像学习编程一样学习深度学习模型开发

作为一个程序员，我们可以像学习编程一样学习深度学习模型开发。我们以 Keras 为例来说明。

我们可以用 5 步 + 4 种基本元素 + 9 种基本层结构，这 5-4-9 模型来总结。

5步法：

1. 构造网络模型2. 编译模型3. 训练模型4. 评估模型5. 使用模型进行预测

4种基本元素：

1. 网络结构：由10种基本层结构和其他层结构组成2. 激活函数：如relu, softmax。口诀: 最后输出用softmax，其余基本都用relu3. 损失函数：categorical_crossentropy多分类对数损失，binary_crossentropy对数损失，mean_squared_error平均方差损失, mean_absolute_error平均绝对值损失4. 优化器：如SGD随机梯度下降, RMSProp, Adagrad, Adam, Adadelta等

9种基本层模型

包括3种主模型：

1. 全连接层Dense2. 卷积层：如conv1d, conv2d3. 循环层：如lstm, gru

3种辅助层：1. Activation层2. Dropout层3. 池化层

3种异构网络互联层：

1. 嵌入层：用于第一层，输入数据到其他网络的转换2. Flatten层：用于卷积层到全连接层之间的过渡3. Permute层：用于RNN与CNN之间的接口

我们通过一张图来理解下它们之间的关系

▌五步法

五步法是用深度学习来解决问题的五个步骤：

1. 构造网络模型2. 编译模型3. 训练模型4. 评估模型5. 使用模型进行预测

在这五步之中，其实关键的步骤主要只有第一步，这一步确定了，后面的参数都可以根据它来设置。

过程化方法构造网络模型

我们先学习最容易理解的，过程化方法构造网络模型的过程。

Keras中提供了Sequential容器来实现过程式构造。只要用Sequential的add方法把层结构加进来就可以了。10种基本层结构我们会在后面详细讲。

例：

from keras.models import Sequentialfrom keras.layers import Dense, Activationmodel = Sequential()model.add(Dense(units=64, input_dim=100))model.add(Activation("relu"))model.add(Dense(units=10))model.add(Activation("softmax"))

对于什么样的问题构造什么样的层结构，我们会在后面的例子中介绍。

编译模型

模型构造好之后，下一步就可以调用Sequential的compile方法来编译它。

model.compile(loss='categorical_crossentropy', optimizer='sgd', metrics=['accuracy'])

编译时需要指定两个基本元素：loss是损失函数，optimizer是优化函数。

如果只想用最基本的功能，只要指定字符串的名字就可以了。如果想配置更多的参数，调用相应的类来生成对象。例：我们想为随机梯度下降配上Nesterov动量，就生成一个SGD的对象就好了：

from keras.optimizers import SGDmodel.compile(loss='categorical_crossentropy', optimizer=SGD(lr=0.01, momentum=0.9, nesterov=True))

lr是学习率，learning rate。

训练模型

调用fit函数，将输出的值X，打好标签的值y，epochs训练轮数，batch_size批次大小设置一下就可以了：

model.fit(x_train, y_train, epochs=5, batch_size=32)

评估模型

模型训练的好不好，训练数据不算数，需要用测试数据来评估一下：

loss_and_metrics = model.evaluate(x_test, y_test, batch_size=128)

用模型来预测

一切训练的目的是在于预测：

classes = model.predict(x_test, batch_size=128)

▌4种基本元素

网络结构

主要用后面的层结构来拼装。网络结构如何设计呢? 可以参考论文，比如这篇中不管是左边的19层的VGG-19，还是右边34层的resnet，只要按图去实现就好了。

激活函数

对于多分类的情况，最后一层是softmax。

其它深度学习层中多用relu。

二分类可以用sigmoid。

另外浅层神经网络也可以用tanh。

损失函数

categorical_crossentropy：多分类对数损失

binary_crossentropy：对数损失

mean_squared_error：均方差

mean_absolute_error：平均绝对值损失

对于多分类来说，主要用categorical_crossentropy。

优化器

SGD：随机梯度下降

Adagrad：Adaptive Gradient自适应梯度下降

Adadelta：对于Adagrad的进一步改进

RMSProp

Adam

本文将着重介绍后两种教程。

深度学习中的函数式编程

前面介绍的各种基本层，除了可以add进Sequential容器串联之外，它们本身也是callable对象，被调用之后，返回的还是callable对象。所以可以将它们视为函数，通过调用的方式来进行串联。

