深度学习100例-卷积神经网络（Inception-ResNet-v2）识别交通标志 | 第14天

一、前期工作

我的环境：

• 语言环境：Python3.6.5
• 编译器：jupyter notebook
• 深度学习环境：TensorFlow2.4.1
• 数据和代码：📌【传送门】

推荐阅读：

✨ 深度学习100例-卷积神经网络（CNN）实现mnist手写数字识别 | 第1天

✨ 深度学习100例 - 卷积神经网络（Inception V3）识别手语 | 第13天

✨ 手把手教你用 CNN 识别验证码 - 深度学习100例 | 第12天

✨ 循环神经网络（LSTM）实现股票预测-深度学习100例 | 第10天

✨ 深度学习100例-卷积神经网络（VGG-16）识别海贼王草帽一伙 | 第6天

🚀 来自专栏：《深度学习100例》

1. 设置GPU

如果使用的是CPU可以注释掉这部分的代码。

import tensorflow as tf

gpus = tf.config.list_physical_devices("GPU")

if gpus:
    tf.config.experimental.set_memory_growth(gpus[0], True)  #设置GPU显存用量按需使用
    tf.config.set_visible_devices([gpus[0]],"GPU")

2. 导入数据

import matplotlib.pyplot as plt
# 支持中文
plt.rcParams['font.sans-serif'] = ['SimHei']  # 用来正常显示中文标签
plt.rcParams['axes.unicode_minus'] = False  # 用来正常显示负号

import os,PIL,pathlib

# 设置随机种子尽可能使结果可以重现
import pandas as pd
import numpy  as np
np.random.seed(1)

# 设置随机种子尽可能使结果可以重现
import tensorflow as tf
tf.random.set_seed(1)

from tensorflow import keras
from tensorflow.keras import layers,models

# 导入图片数据
pictures_dir = "D:/jupyter notebook/DL-100-days/datasets/14_traffic_sign/images"
pictures_dir = pathlib.Path(pictures_dir)

# 导入训练数据的图片路径名及标签
train = pd.read_csv("D:/jupyter notebook/DL-100-days/datasets/14_traffic_sign/annotations.csv")

3. 查看数据

image_count = len(list(pictures_dir.glob('*.png')))

print("图片总数为：",image_count)

图片总数为： 5998

train.head()

	file_name	category
0	000_0001.png	0
1	000_0002.png	0
2	000_0003.png	0
3	000_0010.png	0
4	000_0011.png	0

二、构建一个tf.data.Dataset

1.加载数据

数据集中已经划分好了测试集与训练集，这次只需要进行分别加载就好了。

def preprocess_image(image):
    image = tf.image.decode_jpeg(image, channels=3)  # 编码解码处理
    image = tf.image.resize(image, [299,299])        # 图片调整
    return image/255.0                               # 归一化处理

def load_and_preprocess_image(path):
    image = tf.io.read_file(path)
    return preprocess_image(image)

AUTOTUNE = tf.data.experimental.AUTOTUNE
common_paths = "D:/jupyter notebook/DL-100-days/datasets/14_traffic_sign/images/"

# 训练数据的标签
train_image_label = [i for i in train["category"]]
train_label_ds = tf.data.Dataset.from_tensor_slices(train_image_label)

# 训练数据的路径
train_image_paths = [ common_paths+i for i in train["file_name"]]
# 加载图片路径
train_path_ds = tf.data.Dataset.from_tensor_slices(train_image_paths)
# 加载图片数据
train_image_ds = train_path_ds.map(load_and_preprocess_image, num_parallel_calls=AUTOTUNE)

# 将图片与标签进行对应打包
image_label_ds = tf.data.Dataset.zip((train_image_ds, train_label_ds))
image_label_ds

<ZipDataset shapes: ((299, 299, 3), ()), types: (tf.float32, tf.int32)>

plt.figure(figsize=(20,4))

for i in range(20):
    plt.subplot(2,10,i+1)
    plt.xticks([])
    plt.yticks([])
    plt.grid(False)
    
    # 显示图片
    images = plt.imread(train_image_paths[i])
    plt.imshow(images)
    # 显示标签
    plt.xlabel(train_image_label[i])

plt.show()

2. 配置数据集

• shuffle() ： 打乱数据，关于此函数的详细介绍可以参考：https://zhuanlan.zhihu.com/p/42417456
• prefetch() ：预取数据，加速运行，其详细介绍可以参考我前两篇文章，里面都有讲解。
• cache() ：将数据集缓存到内存当中，加速运行

