第T8周：使用TensorFlow实现猫狗识别

• 🍨 本文为🔗365天深度学习训练营 中的学习记录博客
• 🍖 原作者：K同学啊
  

电脑环境：
语言环境：Python 3.8.0
编译器：Jupyter Notebook
深度学习环境：tensorflow 2.15.0

一、前期工作

1.设置GPU（如果使用的是CPU可以忽略这步）

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")

# 打印显卡信息，确认GPU可用
print(gpus)

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

#隐藏警告
import warnings
warnings.filterwarnings('ignore')

data_dir = "./365-7-data"
data_dir = pathlib.Path(data_dir)

image_count = len(list(data_dir.glob('*/*')))

print("图片总数为：",image_count)

二、数据预处理

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

"""
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=12,
    image_size=(img_height, img_width),
    batch_size=batch_size)

val_ds = tf.keras.preprocessing.image_dataset_from_directory(
    data_dir,
    validation_split=0.2,
    subset="validation",
    seed=12,
    image_size=(img_height, img_width),
    batch_size=batch_size)

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)

输出：

[‘cat’, ‘dog’]

2、再次检查数据

for image_batch, labels_batch in train_ds:
    print(image_batch.shape)
    print(labels_batch.shape)
    break

输出：

(8, 224, 224, 3)
(8,)

3. 配置数据集

AUTOTUNE = tf.data.AUTOTUNE

def preprocess_image(image,label):
    return (image/255.0,label)

# 归一化处理
train_ds = train_ds.map(preprocess_image, num_parallel_calls=AUTOTUNE)
val_ds   = val_ds.map(preprocess_image, num_parallel_calls=AUTOTUNE)

train_ds = train_ds.cache().shuffle(1000).prefetch(buffer_size=AUTOTUNE)
val_ds   = val_ds.cache().prefetch(buffer_size=AUTOTUNE)

4. 可视化数据

plt.figure(figsize=(15, 10))  # 图形的宽为15高为10

for images, labels in train_ds.take(1):
    for i in range(8):
        
        ax = plt.subplot(5, 8, i + 1) 
        plt.imshow(images[i])
        plt.title(class_names[labels[i]])
        
        plt.axis("off")

三、构建CNN网络

from tensorflow.keras import layers, models, Input
from tensorflow.keras.models import Model
from tensorflow.keras.layers import Conv2D, MaxPooling2D, Dense, Flatten, Dropout

def VGG16(nb_classes, input_shape):
    input_tensor = Input(shape=input_shape)
    # 1st block
    x = Conv2D(64, (3,3), activation='relu', padding='same',name='block1_conv1')(input_tensor)
    x = Conv2D(64, (3,3), activation='relu', padding='same',name='block1_conv2')(x)
    x = MaxPooling2D((2,2), strides=(2,2), name = 'block1_pool')(x)
    # 2nd block
    x = Conv2D(128, (3,3), activation='relu', padding='same',name='block2_conv1')(x)
    x = Conv2D(128, (3,3), activation='relu', padding='same',name='block2_conv2')(x)
    x = MaxPooling2D((2,2), strides=(2,2), name = 'block2_pool')(x)
    # 3rd block
    x = Conv2D(256, (3,3), activation='relu', padding='same',name='block3_conv1')(x)
    x = Conv2D(256, (3,3), activation='relu', padding='same',name='block3_conv2')(x)
    x = Conv2D(256, (3,3), activation='relu', padding='same',name='block3_conv3')(x)
    x = MaxPooling2D((2,2), strides=(2,2), name = 'block3_pool')(x)
    # 4th block
    x = Conv2D(512, (3,3), activation='relu', padding='same',name='block4_conv1')(x)
    x = Conv2D(512, (3,3), activation='relu', padding='same',name='block4_conv2')(x)
    x = Conv2D(512, (3,3), activation='relu', padding='same',name='block4_conv3')(x)
    x = MaxPooling2D((2,2), strides=(2,2), name = 'block4_pool')(x)
    # 5th block
    x = Conv2D(512, (3,3), activation='relu', padding='same',name='block5_conv1')(x)
    x = Conv2D(512, (3,3), activation='relu', padding='same',name='block5_conv2')(x)
    x = Conv2D(512, (3,3), activation='relu', padding='same',name='block5_conv3')(x)
    x = MaxPooling2D((2,2), strides=(2,2), name = 'block5_pool')(x)
    # full connection
    x = Flatten()(x)
    x = Dense(4096, activation='relu',  name='fc1')(x)
    x = Dense(4096, activation='relu', name='fc2')(x)
    output_tensor = Dense(nb_classes, activation='softmax', name='predictions')(x)

    model = Model(input_tensor, output_tensor)
    return model

model=VGG16(1000, (img_width, img_height, 3))
model.summary()

