ATCN-RF的风力预测.zip

import inspect from typing import List from tensorflow import keras from keras import backend as K from keras import Model, Input, optimizers from keras import layers from keras.layers import Activation, SpatialDropout1D, Lambda from keras.layers import Layer, Conv1D, Dense, BatchNormalization, LayerNormalization def is_power_of_two(num: int): return num != 0 and ((num & (num - 1)) == 0)#(num&(num-1)==0)num是2的指数 def adjust_dilations(dilations: list):#自适应的膨胀系数 if all([is_power_of_two(i) for i in dilations]): return dilations else: new_dilations = [2 ** i for i in dilations] return new_dilations #定义残差块 class ResidualBlock(Layer): def __init__(self, dilation_rate: int, nb_filters: int, kernel_size: int, padding: str, activation: str = 'relu', dropout_rate: float = 0, kernel_initializer: str = 'he_normal', use_batch_norm: bool = False, use_layer_norm: bool = False, **kwargs): """Defines the residual block for the WaveNet TCN,#为TCN网络定义残差块 Args:#命令行参数个数 x: The previous layer in the model模型中的上一层 training: boolean indicating whether the layer should behave in training mode or in inference mode指示图层在训练模式或推理模式下的行为是具有布尔值 dilation_rate: The dilation power of 2 we are using for this residual block我们用于此残余块的2的膨胀率 nb_filters: The number of convolutional filters to use in this block要在此块中使用的卷积滤波器的数量 kernel_size: The size of the convolutional kernel卷积核的大小 padding: The padding used in the convolutional layers, 'same' or 'causal'. 填充：卷积层中使用的填充，"相同"或"因果"。 activation: The final activation used in o = Activation(x + F(x)) dropout_rate: Float between 0 and 1. Fraction of the input units to drop.浮动在0到1之间。要丢弃的输入单位的分数。 kernel_initializer: Initializer for the kernel weights matrix (Conv1D).内核权重矩阵 （Conv1D） 的初始值设定项。 use_batch_norm: Whether to use batch normalization in the residual layers or not.是否在残差图层中使用批量归一化。 use_layer_norm: Whether to use layer normalization in the residual layers or not.是否在残差图层中使用图层归一化。 kwargs: Any initializers for Layer class.Layer 类的任何初始值设定项。 """ self.dilation_rate = dilation_rate self.nb_filters = nb_filters#过滤器即卷积核尺寸 self.kernel_size = kernel_size self.padding = padding self.activation = activation self.dropout_rate = dropout_rate self.use_batch_norm = use_batch_norm self.use_layer_norm = use_layer_norm self.kernel_initializer = kernel_initializer self.layers = [] self.layers_outputs = [] self.shape_match_conv = None self.res_output_shape = None self.final_activation = None super(ResidualBlock, self).__init__(**kwargs) #定义一个添加层的函数可以很容易的添加神经层,为之后的添加省下不少时间 def _add_and_activate_layer(self, layer): """Helper function for building layer Args: layer: Appends layer to internal layer list and builds it based on the current output shape of ResidualBlocK. Updates current output shape. #添加图层到内部图层列表，并根据当前的输出构建它残差块（ResidualBlocK）的形状，更新当前输出形状。 """ self.layers.append(layer)#将定义好的网络层添加进模型中 self.layers[-1].build(self.res_output_shape) self.res_output_shape = self.layers[-1].compute_output_shape(self.res_output_shape) def build(self, input_shape):#build() 可自定义网络的权重的维度，可以根据输入来指定权重的维度 with K.name_scope(self.name): # name scope used to make sure weights get unique names名称作用域，用于确保权重获得唯一的名称 self.layers = [] self.res_output_shape = input_shape for k in range(2): name = 'conv1D_{}'.format(k) #一维膨胀因果卷积网络，format(k)传递位置参数，conv1D_{k}。 with K.name_scope(name): # name scope used to make sure weights get unique names self._add_and_activate_layer(Conv1D(filters=self.nb_filters, kernel_size=self.kernel_size, dilation_rate=self.dilation_rate, padding=self.padding, name=name, kernel_initializer=self.kernel_initializer)) with K.name_scope('norm_{}'.format(k)):#正则化 if self.use_batch_norm: self._add_and_activate_layer(BatchNormalization()) elif self.use_layer_norm: self._add_and_activate_layer(LayerNormalization()) self._add_and_activate_layer(Activation('relu'))#激活函数 self._add_and_activate_layer(SpatialDropout1D(rate=self.dropout_rate)) if self.nb_filters != input_shape[-1]: # 1x1 conv to match the shapes (channel dimension).# 1x1转换以匹配形状(通道尺寸)，即1x1全卷积连接 name = 'matching_conv1D' with K.name_scope(name): # make and build this layer separately because it directly uses input_shape，单独创建和构建这个层，因为它直接使用input_shape self.shape_match_conv = Conv1D(filters=self.nb_filters, kernel_size=1, padding='same',#padding='SAME'意味着卷积后的输出与输入size保持一致。 name=name, kernel_initializer=self.kernel_initializer) else: name = 'matching_identity' self.shape_match_conv = Lambda(lambda x: x, name=name) #lambda定义一个函数，x=x with K.name_scope(name): self.shape_match_conv.build(input_shape) self.res_output_shape = self.shape_match_conv.compute_output_shape(input_shape)#残差块的输出形状=卷积计算的输出形状（） self.final_activation = Activation(self.activation)#激活函数层 self.final_activation.build(self.res_output_shape) # probably isn't necessary # this is done to force Keras to add the layers in the list to self._layers #这样做是为了迫使Keras将列表中的图层添加到self._layers for layer in self.layers: self.__setattr__(layer.name, layer) self.__setattr__(self.shape_match_conv.name, self.shape_match_conv) self.__setattr__(self.final_activation.name, self.final_activation) super(ResidualBlock, self).build(input_shape) # done to make sure self.built is set True #super(Net, self).__init__()指子类继承了父类的所有属性和方法，父类属性自然会用父类方法来进行初始化 def