運(yùn)行系統(tǒng)：macOS Sequoia 15.0
Python編譯器：PyCharm 2024.1.4 (Community Edition)
Python版本：3.12
TensorFlow版本：2.17.0
Pytorch版本：2.4.1

往期鏈接：

1-5	6-10	11-20	21-30	31-40	41-50

51-60：函數(shù)	61-70：類	71-80：編程范式及設(shè)計(jì)模式

81-90：Python編碼規(guī)范	91-100：Python自帶常用模塊-1

101-105：Python自帶模塊-2	106-110：Python自帶模塊-3

111-115：Python常用第三方包-頻繁使用	116-120：Python常用第三方包-深度學(xué)習(xí)

121-125：Python常用第三方包-爬取數(shù)據(jù)	126-130：Python常用第三方包-為了樂趣

131-135：Python常用第三方包-拓展工具1	136-140：Python常用第三方包-拓展工具2

Python項(xiàng)目實(shí)戰(zhàn)

141-145	146-150	151-155	156-160	161-165	166-170	171-175	176-180

針對(duì)圖像的經(jīng)典卷積網(wǎng)絡(luò)結(jié)構(gòu)進(jìn)化史及可視化（續(xù)）

P181–MobileNet【2017】

模型結(jié)構(gòu)及創(chuàng)新性說明

MobileNet是一系列為移動(dòng)和嵌入式視覺應(yīng)用設(shè)計(jì)的輕量級(jí)卷積神經(jīng)網(wǎng)絡(luò)。以下是MobileNet各個(gè)版本的的主要特點(diǎn)：

（1）MobileNetV1版本

主要特點(diǎn)

• 引入深度可分離卷積（Depthwise Separable Convolution）
• 使用寬度乘子（Width Multiplier）和分辨率乘子（Resolution Multiplier）調(diào)整模型大小和復(fù)雜度

創(chuàng)新點(diǎn)

• 深度可分離卷積將標(biāo)準(zhǔn)卷積分解為深度卷積和逐點(diǎn)卷積，大大減少了計(jì)算量
• 使用ReLU6作為激活函數(shù)，有利于低精度計(jì)算

---

（2）MobileNetV2版本

主要特點(diǎn)

• 引入倒置殘差結(jié)構(gòu)（Inverted Residual Structure）
• 設(shè)計(jì)線性瓶頸（Linear Bottleneck）

創(chuàng)新點(diǎn)

• 倒置殘差結(jié)構(gòu)先擴(kuò)展通道數(shù)，再做深度卷積，最后壓縮回原來的通道數(shù)
• 去掉了最后一個(gè)ReLU，使用線性激活，有助于保留低維特征

---

（3）MobileNetV3

主要特點(diǎn)

• 網(wǎng)絡(luò)結(jié)構(gòu)搜索（NAS）優(yōu)化的網(wǎng)絡(luò)架構(gòu)
• 引入新的激活函數(shù)：h-swish
• 集成Squeeze-and-Excitation (SE) 模塊
• 提供Small和Large兩個(gè)版本

創(chuàng)新點(diǎn)

• 使用NAS自動(dòng)搜索最優(yōu)網(wǎng)絡(luò)結(jié)構(gòu)
• h-swish激活函數(shù)提高了精度，同時(shí)計(jì)算效率高
• SE模塊增強(qiáng)了特征的表達(dá)能力
• 優(yōu)化了網(wǎng)絡(luò)的首尾層，進(jìn)一步提高效率

