amp; Description</th> 
 </tr> 
 <tr> 
 <td>1</td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_dense_layer.htm" rel="nofollow">Dense Layer</a> Dense layer is the regular deeply connected neural network layer. </td> 
 </tr> 
 <tr> 
 <td>2</td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_dropout_layers.htm" rel="nofollow">Dropout Layers</a> Dropout is one of the important concept in the machine learning. </td> 
 </tr> 
 <tr> 
 <td>3</td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_flatten_layers.htm" rel="nofollow">Flatten Layers</a> Flatten is used to flatten the input. </td> 
 </tr> 
 <tr> 
 <td>4</td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_reshape_layers.htm" rel="nofollow">Reshape Layers</a> Reshape is used to change the shape of the input. </td> 
 </tr> 
 <tr> 
 <td>5</td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_permute_layers.htm" rel="nofollow">Permute Layers</a> Permute is also used to change the shape of the input using pattern. </td> 
 </tr> 
 <tr> 
 <td>6</td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_repeatvector_layers.htm" rel="nofollow">RepeatVector Layers</a> RepeatVector is used to repeat the input for set number, n of times. </td> 
 </tr> 
 <tr> 
 <td>7</td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_lambda_layers.htm" rel="nofollow">Lambda Layers</a> Lambda is used to transform the input data using an expression or function. </td> 
 </tr> 
 <tr> 
 <td>8</td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_convolution_layers.htm" rel="nofollow">Convolution Layers</a> Keras contains a lot of layers for creating Convolution based ANN, popularly called as Convolution Neural Network (CNN). </td> 
 </tr> 
 <tr> 
 <td>9</td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_pooling_layer.htm" rel="nofollow">Pooling Layer</a> It is used to perform max pooling operations on temporal data. </td> 
 </tr> 
 <tr> 
 <td>10</td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_locally_connected_layer.htm" rel="nofollow">Locally connected layer</a> Locally connected layers are similar to Conv1D layer but the difference is Conv1D layer weights are shared but here weights are unshared. </td> 
 </tr> 
 <tr> 
 <td>11</td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_merge_layer.htm" rel="nofollow">Merge Layer</a> It is used to merge a list of inputs. </td> 
 </tr> 
 <tr> 
 <td>12</td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_embedding_layer.htm" rel="nofollow">Embedding Layer</a> It performs embedding operations in input layer. </td> 
 </tr> 
 </tbody> 
</table>

<table> 
 <tbody> 
 <tr> 
 <th> 序号 </th> 
 <th> 图层和说明 </th> 
 </tr> 
 <tr> 
 <td> 1个 </td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_dense_layer.htm" rel="nofollow">致密层</a> 密集层是规则的深度连接神经网络层。 </td> 
 </tr> 
 <tr> 
 <td> 2 </td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_dropout_layers.htm" rel="nofollow">辍学层</a> 辍学是机器学习中的重要概念之一。 </td> 
 </tr> 
 <tr> 
 <td> 3 </td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_flatten_layers.htm" rel="nofollow">展平层</a> 展平被用来压平的输入端。 </td> 
 </tr> 
 <tr> 
 <td> 4 </td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_reshape_layers.htm" rel="nofollow">重塑图层</a> 重塑是用来改变所述输入的形状。 </td> 
 </tr> 
 <tr> 
 <td> 5 </td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_permute_layers.htm" rel="nofollow">置换层</a> Permute还用于使用模式更改输入的形状。 </td> 
 </tr> 
 <tr> 
 <td> 6 </td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_repeatvector_layers.htm" rel="nofollow">RepeatVector图层</a> RepeatVector用于将输入重复设置次数n次。 </td> 
 </tr> 
 <tr> 
 <td> 7 </td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_lambda_layers.htm" rel="nofollow">Lambda图层</a> Lambda用于使用表达式或函数来转换输入数据。 </td> 
 </tr> 
 <tr> 
 <td> 8 </td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_convolution_layers.htm" rel="nofollow">卷积层</a> Keras包含用于创建基于卷积的ANN的许多层，通常被称为卷积神经网络(CNN) 。 </td> 
 </tr> 
 <tr> 
 <td> 9 </td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_pooling_layer.htm" rel="nofollow">池化层</a> 它用于对时间数据执行最大池化操作。 </td> 
 </tr> 
 <tr> 
 <td> 10 </td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_locally_connected_layer.htm" rel="nofollow">本地连接层</a> 本地连接的图层与Conv1D图层相似，但区别在于Conv1D图层的权重是共享的，但此处的权重是不共享的。 </td> 
 </tr> 
 <tr> 
 <td> 11 </td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_merge_layer.htm" rel="nofollow">合并层</a> 它用于合并输入列表。 </td> 
 </tr> 
 <tr> 
 <td> 12 </td> 
 <td> <a href="https://www.tutorialspoint.com/keras/keras_embedding_layer.htm" rel="nofollow">嵌入层</a> 它在输入层中执行嵌入操作。 </td> 
 </tr> 
 </tbody> 
</table>

> 翻译自: [https://www.tutorialspoint.com/keras/keras\_layers.htm][https_www.tutorialspoint.com_keras_keras_layers.htm]

keras 怎么add图层

[keras _add]: /images/20221123/851bf844ca92417da5de1d1161a846b2.png
[https_www.tutorialspoint.com_keras_keras_layers.htm]: https://www.tutorialspoint.com/keras/keras_layers.htm