深度学习100例

2023-09-11 11:19| 来源: 网络整理| 查看: 265

文章目录一、前期工作1. 设置GPU2. 导入数据3. 查看数据二、数据预处理1. 加载数据2. 可视化数据3. 再次检查数据4. 配置数据集三、残差网络（ResNet）介绍1. 残差网络解决了什么2. ResNet-50介绍四、构建ResNet-50网络模型五、编译六、训练模型七、模型评估八、保存and加载模型九、预测

一、前期工作

本文将采用ResNet-50实现鸟类图片的识别分类

🚀 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】

🚀 来自专栏：《深度学习100例》

如果你是一名深度学习小白可以先看看我这个专门为你写的专栏：《小白入门深度学习》

小白入门深度学习 | 第一篇：配置深度学习环境小白入门深度学习 | 第二篇：编译器的使用-Jupyter Notebook小白入门深度学习 | 第三篇：深度学习初体验小白入门深度学习 | 第四篇：配置PyTorch环境

🚀 往期精彩-卷积神经网络篇：

🚀 往期精彩-循环神经网络篇：

深度学习100例-循环神经网络（RNN）实现股票预测 | 第9天深度学习100例-循环神经网络（LSTM）实现股票预测 | 第10天

🚀 往期精彩-生成对抗网络篇：

深度学习100例-生成对抗网络（GAN）手写数字生成 | 第18天深度学习100例-生成对抗网络（DCGAN）手写数字生成 | 第19天深度学习100例-生成对抗网络（DCGAN）生成动漫小姐姐 | 第20天

转载请通过左侧联系方式（电脑端可看）、或者站内私信的方式联系我

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 # 设置随机种子尽可能使结果可以重现 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 import pathlib data_dir = "D:/jupyter notebook/DL-100-days/datasets/bird_photos" data_dir = pathlib.Path(data_dir) 3. 查看数据 image_count = len(list(data_dir.glob('*/*'))) print("图片总数为：",image_count) 图片总数为： 565 二、数据预处理文件夹数量Bananaquit166 张Black Throated Bushtiti111 张Black skimmer122 张Cockatoo166张 1. 加载数据

使用image_dataset_from_directory方法将磁盘中的数据加载到tf.data.Dataset中

batch_size = 8 img_height = 224 img_width = 224

TensorFlow版本是2.2.0的同学可能会遇到module 'tensorflow.keras.preprocessing' has no attribute 'image_dataset_from_directory'的报错，升级一下TensorFlow就OK了。

""" 关于image_dataset_from_directory()的详细介绍可以参考文章：https://mtyjkh.blog.csdn.net/article/details/117018789 """ train_ds = tf.keras.preprocessing.image_dataset_from_directory( data_dir, validation_split=0.2, subset="training", seed=123, image_size=(img_height, img_width), batch_size=batch_size) Found 565 files belonging to 4 classes. Using 452 files for training. """ 关于image_dataset_from_directory()的详细介绍可以参考文章：https://mtyjkh.blog.csdn.net/article/details/117018789 """ val_ds = tf.keras.preprocessing.image_dataset_from_directory( data_dir, validation_split=0.2, subset="validation", seed=123, image_size=(img_height, img_width), batch_size=batch_size) Found 565 files belonging to 4 classes. Using 113 files for validation.

我们可以通过class_names输出数据集的标签。标签将按字母顺序对应于目录名称。

class_names = train_ds.class_names print(class_names) ['Bananaquit', 'Black Throated Bushtiti', 'Black skimmer', 'Cockatoo'] 2. 可视化数据 plt.figure(figsize=(10, 5)) # 图形的宽为10高为5 plt.suptitle("微信公众号：K同学啊") for images, labels in train_ds.take(1): for i in range(8): ax = plt.subplot(2, 4, i + 1) plt.imshow(images[i].numpy().astype("uint8")) plt.title(class_names[labels[i]]) plt.axis("off")

plt.imshow(images[1].numpy().astype("uint8"))

在这里插入图片描述

3. 再次检查数据 for image_batch, labels_batch in train_ds: print(image_batch.shape) print(labels_batch.shape) break (8, 224, 224, 3) (8,) Image_batch是形状的张量（8, 224, 224, 3)。这是一批形状240x240x3的8张图片（最后一维指的是彩色通道RGB）。Label_batch是形状（8，）的张量，这些标签对应8张图片 4. 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/42417456prefetch() ：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 AUTOTUNE = tf.data.AUTOTUNE train_ds = train_ds.cache().shuffle(1000).prefetch(buffer_size=AUTOTUNE) val_ds = val_ds.cache().prefetch(buffer_size=AUTOTUNE)