---

模型結(jié)構(gòu)代碼

MobileNet V1版本

import tensorflow as tf
from tensorflow.keras import layers, modelsdef depthwise_conv_block(inputs, pointwise_conv_filters, alpha,depth_multiplier=1, strides=(1, 1), block_id=1):"""Adds a depthwise convolution block.A depthwise convolution block consists of a depthwise conv,batch normalization, ReLU6, pointwise convolution,batch normalization and ReLU6 activation."""channel_axis = -1pointwise_conv_filters = int(pointwise_conv_filters * alpha)x = layers.DepthwiseConv2D((3, 3),padding='same',depth_multiplier=depth_multiplier,strides=strides,use_bias=False,name='conv_dw_%d' % block_id)(inputs)x = layers.BatchNormalization(axis=channel_axis, name='conv_dw_%d_bn' % block_id)(x)x = layers.ReLU(6., name='conv_dw_%d_relu' % block_id)(x)x = layers.Conv2D(pointwise_conv_filters, (1, 1),padding='same',use_bias=False,strides=(1, 1),name='conv_pw_%d' % block_id)(x)x = layers.BatchNormalization(axis=channel_axis, name='conv_pw_%d_bn' % block_id)(x)return layers.ReLU(6., name='conv_pw_%d_relu' % block_id)(x)def MobileNetV1(input_shape=(224, 224, 3),alpha=1.0,depth_multiplier=1,dropout=1e-3,classes=1000):"""Instantiates the MobileNet architecture.Arguments:input_shape: Optional shape tuple, to be specified if you wouldlike to use a model with an input img resolution that is not(224, 224, 3).alpha: Controls the width of the network. This is known as thewidth multiplier in the MobileNet paper.- If `alpha` < 1.0, proportionally decreases the numberof filters in each layer.- If `alpha` > 1.0, proportionally increases the numberof filters in each layer.- If `alpha` = 1, default number of filters from the paperare used at each layer.depth_multiplier: Depth multiplier for depthwise convolution.This is called the resolution multiplier in the MobileNet paper.dropout: Dropout rate.classes: Optional number of classes to classify images into.Returns:A Keras model instance."""img_input = layers.Input(shape=input_shape)x = layers.Conv2D(int(32 * alpha), (3, 3),strides=(2, 2),padding='same',use_bias=False,name='conv1')(img_input)x = layers.BatchNormalization(axis=-1, name='conv1_bn')(x)x = layers.ReLU(6., name='conv1_relu')(x)x = depthwise_conv_block(x, 64, alpha, depth_multiplier, block_id=1)x = depthwise_conv_block(x, 128, alpha, depth_multiplier, strides=(2, 2), block_id=2)x = depthwise_conv_block(x, 128, alpha, depth_multiplier, block_id=3)x = depthwise_conv_block(x, 256, alpha, depth_multiplier, strides=(2, 2), block_id=4)x = depthwise_conv_block(x, 256, alpha, depth_multiplier, block_id=5)x = depthwise_conv_block(x, 512, alpha, depth_multiplier, strides=(2, 2), block_id=6)x = depthwise_conv_block(x, 512, alpha, depth_multiplier, block_id=7)x = depthwise_conv_block(x, 512, alpha, depth_multiplier, block_id=8)x = depthwise_conv_block(x, 512, alpha, depth_multiplier, block_id=9)x = depthwise_conv_block(x, 512, alpha, depth_multiplier, block_id=10)x = depthwise_conv_block(x, 512, alpha, depth_multiplier, block_id=11)x = depthwise_conv_block(x, 1024, alpha, depth_multiplier, strides=(2, 2), block_id=12)x = depthwise_conv_block(x, 1024, alpha, depth_multiplier, block_id=13)x = layers.GlobalAveragePooling2D()(x)x = layers.Reshape((1, 1, int(1024 * alpha)))(x)x = layers.Dropout(dropout, name='dropout')(x)x = layers.Conv2D(classes, (1, 1),padding='same',name='conv_preds')(x)x = layers.Reshape((classes,), name='reshape_2')(x)x = layers.Activation('softmax', name='act_softmax')(x)model = models.Model(img_input, x, name='mobilenet_v1')return model# 創(chuàng)建MobileNet V1模型
model = MobileNetV1(input_shape=(224, 224, 3), classes=1000)# 打印模型摘要
model.summary()

可以通過調(diào)整alpha參數(shù)來創(chuàng)建不同大小的MobileNetV1模型：

custom_model = MobileNetV1(input_shape=(224, 224, 3), classes=10, alpha=0.75)
custom_model.summary()