四、编译

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

五、训练模型

from tqdm import tqdm
import tensorflow.keras.backend as K

epochs = 10
lr     = 1e-4

# 记录训练数据，方便后面的分析
history_train_loss     = []
history_train_accuracy = []
history_val_loss       = []
history_val_accuracy   = []

for epoch in range(epochs):
    train_total = len(train_ds)
    val_total   = len(val_ds)
    
    """
    total：预期的迭代数目
    ncols：控制进度条宽度
    mininterval：进度更新最小间隔，以秒为单位（默认值：0.1）
    """
    with tqdm(total=train_total, desc=f'Epoch {epoch + 1}/{epochs}',mininterval=1,ncols=100) as pbar:
        
        lr = lr*0.92
        K.set_value(model.optimizer.lr, lr)

        for image,label in train_ds:   
            """
            训练模型，简单理解train_on_batch就是：它是比model.fit()更高级的一个用法

            想详细了解 train_on_batch 的同学，
            可以看看我的这篇文章：https://www.yuque.com/mingtian-fkmxf/hv4lcq/ztt4gy
            """
            history = model.train_on_batch(image,label)

            train_loss     = history[0]
            train_accuracy = history[1]
            
            pbar.set_postfix({"loss": "%.4f"%train_loss,
                              "accuracy":"%.4f"%train_accuracy,
                              "lr": K.get_value(model.optimizer.lr)})
            pbar.update(1)
        history_train_loss.append(train_loss)
        history_train_accuracy.append(train_accuracy)
            
    print('开始验证！')
    
    with tqdm(total=val_total, desc=f'Epoch {epoch + 1}/{epochs}',mininterval=0.3,ncols=100) as pbar:

        for image,label in val_ds:      
            
            history = model.test_on_batch(image,label)
            
            val_loss     = history[0]
            val_accuracy = history[1]
            
            pbar.set_postfix({"loss": "%.4f"%val_loss,
                              "accuracy":"%.4f"%val_accuracy})
            pbar.update(1)
        history_val_loss.append(val_loss)
        history_val_accuracy.append(val_accuracy)
            
    print('结束验证！')
    print("验证loss为：%.4f"%val_loss)
    print("验证准确率为：%.4f"%val_accuracy)

输出：

Epoch 1/10: 100%|████████| 340/340 [01:53<00:00,  2.99it/s, loss=0.8901, accuracy=0.1250, lr=9.2e-5]
开始验证！
Epoch 1/10: 100%|█████████████████████| 85/85 [00:03<00:00, 23.67it/s, loss=0.6123, accuracy=0.6250]
结束验证！
验证loss为：0.6123
验证准确率为：0.6250
Epoch 2/10: 100%|███████| 340/340 [00:22<00:00, 15.12it/s, loss=0.1449, accuracy=1.0000, lr=8.46e-5]
开始验证！
Epoch 2/10: 100%|█████████████████████| 85/85 [00:03<00:00, 25.99it/s, loss=0.2008, accuracy=0.8750]
结束验证！
验证loss为：0.2008
验证准确率为：0.8750
Epoch 3/10: 100%|███████| 340/340 [00:22<00:00, 15.23it/s, loss=0.0083, accuracy=1.0000, lr=7.79e-5]
开始验证！
Epoch 3/10: 100%|█████████████████████| 85/85 [00:03<00:00, 25.47it/s, loss=0.0298, accuracy=1.0000]
结束验证！
验证loss为：0.0298
验证准确率为：1.0000
Epoch 4/10: 100%|███████| 340/340 [00:22<00:00, 14.86it/s, loss=0.0321, accuracy=1.0000, lr=7.16e-5]
开始验证！
Epoch 4/10: 100%|█████████████████████| 85/85 [00:03<00:00, 25.84it/s, loss=0.0092, accuracy=1.0000]
结束验证！
验证loss为：0.0092
验证准确率为：1.0000
Epoch 5/10: 100%|███████| 340/340 [00:22<00:00, 15.03it/s, loss=0.3167, accuracy=0.8750, lr=6.59e-5]
开始验证！
Epoch 5/10: 100%|█████████████████████| 85/85 [00:03<00:00, 26.73it/s, loss=0.0381, accuracy=1.0000]
结束验证！
验证loss为：0.0381
验证准确率为：1.0000
Epoch 6/10: 100%|███████| 340/340 [00:22<00:00, 15.38it/s, loss=0.0323, accuracy=1.0000, lr=6.06e-5]
开始验证！
Epoch 6/10: 100%|█████████████████████| 85/85 [00:03<00:00, 25.85it/s, loss=0.0002, accuracy=1.0000]
结束验证！
验证loss为：0.0002
验证准确率为：1.0000
Epoch 7/10: 100%|███████| 340/340 [00:22<00:00, 15.04it/s, loss=0.0005, accuracy=1.0000, lr=5.58e-5]
开始验证！
Epoch 7/10: 100%|█████████████████████| 85/85 [00:03<00:00, 26.34it/s, loss=0.0040, accuracy=1.0000]
结束验证！
验证loss为：0.0040
验证准确率为：1.0000
Epoch 8/10: 100%|███████| 340/340 [00:21<00:00, 15.47it/s, loss=0.0018, accuracy=1.0000, lr=5.13e-5]
开始验证！
Epoch 8/10: 100%|█████████████████████| 85/85 [00:03<00:00, 26.12it/s, loss=0.0171, accuracy=1.0000]
结束验证！
验证loss为：0.0171
验证准确率为：1.0000
Epoch 9/10: 100%|███████| 340/340 [00:22<00:00, 15.38it/s, loss=0.0000, accuracy=1.0000, lr=4.72e-5]
开始验证！
Epoch 9/10: 100%|█████████████████████| 85/85 [00:03<00:00, 26.08it/s, loss=0.0009, accuracy=1.0000]
结束验证！
验证loss为：0.0009
验证准确率为：1.0000
Epoch 10/10: 100%|██████| 340/340 [00:21<00:00, 15.49it/s, loss=0.0050, accuracy=1.0000, lr=4.34e-5]
开始验证！
Epoch 10/10: 100%|████████████████████| 85/85 [00:03<00:00, 26.46it/s, loss=0.0001, accuracy=1.0000]
结束验证！
验证loss为：0.0001
验证准确率为：1.0000