如果报AttributeError: module 'tensorflow._api.v2.data' has no attribute 'AUTOTUNE'错误，将AUTOTUNE = tf.data.AUTOTUNE更换为AUTOTUNE = tf.data.experimental.AUTOTUNE

三、残差网络（ResNet）介绍 1. 残差网络解决了什么

残差网络是为了解决神经网络隐藏层过多时，而引起的网络退化问题。退化（degradation）问题是指：当网络隐藏层变多时，网络的准确度达到饱和然后急剧退化，而且这个退化不是由于过拟合引起的。

拓展：深度神经网络的“两朵乌云”

梯度弥散/爆炸

简单来讲就是网络太深了，会导致模型训练难以收敛。这个问题可以被标准初始化和中间层正规化的方法有效控制。（现阶段知道这么一回事就好了）

网络退化

随着网络深度增加，网络的表现先是逐渐增加至饱和，然后迅速下降，这个退化不是由于过拟合引起的。

2. ResNet-50介绍

ResNet-50有两个基本的块，分别名为Conv Block和Identity Block

Conv Block结构：

在这里插入图片描述

Identity Block结构：

在这里插入图片描述

ResNet-50总体结构：

在这里插入图片描述

四、构建ResNet-50网络模型

下面是本文的重点，可以试着按照上面三张图自己构建一下ResNet-50

from keras import layers from keras.layers import Input,Activation,BatchNormalization,Flatten from keras.layers import Dense,Conv2D,MaxPooling2D,ZeroPadding2D,AveragePooling2D from keras.models import Model def identity_block(input_tensor, kernel_size, filters, stage, block): filters1, filters2, filters3 = filters name_base = str(stage) + block + '_identity_block_' x = Conv2D(filters1, (1, 1), name=name_base + 'conv1')(input_tensor) x = BatchNormalization(name=name_base + 'bn1')(x) x = Activation('relu', name=name_base + 'relu1')(x) x = Conv2D(filters2, kernel_size,padding='same', name=name_base + 'conv2')(x) x = BatchNormalization(name=name_base + 'bn2')(x) x = Activation('relu', name=name_base + 'relu2')(x) x = Conv2D(filters3, (1, 1), name=name_base + 'conv3')(x) x = BatchNormalization(name=name_base + 'bn3')(x) x = layers.add([x, input_tensor] ,name=name_base + 'add') x = Activation('relu', name=name_base + 'relu4')(x) return x def conv_block(input_tensor, kernel_size, filters, stage, block, strides=(2, 2)): filters1, filters2, filters3 = filters res_name_base = str(stage) + block + '_conv_block_res_' name_base = str(stage) + block + '_conv_block_' x = Conv2D(filters1, (1, 1), strides=strides, name=name_base + 'conv1')(input_tensor) x = BatchNormalization(name=name_base + 'bn1')(x) x = Activation('relu', name=name_base + 'relu1')(x) x = Conv2D(filters2, kernel_size, padding='same', name=name_base + 'conv2')(x) x = BatchNormalization(name=name_base + 'bn2')(x) x = Activation('relu', name=name_base + 'relu2')(x) x = Conv2D(filters3, (1, 1), name=name_base + 'conv3')(x) x = BatchNormalization(name=name_base + 'bn3')(x) shortcut = Conv2D(filters3, (1, 1), strides=strides, name=res_name_base + 'conv')(input_tensor) shortcut = BatchNormalization(name=res_name_base + 'bn')(shortcut) x = layers.add([x, shortcut], name=name_base+'add') x = Activation('relu', name=name_base+'relu4')(x) return x def ResNet50(input_shape=[224,224,3],classes=1000): img_input = Input(shape=input_shape) x = ZeroPadding2D((3, 3))(img_input) x = Conv2D(64, (7, 7), strides=(2, 2), name='conv1')(x) x = BatchNormalization(name='bn_conv1')(x) x = Activation('relu')(x) x = MaxPooling2D((3, 3), strides=(2, 2))(x) x = conv_block(x, 3, [64, 64, 256], stage=2, block='a', strides=(1, 1)) x = identity_block(x, 3, [64, 64, 256], stage=2, block='b') x = identity_block(x, 3, [64, 64, 256], stage=2, block='c') x = conv_block(x, 3, [128, 128, 512], stage=3, block='a') x = identity_block(x, 3, [128, 128, 512], stage=3, block='b') x = identity_block(x, 3, [128, 128, 512], stage=3, block='c') x = identity_block(x, 3, [128, 128, 512], stage=3, block='d') x = conv_block(x, 3, [256, 256, 1024], stage=4, block='a') x = identity_block(x, 3, [256, 256, 1024], stage=4, block='b') x = identity_block(x, 3, [256, 256, 1024], stage=4, block='c') x = identity_block(x, 3, [256, 256, 1024], stage=4, block='d') x = identity_block(x, 3, [256, 256, 1024], stage=4, block='e') x = identity_block(x, 3, [256, 256, 1024], stage=4, block='f') x = conv_block(x, 3, [512, 512, 2048], stage=5, block='a') x = identity_block(x, 3, [512, 512, 2048], stage=5, block='b') x = identity_block(x, 3, [512, 512, 2048], stage=5, block='c') x = AveragePooling2D((7, 7), name='avg_pool')(x) x = Flatten()(x) x = Dense(classes, activation='softmax', name='fc1000')(x) model = Model(img_input, x, name='resnet50') # 加载预训练模型 