来个官方例子：

from keras.layers import Input, Densefrom keras.models import Modelinputs = Input(shape=(784,))x = Dense(64, activation='relu')(inputs)x = Dense(64, activation='relu')(x)predictions = Dense(10, activation='softmax')(x)model = Model(inputs=inputs, outputs=predictions)model.compile(optimizer='rmsprop', loss='categorical_crossentropy', metrics=['accuracy'])model.fit(data, labels)

为什么要用函数式编程？

答案是，复杂的网络结构并不是都是线性的add进容器中的。并行的，重用的，什么情况都有。这时候callable的优势就发挥出来了。

比如下面的Google Inception模型，就是带并联的：

我们的代码自然是以并联应对并联了，一个输入input_img被三个模型所重用：

from keras.layers import Conv2D, MaxPooling2D, Inputinput_img = Input(shape=(256, 256, 3))tower_1 = Conv2D(64, (1, 1), padding='same', activation='relu')(input_img)tower_1 = Conv2D(64, (3, 3), padding='same', activation='relu')(tower_1)tower_2 = Conv2D(64, (1, 1), padding='same', activation='relu')(input_img)tower_2 = Conv2D(64, (5, 5), padding='same', activation='relu')(tower_2)tower_3 = MaxPooling2D((3, 3), strides=(1, 1), padding='same')(input_img)tower_3 = Conv2D(64, (1, 1), padding='same', activation='relu')(tower_3)output = keras.layers.concatenate([tower_1, tower_2, tower_3], axis=1)

▌案例教程

CNN处理MNIST手写识别

光说不练是假把式。我们来看看符合五步法的处理MNIST的例子。

首先解析一下核心模型代码，因为模型是线性的，我们还是用Sequential容器

model = Sequential()

核心是两个卷积层：

model.add(Conv2D(32, kernel_size=(3, 3), activation='relu', input_shape=input_shape))model.add(Conv2D(64, (3, 3), activation='relu'))

为了防止过拟合，我们加上一个最大池化层，再加上一个Dropout层：

model.add(MaxPooling2D(pool_size=(2, 2)))model.add(Dropout(0.25))

下面要进入全连接层输出了，这两个中间的数据转换需要一个Flatten层：

model.add(Flatten())

下面是全连接层，激活函数是relu。

还怕过拟合，再来个Dropout层！

model.add(Dense(128, activation='relu'))model.add(Dropout(0.5))

最后通过一个softmax激活函数的全连接网络输出：

model.add(Dense(num_classes, activation='softmax'))

下面是编译这个模型，损失函数是categorical_crossentropy多类对数损失函数，优化器选用Adadelta。

model.compile(loss=keras.losses.categorical_crossentropy, optimizer=keras.optimizers.Adadelta(), metrics=['accuracy'])

下面是可以运行的完整代码：

from __future__ import print_functionimport kerasfrom keras.datasets import mnistfrom keras.models import Sequentialfrom keras.layers import Dense, Dropout, Flattenfrom keras.layers import Conv2D, MaxPooling2Dfrom keras import backend as Kbatch_size = 128num_classes = 10epochs = 12# input image dimensionsimg_rows, img_cols = 28, 28# the data, split between train and test sets(x_train, y_train), (x_test, y_test) = mnist.load_data()if K.image_data_format() == 'channels_first': x_train = x_train.reshape(x_train.shape[0], 1, img_rows, img_cols) x_test = x_test.reshape(x_test.shape[0], 1, img_rows, img_cols) input_shape = (1, img_rows, img_cols)else: x_train = x_train.reshape(x_train.shape[0], img_rows, img_cols, 1) x_test = x_test.reshape(x_test.shape[0], img_rows, img_cols, 1) input_shape = (img_rows, img_cols, 1) x_train = x_train.astype('float32')x_test = x_test.astype('float32')x_train /= 255x_test /= 255print('x_train shape:', x_train.shape)print(x_train.shape[0], 'train samples')print(x_test.shape[0], 'test samples')# convert class vectors to binary class matricesy_train = keras.utils.to_categorical(y_train, num_classes)y_test = keras.utils.to_categorical(y_test, num_classes)model = Sequential()model.add(Conv2D(32, kernel_size=(3, 3), activation='relu', input_shape=input_shape))model.add(Conv2D(64, (3, 3), activation='relu'))model.add(MaxPooling2D(pool_size=(2, 2)))model.add(Dropout(0.25))model.add(Flatten())model.add(Dense(128, activation='relu'))model.add(Dropout(0.5))model.add(Dense(num_classes, activation='softmax'))model.compile(loss=keras.losses.categorical_crossentropy, optimizer=keras.optimizers.Adadelta(), metrics=['accuracy'])model.fit(x_train, y_train, batch_size=batch_size, epochs=epochs, verbose=1, validation_data=(x_test, y_test))score = model.evaluate(x_test, y_test, verbose=0)print('Test loss:', score[0])print('Test accuracy:', score[1])