我的电脑GPU配置是 NVIDIA GeForce RTX 3080 ,BATCH_SIZE = 8显存就会报OOM显存不足的错误，在BATCH_SIZE = 6时运行正常，每一个epoch时间约为130秒，大家请根据自己的电脑配置动态调整 BATCH_SIZE。采用CPU训练的同学，我建议将 BATCH_SIZE 调整为2或者1。

BATCH_SIZE = 6

# 将训练数据集拆分成训练集与验证集
train_ds = image_label_ds.take(5000).shuffle(1000)  # 前1500个batch
val_ds   = image_label_ds.skip(5000).shuffle(1000)  # 跳过前1500，选取后面的

train_ds = train_ds.batch(BATCH_SIZE)
train_ds = train_ds.prefetch(buffer_size=AUTOTUNE)

val_ds = val_ds.batch(BATCH_SIZE)
val_ds = val_ds.prefetch(buffer_size=AUTOTUNE)
val_ds

<PrefetchDataset shapes: ((None, 299, 299, 3), (None,)), types: (tf.float32, tf.int32)>

# 查看数据 shape 进行检查
for image_batch, labels_batch in train_ds:
    print(image_batch.shape)
    print(labels_batch.shape)
    break

(6, 299, 299, 3)
(6,)

# 再次查看数据，确认是否被打乱
plt.figure(figsize=(8,8))

for images, labels in train_ds.take(1):
    for i in range(6):
        
        ax = plt.subplot(4, 3, i + 1)  
        plt.imshow(images[i])
        plt.title(labels[i].numpy())  #使用.numpy()将张量转换为 NumPy 数组
        
        plt.axis("off")

三、Inception-ResNet-v2介绍

关于Inception系列的介绍可以见：https://baike.baidu.com/item/Inception%E7%BB%93%E6%9E%84 ，个人认为这些在现阶段只需要将模型走一遍（学会搭建），后期如果需要的话，可以再回头来进行详细研究。

这个模型相比之前写过的一些模型可能较为复杂一些，先放一张图整体感受一下它

关于上面卷积的计算还比较蒙的同学可以参考我这篇文章哈：卷积的计算

四、构建Inception-ResNet-v2网络

1.自己搭建

下面是本文的重点 InceptionResNetV2 网络模型的构建，可以试着按照上面的图自己构建一下 InceptionResNetV2，这部分我主要是参考官网的构建过程，将其单独拎了出来。

from tensorflow.keras import layers, models, Input
from tensorflow.keras.models import Model
from tensorflow.keras.layers import Conv2D, Dense, Flatten, Dropout,BatchNormalization,Activation
from tensorflow.keras.layers import MaxPooling2D, AveragePooling2D, Concatenate, Lambda,GlobalAveragePooling2D
from tensorflow.keras import backend as K

def conv2d_bn(x,filters,kernel_size,strides=1,padding='same',activation='relu',use_bias=False,name=None):
    
    x = Conv2D(filters,kernel_size,strides=strides,padding=padding,use_bias=use_bias,name=name)(x)
    
    if not use_bias:
        bn_axis = 1 if K.image_data_format() == 'channels_first' else 3
        bn_name = None if name is None else name + '_bn'
        x = BatchNormalization(axis=bn_axis, scale=False, name=bn_name)(x)
    if activation is not None:
        ac_name = None if name is None else name + '_ac'
        x = Activation(activation, name=ac_name)(x)
    return x

def inception_resnet_block(x, scale, block_type, block_idx, activation='relu'):
    if block_type == 'block35':
        branch_0 = conv2d_bn(x, 32, 1)
        branch_1 = conv2d_bn(x, 32, 1)
        branch_1 = conv2d_bn(branch_1, 32, 3)
        branch_2 = conv2d_bn(x, 32, 1)
        branch_2 = conv2d_bn(branch_2, 48, 3)
        branch_2 = conv2d_bn(branch_2, 64, 3)
        branches = [branch_0, branch_1, branch_2]
    elif block_type == 'block17':
        branch_0 = conv2d_bn(x, 192, 1)
        branch_1 = conv2d_bn(x, 128, 1)
        branch_1 = conv2d_bn(branch_1, 160, [1, 7])
        branch_1 = conv2d_bn(branch_1, 192, [7, 1])
        branches = [branch_0, branch_1]
    elif block_type == 'block8':
        branch_0 = conv2d_bn(x, 192, 1)
        branch_1 = conv2d_bn(x, 192, 1)
        branch_1 = conv2d_bn(branch_1, 224, [1, 3])
        branch_1 = conv2d_bn(branch_1, 256, [3, 1])
        branches = [branch_0, branch_1]
    else:
        raise ValueError('Unknown Inception-ResNet block type. '
                         'Expects "block35", "block17" or "block8", '
                         'but got: ' + str(block_type))

