猫狗分类初级

import os, shutil

# 下载数据集，解压
original_dataset_dir = 'train/'

# 设置数据集各级目录名称
base_dir = 'data'
data_dirs = ['train', 'validation', 'test']
class_dirs = ['cat', 'dog']

# 分别设置训练。验证和测试数据集的图像文件区间
images = {
    'train': [0, 1000],  # 1000张图片做训练
    'validation': [1000, 1500],  # 500张图片做验证
    'test': [1500, 2000],    # 500张图片做测试
}

for d in data_dirs:
    for c in class_dirs:
        # 构造数据集目录， 格式实例：’data/train/dog'
        dataset_dir = f"{base_dir}/{d}/{c}/"
        os.makedirs(dataset_dir, exist_ok=True)
        # 构造图像文件名称，格式示例：‘cat.1.jpg'
        fnames = [f'{c}.{i}.jpg' for i in range(images[d][0], images[d][1])]
        # 将挑选的图像文件拷贝到对应的数据集目录下
        for fname in fnames:
            shutil.copyfile(original_dataset_dir+fname, dataset_dir+fname)

from keras import models, layers, optimizers

model = models.Sequential()

model.add(layers.Conv2D(32, (3, 3), activation='relu', input_shape=(150, 150, 3)))
model.add(layers.MaxPooling2D((2, 2)))
model.add(layers.Conv2D(64, (3, 3), activation='relu'))
model.add(layers.MaxPooling2D((2, 2)))
model.add(layers.Conv2D(128, (3, 3), activation='relu'))
model.add(layers.MaxPooling2D((2, 2)))
model.add(layers.Conv2D(128, (3, 3), activation='relu'))
model.add(layers.MaxPooling2D((2, 2)))
model.add(layers.Flatten())

model.add(layers.Dense(512, activation='relu'))     # 加入两层全连接层用于分类
model.add(layers.Dense(1, activation='sigmoid'))

model.summary()
# 模型编译
model.compile(loss='binary_crossentropy',
              optimizer=optimizers.RMSprop(lr=1e4-4),
              metrics=['acc'])

from keras.preprocessing.image import ImageDataGenerator

# 使用 ImageDataGenerator 从目录中读取图像
train_datagen = ImageDataGenerator(rescale=1./255)  # 值标准化，解码为RGB
test_datagen = ImageDataGenerator(rescale=1./255)

# 创建训练集数据生成器， 猫和狗各1000张
train_generator = train_datagen.flow_from_directory(
    'data/train',
    target_size=(150, 150),
    batch_size=20,
    class_mode='binary'
)
# 创建测试集数据生成器， 猫和狗各500张
test_generator = test_datagen.flow_from_directory(
    'data/test',
    target_size=(150, 150),
    batch_size=20,
    class_mode='binary'
)

for data_batch, labels_batch in train_generator:
    print(f'data_batch.shape:{data_batch.shape}')
    print(f'label_batch.shape:{labels_batch.shape}')
    break

# 模型训练
history = model.fit_generator(train_generator,
                              steps_per_epoch=50,
                              epochs=30,    # 原本是30轮的，电脑承受不住，改成10轮了
                              validation_data=test_generator,
                              validation_steps=50)

model.save('cats_and_dogs_small_1.h5')

# 可视化对比

import matplotlib.pyplot as plt

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

epochs = range(1, len(acc) + 1)

plt.plot(epochs, acc, 'bo', label='Training acc')
plt.plot(epochs, val_acc, 'b', label='Validataion acc')
plt.title('Training and validataion accuracy')
plt.legend()

plt.figure()

plt.plot(epochs, loss, 'bo', label= 'Training loss')
plt.plot(epochs, val_loss, 'b', label= 'Validation loss')
plt.title("Training and validataion loss")
plt.legend()

plt.show()