這將創(chuàng)建一個(gè)稍微窄一些（alpha=0.75）的MobileNet模型，用于10類分類任務(wù)。

MobileNet V2版本

import tensorflow as tf
from tensorflow.keras import layers, modelsdef inverted_residual_block(inputs, filters, stride, expand_ratio, alpha):input_channels = inputs.shape[-1]pointwise_filters = int(filters * alpha)# Expansion phasex = layers.Conv2D(int(input_channels * expand_ratio), kernel_size=1, padding='same', use_bias=False)(inputs)x = layers.BatchNormalization()(x)x = layers.ReLU(6.)(x)# Depthwise Convolutionx = layers.DepthwiseConv2D(kernel_size=3, strides=stride, padding='same', use_bias=False)(x)x = layers.BatchNormalization()(x)x = layers.ReLU(6.)(x)# Projectionx = layers.Conv2D(pointwise_filters, kernel_size=1, padding='same', use_bias=False)(x)x = layers.BatchNormalization()(x)# Residual connection if possibleif stride == 1 and input_channels == pointwise_filters:return layers.Add()([inputs, x])return xdef MobileNetV2(input_shape=(224, 224, 3), num_classes=1000, alpha=1.0, include_top=True):inputs = layers.Input(shape=input_shape)# First Convolution Layerx = layers.Conv2D(int(32 * alpha), kernel_size=3, strides=(2, 2), padding='same', use_bias=False)(inputs)x = layers.BatchNormalization()(x)x = layers.ReLU(6.)(x)# Inverted Residual Blocksx = inverted_residual_block(x, filters=16, stride=1, expand_ratio=1, alpha=alpha)x = inverted_residual_block(x, filters=24, stride=2, expand_ratio=6, alpha=alpha)x = inverted_residual_block(x, filters=24, stride=1, expand_ratio=6, alpha=alpha)x = inverted_residual_block(x, filters=32, stride=2, expand_ratio=6, alpha=alpha)x = inverted_residual_block(x, filters=32, stride=1, expand_ratio=6, alpha=alpha)x = inverted_residual_block(x, filters=32, stride=1, expand_ratio=6, alpha=alpha)x = inverted_residual_block(x, filters=64, stride=2, expand_ratio=6, alpha=alpha)x = inverted_residual_block(x, filters=64, stride=1, expand_ratio=6, alpha=alpha)x = inverted_residual_block(x, filters=64, stride=1, expand_ratio=6, alpha=alpha)x = inverted_residual_block(x, filters=64, stride=1, expand_ratio=6, alpha=alpha)x = inverted_residual_block(x, filters=96, stride=1, expand_ratio=6, alpha=alpha)x = inverted_residual_block(x, filters=96, stride=1, expand_ratio=6, alpha=alpha)x = inverted_residual_block(x, filters=96, stride=1, expand_ratio=6, alpha=alpha)x = inverted_residual_block(x, filters=160, stride=2, expand_ratio=6, alpha=alpha)x = inverted_residual_block(x, filters=160, stride=1, expand_ratio=6, alpha=alpha)x = inverted_residual_block(x, filters=160, stride=1, expand_ratio=6, alpha=alpha)x = inverted_residual_block(x, filters=320, stride=1, expand_ratio=6, alpha=alpha)# Last Convolution Layerx = layers.Conv2D(int(1280 * alpha), kernel_size=1, use_bias=False)(x)x = layers.BatchNormalization()(x)x = layers.ReLU(6.)(x)if include_top:x = layers.GlobalAveragePooling2D()(x)x = layers.Dense(num_classes, activation='softmax')(x)model = models.Model(inputs, x, name='MobileNetV2')return model# 創(chuàng)建MobileNet V2模型
model = MobileNetV2(input_shape=(224, 224, 3), num_classes=1000)# 打印模型摘要
model.summary()