    block_name = block_type + '_' + str(block_idx)
    mixed = Concatenate(name=block_name + '_mixed')(branches)
    up = conv2d_bn(mixed,K.int_shape(x)[3],1,activation=None,use_bias=True,name=block_name + '_conv')

    x = Lambda(lambda inputs, scale: inputs[0] + inputs[1] * scale,
               output_shape=K.int_shape(x)[1:],
               arguments={'scale': scale},
               name=block_name)([x, up])
    if activation is not None:
        x = Activation(activation, name=block_name + '_ac')(x)
    return x


def InceptionResNetV2(input_shape=[299,299,3],classes=1000):

    inputs = Input(shape=input_shape)

    # Stem block
    x = conv2d_bn(inputs, 32, 3, strides=2, padding='valid')
    x = conv2d_bn(x, 32, 3, padding='valid')
    x = conv2d_bn(x, 64, 3)
    x = MaxPooling2D(3, strides=2)(x)
    x = conv2d_bn(x, 80, 1, padding='valid')
    x = conv2d_bn(x, 192, 3, padding='valid')
    x = MaxPooling2D(3, strides=2)(x)

    # Mixed 5b (Inception-A block)
    branch_0 = conv2d_bn(x, 96, 1)
    branch_1 = conv2d_bn(x, 48, 1)
    branch_1 = conv2d_bn(branch_1, 64, 5)
    branch_2 = conv2d_bn(x, 64, 1)
    branch_2 = conv2d_bn(branch_2, 96, 3)
    branch_2 = conv2d_bn(branch_2, 96, 3)
    branch_pool = AveragePooling2D(3, strides=1, padding='same')(x)
    branch_pool = conv2d_bn(branch_pool, 64, 1)
    branches = [branch_0, branch_1, branch_2, branch_pool]

    x = Concatenate(name='mixed_5b')(branches)

    # 10次 Inception-ResNet-A block
    for block_idx in range(1, 11):
        x = inception_resnet_block(x, scale=0.17, block_type='block35', block_idx=block_idx)

    # Reduction-A block
    branch_0 = conv2d_bn(x, 384, 3, strides=2, padding='valid')
    branch_1 = conv2d_bn(x, 256, 1)
    branch_1 = conv2d_bn(branch_1, 256, 3)
    branch_1 = conv2d_bn(branch_1, 384, 3, strides=2, padding='valid')
    branch_pool = MaxPooling2D(3, strides=2, padding='valid')(x)
    branches = [branch_0, branch_1, branch_pool]
    x = Concatenate(name='mixed_6a')(branches)

    # 20次 Inception-ResNet-B block
    for block_idx in range(1, 21):
        x = inception_resnet_block(x, scale=0.1, block_type='block17', block_idx=block_idx)


    # Reduction-B block
    branch_0 = conv2d_bn(x, 256, 1)
    branch_0 = conv2d_bn(branch_0, 384, 3, strides=2, padding='valid')
    branch_1 = conv2d_bn(x, 256, 1)
    branch_1 = conv2d_bn(branch_1, 288, 3, strides=2, padding='valid')
    branch_2 = conv2d_bn(x, 256, 1)
    branch_2 = conv2d_bn(branch_2, 288, 3)
    branch_2 = conv2d_bn(branch_2, 320, 3, strides=2, padding='valid')
    branch_pool = MaxPooling2D(3, strides=2, padding='valid')(x)
    branches = [branch_0, branch_1, branch_2, branch_pool]
    x = Concatenate(name='mixed_7a')(branches)

    # 10次 Inception-ResNet-C block
    for block_idx in range(1, 10):
        x = inception_resnet_block(x, scale=0.2, block_type='block8', block_idx=block_idx)
    x = inception_resnet_block(x, scale=1., activation=None, block_type='block8', block_idx=10)


    x = conv2d_bn(x, 1536, 1, name='conv_7b')

    x = GlobalAveragePooling2D(name='avg_pool')(x)
    x = Dense(classes, activation='softmax', name='predictions')(x)