model.load_weights("resnet50_weights_tf_dim_ordering_tf_kernels.h5") return model model = ResNet50() model.summary() Model: "resnet50" __________________________________________________________________________________________________ Layer (type) Output Shape Param # Connected to ================================================================================================== input_1 (InputLayer) [(None, 224, 224, 3) 0 __________________________________________________________________________________________________ zero_padding2d (ZeroPadding2D) (None, 230, 230, 3) 0 input_1[0][0] __________________________________________________________________________________________________ conv1 (Conv2D) (None, 112, 112, 64) 9472 zero_padding2d[0][0] __________________________________________________________________________________________________ bn_conv1 (BatchNormalization) (None, 112, 112, 64) 256 conv1[0][0] __________________________________________________________________________________________________ activation (Activation) (None, 112, 112, 64) 0 bn_conv1[0][0] __________________________________________________________________________________________________ max_pooling2d (MaxPooling2D) (None, 55, 55, 64) 0 activation[0][0] __________________________________________________________________________________________________ 2a_conv_block_conv1 (Conv2D) (None, 55, 55, 64) 4160 max_pooling2d[0][0] __________________________________________________________________________________________________ 2a_conv_block_bn1 (BatchNormali (None, 55, 55, 64) 256 2a_conv_block_conv1[0][0] __________________________________________________________________________________________________ 2a_conv_block_relu1 (Activation (None, 55, 55, 64) 0 2a_conv_block_bn1[0][0] __________________________________________________________________________________________________ 2a_conv_block_conv2 (Conv2D) (None, 55, 55, 64) 36928 2a_conv_block_relu1[0][0] __________________________________________________________________________________________________ 2a_conv_block_bn2 (BatchNormali (None, 55, 55, 64) 256 2a_conv_block_conv2[0][0] __________________________________________________________________________________________________ 2a_conv_block_relu2 (Activation (None, 55, 55, 64) 0 2a_conv_block_bn2[0][0] __________________________________________________________________________________________________ 2a_conv_block_conv3 (Conv2D) (None, 55, 55, 256) 16640 2a_conv_block_relu2[0][0] __________________________________________________________________________________________________ 2a_conv_block_res_conv (Conv2D) (None, 55, 55, 256) 16640 max_pooling2d[0][0] __________________________________________________________________________________________________ 2a_conv_block_bn3 (BatchNormali (None, 55, 55, 256) 1024 2a_conv_block_conv3[0][0] __________________________________________________________________________________________________ 2a_conv_block_res_bn (BatchNorm (None, 55, 55, 256) 1024 2a_conv_block_res_conv[0][0] __________________________________________________________________________________________________ 2a_conv_block_add (Add) (None, 55, 55, 256) 0 2a_conv_block_bn3[0][0] 2a_conv_block_res_bn[0][0] __________________________________________________________________________________________________ 2a_conv_block_relu4 (Activation (None, 55, 55, 256) 0 2a_conv_block_add[0][0] __________________________________________________________________________________________________ 2b_identity_block_conv1 (Conv2D (None, 55, 55, 64) 16448 2a_conv_block_relu4[0][0] __________________________________________________________________________________________________ 2b_identity_block_bn1 (BatchNor (None, 55, 55, 64) 256 2b_identity_block_conv1[0][0] ============================================================= 此处省略了若干行，此处省略了若干行，此处省略了若干行 ============================================================= __________________________________________________________________________________________________ 5c_identity_block_relu2 (Activa (None, 7, 