MobileNet V3 版本

Small版本

import tensorflow as tf
from tensorflow.keras import layers, modelsclass HSwish(layers.Layer):def call(self, x):return x * tf.nn.relu6(x + 3) / 6class HSigmoid(layers.Layer):def call(self, x):return tf.nn.relu6(x + 3) / 6def squeeze_excite_block(inputs, se_ratio=0.25):x = layers.GlobalAveragePooling2D()(inputs)filters = inputs.shape[-1]x = layers.Dense(max(1, int(filters * se_ratio)), activation='relu')(x)x = layers.Dense(filters, activation=HSigmoid())(x)x = layers.Reshape((1, 1, filters))(x)return layers.multiply([inputs, x])def bneck(inputs, out_channels, exp_channels, kernel_size, stride, se_ratio, activation, alpha=1.0):x = layers.Conv2D(int(exp_channels * alpha), 1, padding='same', use_bias=False)(inputs)x = layers.BatchNormalization()(x)x = activation(x)x = layers.DepthwiseConv2D(kernel_size, stride, padding='same', use_bias=False)(x)x = layers.BatchNormalization()(x)x = activation(x)if se_ratio:x = squeeze_excite_block(x, se_ratio)x = layers.Conv2D(int(out_channels * alpha), 1, padding='same', use_bias=False)(x)x = layers.BatchNormalization()(x)if stride == 1 and inputs.shape[-1] == int(out_channels * alpha):return layers.Add()([inputs, x])return xdef MobileNetV3Small(input_shape=(224, 224, 3), num_classes=1000, alpha=1.0, include_top=True):inputs = layers.Input(shape=input_shape)x = layers.Conv2D(16, 3, strides=2, padding='same', use_bias=False)(inputs)x = layers.BatchNormalization()(x)x = HSwish()(x)x = bneck(x, 16, 16, 3, 2, 0.25, layers.ReLU(), alpha)x = bneck(x, 24, 72, 3, 2, None, layers.ReLU(), alpha)x = bneck(x, 24, 88, 3, 1, None, layers.ReLU(), alpha)x = bneck(x, 40, 96, 5, 2, 0.25, HSwish(), alpha)x = bneck(x, 40, 240, 5, 1, 0.25, HSwish(), alpha)x = bneck(x, 40, 240, 5, 1, 0.25, HSwish(), alpha)x = bneck(x, 48, 120, 5, 1, 0.25, HSwish(), alpha)x = bneck(x, 48, 144, 5, 1, 0.25, HSwish(), alpha)x = bneck(x, 96, 288, 5, 2, 0.25, HSwish(), alpha)x = bneck(x, 96, 576, 5, 1, 0.25, HSwish(), alpha)x = bneck(x, 96, 576, 5, 1, 0.25, HSwish(), alpha)x = layers.Conv2D(int(576 * alpha), 1, padding='same', use_bias=False)(x)x = layers.BatchNormalization()(x)x = HSwish()(x)x = layers.GlobalAveragePooling2D()(x)x = layers.Reshape((1, 1, int(576 * alpha)))(x)x = layers.Conv2D(int(1024 * alpha), 1, padding='same')(x)x = HSwish()(x)if include_top:x = layers.Conv2D(num_classes, 1, padding='same', activation='softmax')(x)x = layers.Reshape((num_classes,))(x)model = models.Model(inputs, x, name='MobileNetV3Small')return model# 創(chuàng)建MobileNet V3 Small模型
model = MobileNetV3Small(input_shape=(224, 224, 3), num_classes=1000)# 打印模型摘要
model.summary()