    # 创建模型
    model = Model(inputs, x, name='inception_resnet_v2')

    return model

model = InceptionResNetV2([299,299,3],58)
model.summary()

Model: "inception_resnet_v2"
__________________________________________________________________________________________________
Layer (type)                    Output Shape         Param #     Connected to                     
==================================================================================================
input_1 (InputLayer)            [(None, 299, 299, 3) 0                                            
__________________________________________________________________________________________________
conv2d (Conv2D)                 (None, 149, 149, 32) 864         input_1[0][0]                    
__________________________________________________________________________________________________
batch_normalization (BatchNorma (None, 149, 149, 32) 96          conv2d[0][0]                     
__________________________________________________________________________________________________
....................
__________________________________________________________________________________________________
conv_7b (Conv2D)                (None, 8, 8, 1536)   3194880     block8_10[0][0]                  
__________________________________________________________________________________________________
conv_7b_bn (BatchNormalization) (None, 8, 8, 1536)   4608        conv_7b[0][0]                    
__________________________________________________________________________________________________
conv_7b_ac (Activation)         (None, 8, 8, 1536)   0           conv_7b_bn[0][0]                 
__________________________________________________________________________________________________
avg_pool (GlobalAveragePooling2 (None, 1536)         0           conv_7b_ac[0][0]                 
__________________________________________________________________________________________________
predictions (Dense)             (None, 58)           89146       avg_pool[0][0]                   
==================================================================================================
Total params: 54,425,882
Trainable params: 54,365,338
Non-trainable params: 60,544
__________________________________________________________________________________________________

2.官方模型

# import tensorflow as tf
# # 如果使用官方模型需要将图片shape调整为 [299,299,3]，目前图片的shape是 [150,150,3]
# model = tf.keras.applications.inception_resnet_v2.InceptionResNetV2()
# model.summary()

五、设置动态学习率

这里先罗列一下学习率大与学习率小的优缺点。

• 学习率大

  • 优点：
    1、加快学习速率。
    2、有助于跳出局部最优值。
  • 缺点：
    1、导致模型训练不收敛。
    2、单单使用大学习率容易导致模型不精确。

• 学习率小

  • 优点：
    1、有助于模型收敛、模型细化。
    2、提高模型精度。
  • 缺点：
    1、很难跳出局部最优值。
    2、收敛缓慢。

注意：这里设置的动态学习率为：指数衰减型（ExponentialDecay）。在每一个epoch开始前，学习率（learning_rate）都将会重置为初始学习率（initial_learning_rate），然后再重新开始衰减。计算公式如下：

learning_rate = initial_learning_rate * decay_rate ^ (step / decay_steps)

# 设置初始学习率
initial_learning_rate = 1e-4

lr_schedule = tf.keras.optimizers.schedules.ExponentialDecay(
        initial_learning_rate, 
        decay_steps=100,      # 敲黑板！！！这里是指 steps，不是指epochs
        decay_rate=0.96,  # lr经过一次衰减就会变成 decay_rate*lr
        staircase=True)

# 将指数衰减学习率送入优化器
optimizer = tf.keras.optimizers.Adam(learning_rate=lr_schedule)

六、编译

在准备对模型进行训练之前，还需要再对其进行一些设置。以下内容是在模型的编译步骤中添加的：

• 损失函数（loss）：用于衡量模型在训练期间的准确率。
• 优化器（optimizer）：决定模型如何根据其看到的数据和自身的损失函数进行更新。
• 指标（metrics）：用于监控训练和测试步骤。以下示例使用了准确率，即被正确分类的图像的比率。

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

七、训练模型

Inception-ResNet-v2 模型相对之前的模型较为复杂，故而运行耗时也更长，我这边每一个epoch运行时间是130s左右。我的GPU配置是 NVIDIA GeForce RTX 3080。建议大家先将 epochs 调整为1跑通程序。

epochs = 10

history = model.fit(
    train_ds,
    validation_data=val_ds,
    epochs=epochs
)