下面我们来个surprise例子，处理一下各种语言之间的翻译。

机器翻译：多语种互译

英译汉，汉译英之类的事情，在学生时代是不是一直难为这你呢？

现在不用担心了，只要有两种语言的对照表，我们就可以训练一个模型来像做一个机器翻译。

首先得下载一个字典：http://www.manythings.org/anki/

然后我们还是老办法，我们先看一下核心代码。没啥说的，这类序列化处理的问题用的一定是RNN，通常都是用LSTM.

encoder_inputs = Input(shape=(None, num_encoder_tokens))encoder = LSTM(latent_dim, return_state=True)encoder_outputs, state_h, state_c = encoder(encoder_inputs)encoder_states = [state_h, state_c]decoder_inputs = Input(shape=(None, num_decoder_tokens))decoder_lstm = LSTM(latent_dim, return_sequences=True, return_state=True)decoder_outputs, _, _ = decoder_lstm(decoder_inputs, initial_state=encoder_states)decoder_dense = Dense(num_decoder_tokens, activation='softmax')decoder_outputs = decoder_dense(decoder_outputs)model = Model([encoder_inputs, decoder_inputs], decoder_outputs)

优化器选用rmsprop，损失函数还是categorical_crossentropy.

validation_split是将一个集合随机分成训练集和测试集。

# Run trainingmodel.compile(optimizer='rmsprop', loss='categorical_crossentropy')model.fit([encoder_input_data, decoder_input_data], decoder_target_data, batch_size=batch_size, epochs=epochs, validation_split=0.2)

最后，训练一个模型不容易，我们将其存储起来。

model.save('s2s.h5')

最后，附上完整的实现了机器翻译功能的代码，加上注释和空行有100多行，供有需要的同学取用。

from __future__ import print_functionfrom keras.models import Modelfrom keras.layers import Input, LSTM, Denseimport numpy as npbatch_size = 64 # Batch size for training.epochs = 100 # Number of epochs to train for.latent_dim = 256 # Latent dimensionality of the encoding space.num_samples = 10000 # Number of samples to train on.# Path to the data txt file on disk.data_path = 'fra-eng/fra.txt'# Vectorize the data.input_texts = []target_texts = []input_characters = set()target_characters = set()with open(data_path, 'r', encoding='utf-8') as f: lines = f.read().split(' ')for line in lines[: min(num_samples, len(lines) - 1)]: input_text, target_text = line.split(' ')

# We use "tab" as the "start sequence" character # for the targets, and " " as "end sequence" character.

target_text = ' ' + target_text + ' ' input_texts.append(input_text) target_texts.append(target_text) for char in input_text:

if char not in input_characters:

input_characters.add(char) for char in target_text:

if char not in target_characters:

target_characters.add(char)input_characters = sorted(list(input_characters))target_characters = sorted(list(target_characters))num_encoder_tokens = len(input_characters)num_decoder_tokens = len(target_characters)max_encoder_seq_length = max([len(txt) for txt in input_texts])max_decoder_seq_length = max([len(txt) for txt in target_texts])print('Number of samples:', len(input_texts))print('Number of unique input tokens:', num_encoder_tokens)print('Number of unique output tokens:', num_decoder_tokens)print('Max sequence length for inputs:', max_encoder_seq_length)print('Max sequence length for outputs:', max_decoder_seq_length)input_token_index = dict( [(char, i) for i, char in enumerate(input_characters)])target_token_index = dict( [(char, i) for i, char in enumerate(target_characters)])encoder_input_data = np.zeros( (len(input_texts), max_encoder_seq_length, num_encoder_tokens), dtype='float32')decoder_input_data = np.zeros( (len(input_texts), max_decoder_seq_length, num_decoder_tokens), dtype='float32')decoder_target_data = np.zeros( (len(input_texts), max_decoder_seq_length, num_decoder_tokens),

dtype='float32')for i, (input_text, target_text) in enumerate(zip(input_texts, target_texts)): for t, char in enumerate(input_text): encoder_input_data[i, t, input_token_index[char]] = 1. for t, char in enumerate(target_text):