Large版本

import tensorflow as tf
from tensorflow.keras import layers, modelsclass HSwish(layers.Layer):def call(self, x):return x * tf.nn.relu6(x + 3) / 6class HSigmoid(layers.Layer):def call(self, x):return tf.nn.relu6(x + 3) / 6def squeeze_excite_block(inputs, se_ratio=0.25):x = layers.GlobalAveragePooling2D()(inputs)filters = inputs.shape[-1]x = layers.Dense(max(1, int(filters * se_ratio)), activation='relu')(x)x = layers.Dense(filters, activation=HSigmoid())(x)x = layers.Reshape((1, 1, filters))(x)return layers.multiply([inputs, x])def bneck(inputs, out_channels, exp_channels, kernel_size, stride, se_ratio, activation, alpha=1.0):x = layers.Conv2D(int(exp_channels * alpha), 1, padding='same', use_bias=False)(inputs)x = layers.BatchNormalization()(x)x = activation(x)x = layers.DepthwiseConv2D(kernel_size, stride, padding='same', use_bias=False)(x)x = layers.BatchNormalization()(x)x = activation(x)if se_ratio:x = squeeze_excite_block(x, se_ratio)x = layers.Conv2D(int(out_channels * alpha), 1, padding='same', use_bias=False)(x)x = layers.BatchNormalization()(x)if stride == 1 and inputs.shape[-1] == int(out_channels * alpha):return layers.Add()([inputs, x])return xdef MobileNetV3Large(input_shape=(224, 224, 3), num_classes=1000, alpha=1.0, include_top=True):inputs = layers.Input(shape=input_shape)x = layers.Conv2D(16, 3, strides=2, padding='same', use_bias=False)(inputs)x = layers.BatchNormalization()(x)x = HSwish()(x)x = bneck(x, 16, 16, 3, 1, None, layers.ReLU(), alpha)x = bneck(x, 24, 64, 3, 2, None, layers.ReLU(), alpha)x = bneck(x, 24, 72, 3, 1, None, layers.ReLU(), alpha)x = bneck(x, 40, 72, 5, 2, 0.25, layers.ReLU(), alpha)x = bneck(x, 40, 120, 5, 1, 0.25, layers.ReLU(), alpha)x = bneck(x, 40, 120, 5, 1, 0.25, layers.ReLU(), alpha)x = bneck(x, 80, 240, 3, 2, None, HSwish(), alpha)x = bneck(x, 80, 200, 3, 1, None, HSwish(), alpha)x = bneck(x, 80, 184, 3, 1, None, HSwish(), alpha)x = bneck(x, 80, 184, 3, 1, None, HSwish(), alpha)x = bneck(x, 112, 480, 3, 1, 0.25, HSwish(), alpha)x = bneck(x, 112, 672, 3, 1, 0.25, HSwish(), alpha)x = bneck(x, 160, 672, 5, 2, 0.25, HSwish(), alpha)x = bneck(x, 160, 960, 5, 1, 0.25, HSwish(), alpha)x = bneck(x, 160, 960, 5, 1, 0.25, HSwish(), alpha)x = layers.Conv2D(int(960 * alpha), 1, padding='same', use_bias=False)(x)x = layers.BatchNormalization()(x)x = HSwish()(x)x = layers.GlobalAveragePooling2D()(x)x = layers.Reshape((1, 1, int(960 * alpha)))(x)x = layers.Conv2D(int(1280 * alpha), 1, padding='same')(x)x = HSwish()(x)if include_top:x = layers.Conv2D(num_classes, 1, padding='same', activation='softmax')(x)x = layers.Reshape((num_classes,))(x)model = models.Model(inputs, x, name='MobileNetV3Large')return model# 創(chuàng)建MobileNet V3 Large模型
model = MobileNetV3Large(input_shape=(224, 224, 3), num_classes=1000)# 打印模型摘要
model.summary()

P182–EfficientNet【2019】

模型結(jié)構(gòu)及創(chuàng)新性說明

EfficientNet是由Google研究人員在2019年提出的一系列卷積神經(jīng)網(wǎng)絡(luò)模型，旨在提高模型效率和準(zhǔn)確性。以下是EfficientNet的主要特點(diǎn)：

模型結(jié)構(gòu)

• 基于MobileNetV2的倒置殘差結(jié)構(gòu)
• 使用Squeeze-and-Excitation (SE) 塊
• 采用復(fù)合縮放方法

創(chuàng)新性：

• 提出了復(fù)合縮放方法，同時(shí)縮放網(wǎng)絡(luò)的寬度、深度和分辨率
• 通過神經(jīng)架構(gòu)搜索(NAS)優(yōu)化基礎(chǔ)網(wǎng)絡(luò)結(jié)構(gòu)
• 在同等計(jì)算資源下，實(shí)現(xiàn)了更高的準(zhǔn)確率