Epoch 1/10
834/834 [==============================] - 154s 163ms/step - loss: 2.5214 - accuracy: 0.3563 - val_loss: 1.3834 - val_accuracy: 0.6168
Epoch 2/10
834/834 [==============================] - 133s 159ms/step - loss: 0.9230 - accuracy: 0.7522 - val_loss: 0.5457 - val_accuracy: 0.8531
Epoch 3/10
834/834 [==============================] - 133s 159ms/step - loss: 0.3952 - accuracy: 0.9105 - val_loss: 0.3391 - val_accuracy: 0.9064
Epoch 4/10
834/834 [==============================] - 134s 160ms/step - loss: 0.1876 - accuracy: 0.9655 - val_loss: 0.2481 - val_accuracy: 0.9296
Epoch 5/10
834/834 [==============================] - 131s 156ms/step - loss: 0.1071 - accuracy: 0.9862 - val_loss: 0.1265 - val_accuracy: 0.9716
Epoch 6/10
834/834 [==============================] - 128s 153ms/step - loss: 0.0587 - accuracy: 0.9954 - val_loss: 0.0911 - val_accuracy: 0.9794
Epoch 7/10
834/834 [==============================] - 132s 158ms/step - loss: 0.0429 - accuracy: 0.9976 - val_loss: 0.0941 - val_accuracy: 0.9777
Epoch 8/10
834/834 [==============================] - 132s 158ms/step - loss: 0.0306 - accuracy: 0.9980 - val_loss: 0.0955 - val_accuracy: 0.9777
Epoch 9/10
834/834 [==============================] - 133s 158ms/step - loss: 0.0248 - accuracy: 0.9997 - val_loss: 0.0864 - val_accuracy: 0.9794
Epoch 10/10
834/834 [==============================] - 132s 158ms/step - loss: 0.0216 - accuracy: 0.9988 - val_loss: 0.0750 - val_accuracy: 0.9794

八、模型评估

acc = history.history['accuracy']
val_acc = history.history['val_accuracy']

loss = history.history['loss']
val_loss = history.history['val_loss']

epochs_range = range(epochs)

plt.figure(figsize=(12, 4))
plt.subplot(1, 2, 1)

plt.plot(epochs_range, acc, label='Training Accuracy')
plt.plot(epochs_range, val_acc, label='Validation Accuracy')
plt.legend(loc='lower right')
plt.title('Training and Validation Accuracy')

plt.subplot(1, 2, 2)
plt.plot(epochs_range, loss, label='Training Loss')
plt.plot(epochs_range, val_loss, label='Validation Loss')
plt.legend(loc='upper right')
plt.title('Training and Validation Loss')
plt.show()

九、模型的保存与加载

# 保存模型
model.save('model/14_model.h5')

# 加载模型
new_model = keras.models.load_model('model/14_model.h5')

十、预测

# 采用加载的模型（new_model）来看预测结果

plt.figure(figsize=(10, 5))  # 图形的宽为10高为5

for images, labels in val_ds.take(1):
    for i in range(6):
        ax = plt.subplot(2, 3, i + 1)  
        
        # 显示图片
        plt.imshow(images[i])
        
        # 需要给图片增加一个维度
        img_array = tf.expand_dims(images[i], 0) 
        
        # 使用模型预测路标
        predictions = new_model.predict(img_array)
        plt.title(np.argmax(predictions))

        plt.axis("off")

---

推荐阅读：

✨ 深度学习100例-卷积神经网络（CNN）实现mnist手写数字识别 | 第1天

✨ 深度学习100例 - 卷积神经网络（Inception V3）识别手语 | 第13天

✨ 手把手教你用 CNN 识别验证码 - 深度学习100例 | 第12天

✨ 循环神经网络（LSTM）实现股票预测-深度学习100例 | 第10天

✨ 深度学习100例-卷积神经网络（VGG-16）识别海贼王草帽一伙 | 第6天

🚀 选自专栏：《深度学习100例》

未完～

• ✨微信众号（K同学啊）后台回复【DL+天数】可获取《深度学习100例》的数据
• ✨微信交流群：加我微信（mtyjkh_）拉你进群，不懂就问，备注：CSDN+来意

我的微信：

---

最后再送大家一本，帮助大家拿到 BAT 等一线大厂 offer 的数据结构刷题笔记，是谷歌和阿里的大佬写的，对于算法薄弱或者需要提高的同学都十分受用（提取码：9go2 ）：

谷歌和阿里大佬的Leetcode刷题笔记

以及我整理的7K+本开源电子书，总有一本可以帮到你 💖（提取码：4eg0）

7K+本开源电子书