六、模型评估

epochs_range = range(epochs)

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

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

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

七、预测

import numpy as np

# 采用加载的模型（new_model）来看预测结果
plt.figure(figsize=(18, 3))  # 图形的宽为18高为5
plt.suptitle("预测结果展示")

for images, labels in val_ds.take(1):
    for i in range(8):
        ax = plt.subplot(1,8, i + 1)  
        
        # 显示图片
        plt.imshow(images[i].numpy())
        
        # 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 = model.predict(img_array)
        plt.title(class_names[np.argmax(predictions)])

        plt.axis("off")

输出：

1/1 [==============================] - 0s 247ms/step
1/1 [==============================] - 0s 19ms/step
1/1 [==============================] - 0s 21ms/step
1/1 [==============================] - 0s 23ms/step
1/1 [==============================] - 0s 20ms/step
1/1 [==============================] - 0s 21ms/step
1/1 [==============================] - 0s 22ms/step
1/1 [==============================] - 0s 19ms/step

八、知识点

1、训练方式

这是我们之前的训练方法。

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

本次使用的训练函数是model.train_on_batch。
函数原型：

Model.train_on_batch(x, y=None, sample_weight=None, class_weight=None, return_dict=False)

• sample_weight：与x长度相同的可选数组，包含适用于每个样本的模型损失的权重。在时态数据的情况下，您可以传递一个具有形状(samples, sequence_length)的2D数组，以便对每个样本的每个时间步应用不同的权重。
• class_weight：可选的字典。将类索引(整数)映射到权值(浮点数)，以应用于训练期间该类样本的模型损失。这对于告诉模型“更多地关注”来自代表性不足的类的样本是有用的。
• return_dict：如果为True，则损失和度量结果将作为字典返回，其中每个键是度量的名称。如果为False，它们将作为列表返回。

2、tqdm

tqdm是一个用于在终端中显示进度条的Python库。它提供了一种简单的方式来跟踪迭代过程的进度，无论是在循环中处理大量数据还是在长时间运行的任务中。

2.1、基本用法：

• 在for循环中使用：

from tqdm import tqdm
import time

for i in tqdm(range(10)):
    time.sleep(1)# 模拟任务执行时间

100%|██████████| 10/10 [00:10<00:00,  1.00s/it]

• 自定义进度条样式

desc：设置进度条的前缀文本；ncols：设置进度条的长度

from tqdm import tqdm
import time
for i in tqdm(range(10), desc="Processing", ncols=80):
    time.sleep(0.5)   

Processing: 100%|███████████████████████████████| 10/10 [00:05<00:00,  1.99it/s]

2.2、手动进度更新：

tqdm可以手动更新，将其对象赋给一个变量，然后调用.update(N)方法来更新进度，tqdm()有个可选的参数设置迭代总数，然后通过update方法进行累加，每次执行update都会打印一次当前进度。

示例：新建一个tqdm实例，total=100表示迭代总数为100

percent = tqdm(total=100)

输出：

  0%|          | 0/100 [00:03<?, ?it/s]

调用update(N)方法，表示完成N次迭代，进度条则会显示对应的百分比

percent.update(1)

输出：

  1%|          | 1/100 [00:47<1:18:17, 47.45s/it]

再次调用会进行累加：

percent.update(90)

输出：

 91%|█████████ | 91/100 [01:35<00:08,  1.12it/s]