# decoder_target_data is ahead of decoder_input_data by one timestep decoder_input_data[i, t, target_token_index[char]] = 1.

if t > 0:

# decoder_target_data will be ahead by one timestep

# and will not include the start character.

decoder_target_data[i, t - 1, target_token_index[char]] = 1.# Define an input sequence and process it.encoder_inputs = Input(shape=(None, num_encoder_tokens))encoder = LSTM(latent_dim, return_state=True)encoder_outputs, state_h, state_c = encoder(encoder_inputs)# We discard `encoder_outputs` and only keep the states.encoder_states = [state_h, state_c]# Set up the decoder, using `encoder_states` as initial state.decoder_inputs = Input(shape=(None, num_decoder_tokens))# We set up our decoder to return full output sequences,# and to return internal states as well. We don't use the# return states in the training model, but we will use them in inference.decoder_lstm = LSTM(latent_dim, return_sequences=True, return_state=True)decoder_outputs, _, _ = decoder_lstm(decoder_inputs,

initial_state=encoder_states)decoder_dense = Dense(num_decoder_tokens, activation='softmax')decoder_outputs = decoder_dense(decoder_outputs)# Define the model that will turn# `encoder_input_data` & `decoder_input_data` into `decoder_target_data`model = Model([encoder_inputs, decoder_inputs], decoder_outputs)# Run trainingmodel.compile(optimizer='rmsprop', loss='categorical_crossentropy')model.fit([encoder_input_data, decoder_input_data], decoder_target_data, batch_size=batch_size, epochs=epochs,

validation_split=0.2)# Save modelmodel.save('s2s.h5')encoder_model = Model(encoder_inputs, encoder_states)decoder_state_input_h = Input(shape=(latent_dim,))decoder_state_input_c = Input(shape=(latent_dim,))decoder_states_inputs = [decoder_state_input_h, decoder_state_input_c]decoder_outputs, state_h, state_c = decoder_lstm( decoder_inputs, initial_state=decoder_states_inputs)decoder_states = [state_h, state_c]decoder_outputs = decoder_dense(decoder_outputs)decoder_model = Model( [decoder_inputs] + decoder_states_inputs, [decoder_outputs] + decoder_states)# Reverse-lookup token index to decode sequences back to# something readable.reverse_input_char_index = dict( (i, char) for char, i in input_token_index.items())reverse_target_char_index = dict( (i, char) for char, i in target_token_index.items())def decode_sequence(input_seq): # Encode the input as state vectors. states_value = encoder_model.predict(input_seq) # Generate empty target sequence of length 1.

target_seq = np.zeros((1, 1, num_decoder_tokens)) # Populate the first character of target sequence with the start character. target_seq[0, 0, target_token_index[' ']] = 1.

# Sampling loop for a batch of sequences # (to simplify, here we assume a batch of size 1). stop_condition = False decoded_sentence = '' while not stop_condition:

output_tokens, h, c = decoder_model.predict(

[target_seq] + states_value)

# Sample a token

sampled_token_index = np.argmax(output_tokens[0, -1, :])

sampled_char = reverse_target_char_index[sampled_token_index]

decoded_sentence += sampled_char

# Exit condition: either hit max length

# or find stop character.

if (sampled_char == ' ' or

len(decoded_sentence) > max_decoder_seq_length):

stop_condition = True

# Update the target sequence (of length 1).

target_seq = np.zeros((1, 1, num_decoder_tokens))

target_seq[0, 0, sampled_token_index] = 1.

# Update states

states_value = [h, c] return decoded_sentencefor seq_index in range(100):

# Take one sequence (part of the training set)

# for trying out decoding. input_seq = encoder_input_data[seq_index: seq_index + 1] decoded_sentence = decode_sequence(input_seq) print('-') print('Input sentence:', input_texts[seq_index]) print('Decoded sentence:', decoded_sentence)