模型結(jié)構(gòu)代碼

B0版本

import matplotlib.pyplot as plt
import tensorflow as tf
from keras.utils import plot_model
from tensorflow.keras import layers, models# macos系統(tǒng)顯示中文
plt.rcParams['font.sans-serif'] = ['Arial Unicode MS']def swish(x):return x * tf.nn.sigmoid(x)def se_block(inputs, se_ratio):channels = inputs.shape[-1]x = layers.GlobalAveragePooling2D()(inputs)x = layers.Dense(max(1, int(channels * se_ratio)), activation=swish)(x)x = layers.Dense(channels, activation='sigmoid')(x)return layers.Multiply()([inputs, x])def mbconv_block(inputs, out_channels, expand_ratio, stride, kernel_size, se_ratio):channels = inputs.shape[-1]x = inputs# Expansion phaseif expand_ratio != 1:expand_channels = channels * expand_ratiox = layers.Conv2D(expand_channels, 1, padding='same', use_bias=False)(x)x = layers.BatchNormalization()(x)x = layers.Activation(swish)(x)# Depthwise Convx = layers.DepthwiseConv2D(kernel_size, stride, padding='same', use_bias=False)(x)x = layers.BatchNormalization()(x)x = layers.Activation(swish)(x)# Squeeze and Excitationif se_ratio:x = se_block(x, se_ratio)# Output phasex = layers.Conv2D(out_channels, 1, padding='same', use_bias=False)(x)x = layers.BatchNormalization()(x)if stride == 1 and channels == out_channels:x = layers.Add()([inputs, x])return xdef efficientnet(width_coefficient, depth_coefficient, resolution, dropout_rate):base_architecture = [# expansion, channels, repeats, stride, kernel_size[1, 16, 1, 1, 3],[6, 24, 2, 2, 3],[6, 40, 2, 2, 5],[6, 80, 3, 2, 3],[6, 112, 3, 1, 5],[6, 192, 4, 2, 5],[6, 320, 1, 1, 3]]inputs = layers.Input(shape=(resolution, resolution, 3))x = layers.Conv2D(32, 3, strides=2, padding='same', use_bias=False)(inputs)x = layers.BatchNormalization()(x)x = layers.Activation(swish)(x)for i, (expansion, channels, repeats, stride, kernel_size) in enumerate(base_architecture):channels = int(channels * width_coefficient)repeats = int(repeats * depth_coefficient)for j in range(repeats):x = mbconv_block(x, channels, expansion, stride if j == 0 else 1, kernel_size, se_ratio=0.25)x = layers.Conv2D(1280, 1, padding='same', use_bias=False)(x)x = layers.BatchNormalization()(x)x = layers.Activation(swish)(x)x = layers.GlobalAveragePooling2D()(x)if dropout_rate > 0:x = layers.Dropout(dropout_rate)(x)outputs = layers.Dense(1000, activation='softmax')(x)model = tf.keras.Model(inputs, outputs)return model# EfficientNet-B0 configuration
def efficientnet_b0():return efficientnet(width_coefficient=1.0,depth_coefficient=1.0,resolution=224,dropout_rate=0.2)# Create the model
model_b0 = efficientnet_b0()# Print model summary
model_b0.summary()# 將模型結(jié)構(gòu)輸出到pdf
plot_model(model_b0, to_file='model_b0.pdf', show_shapes=True,show_layer_names=True)

B1–B7版本

def efficientnet_b1():return efficientnet(width_coefficient=1.0, depth_coefficient=1.1, resolution=240, dropout_rate=0.2)def efficientnet_b2():return efficientnet(width_coefficient=1.1, depth_coefficient=1.2, resolution=260, dropout_rate=0.3)def efficientnet_b3():return efficientnet(width_coefficient=1.2, depth_coefficient=1.4, resolution=300, dropout_rate=0.3)def efficientnet_b4():return efficientnet(width_coefficient=1.4, depth_coefficient=1.8, resolution=380, dropout_rate=0.4)def efficientnet_b5():return efficientnet(width_coefficient=1.6, depth_coefficient=2.2, resolution=456, dropout_rate=0.4)def efficientnet_b6():return efficientnet(width_coefficient=1.8, depth_coefficient=2.6, resolution=528, dropout_rate=0.5)def efficientnet_b7():return efficientnet(width_coefficient=2.0, depth_coefficient=3.1, resolution=600, dropout_rate=0.5)