7, 512) 0 5c_identity_block_bn2[0][0] __________________________________________________________________________________________________ 5c_identity_block_conv3 (Conv2D (None, 7, 7, 2048) 1050624 5c_identity_block_relu2[0][0] __________________________________________________________________________________________________ 5c_identity_block_bn3 (BatchNor (None, 7, 7, 2048) 8192 5c_identity_block_conv3[0][0] __________________________________________________________________________________________________ 5c_identity_block_add (Add) (None, 7, 7, 2048) 0 5c_identity_block_bn3[0][0] 5b_identity_block_relu4[0][0] __________________________________________________________________________________________________ 5c_identity_block_relu4 (Activa (None, 7, 7, 2048) 0 5c_identity_block_add[0][0] __________________________________________________________________________________________________ avg_pool (AveragePooling2D) (None, 1, 1, 2048) 0 5c_identity_block_relu4[0][0] __________________________________________________________________________________________________ flatten (Flatten) (None, 2048) 0 avg_pool[0][0] __________________________________________________________________________________________________ fc1000 (Dense) (None, 1000) 2049000 flatten[0][0] ================================================================================================== Total params: 25,636,712 Trainable params: 25,583,592 Non-trainable params: 53,120 __________________________________________________________________________________________________ 五、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。 # 设置优化器，我这里改变了学习率。 opt = tf.keras.optimizers.Adam(learning_rate=1e-7) model.compile(optimizer="adam", loss='sparse_categorical_crossentropy', metrics=['accuracy']) 六、训练模型 epochs = 10 history = model.fit( train_ds, validation_data=val_ds, epochs=epochs ) Epoch 1/10 57/57 [==============================] - 12s 86ms/step - loss: 2.4313 - accuracy: 0.6548 - val_loss: 213.7383 - val_accuracy: 0.3186 Epoch 2/10 57/57 [==============================] - 3s 52ms/step - loss: 0.4293 - accuracy: 0.8557 - val_loss: 9.0470 - val_accuracy: 0.2566 Epoch 3/10 57/57 [==============================] - 3s 52ms/step - loss: 0.2309 - accuracy: 0.9183 - val_loss: 1.4181 - val_accuracy: 0.7080 Epoch 4/10 57/57 [==============================] - 3s 53ms/step - loss: 0.1721 - accuracy: 0.9535 - val_loss: 2.5627 - val_accuracy: 0.6726 Epoch 5/10 57/57 [==============================] - 3s 53ms/step - loss: 0.0795 - accuracy: 0.9701 - val_loss: 0.2747 - val_accuracy: 0.8938 Epoch 6/10 57/57 [==============================] - 3s 52ms/step - loss: 0.0435 - accuracy: 0.9899 - val_loss: 0.1483 - val_accuracy: 0.9381 Epoch 7/10 57/57 [==============================] - 3s 52ms/step - loss: 0.0308 - accuracy: 0.9970 - val_loss: 0.1705 - val_accuracy: 0.9381 Epoch 8/10 57/57 [==============================] - 3s 52ms/step - loss: 0.0019 - accuracy: 1.0000 - val_loss: 0.0674 - val_accuracy: 0.9735 Epoch 9/10 57/57 [==============================] - 3s 52ms/step - loss: 8.2391e-04 - accuracy: 1.0000 - val_loss: 0.0720 - val_accuracy: 0.9735 Epoch 10/10 57/57 [==============================] - 3s 52ms/step - loss: 6.0079e-04 - accuracy: 1.0000 - val_loss: 0.0762 - val_accuracy: 0.9646 七、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.suptitle("微信公众号：K同学啊") 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()

八、保存and加载模型

这是最简单的模型保存与加载方法哈

# 保存模型 model.save('model/my_model.h5') # 加载模型 new_model = keras.models.load_model('model/my_model.h5') 九、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 plt.suptitle("微信公众号：K同学啊") for images, labels in val_ds.take(1): for i in range(8): ax = plt.subplot(2, 4, i + 1) # 显示图片 plt.imshow(images[i].numpy().astype("uint8")) # 需要给图片增加一个维度 img_array = tf.expand_dims(images[i], 0) # 使用模型预测图片中的人物 predictions = new_model.predict(img_array) plt.title(class_names[np.argmax(predictions)]) plt.axis("off")

在这里插入图片描述

其他精彩内容：

《深度学习100例》专栏直达：【传送门】

如果觉得本文对你有帮助记得点个关注，给个赞，加个收藏

【本文地址】

公司简介

联系我们