【机器学习13】狗猫数据集的两阶段分类实验
1.Anaconda 安装Tensorflow
pip install tensorflow -i https://pypi.tuna.tsinghua.edu.cn/simple/
2.Anaconda安装keras
pip install --upgrade tensorflow -i https://pypi.tuna.tsinghua.edu.cn/simple
pip install keras -i https://pypi.tuna.tsinghua.edu.cn/simple
3.直接训练
import keras
keras.__version__
'2.3.1'
import os, shutil
# The path to the directory where the original
# dataset was uncompressed
original_dataset_dir = './mycatdog'
# The directory where we will
# store our smaller dataset
base_dir = './out1'
os.mkdir(base_dir)
# Directories for our training,
# validation and test splits
train_dir = os.path.join(base_dir, 'train')
os.mkdir(train_dir)
validation_dir = os.path.join(base_dir, 'validation')
os.mkdir(validation_dir)
test_dir = os.path.join(base_dir, 'test')
os.mkdir(test_dir)
# Directory with our training cat pictures
train_cats_dir = os.path.join(train_dir, 'cats')
os.mkdir(train_cats_dir)
# Directory with our training dog pictures
train_dogs_dir = os.path.join(train_dir, 'dogs')
os.mkdir(train_dogs_dir)
# Directory with our validation cat pictures
validation_cats_dir = os.path.join(validation_dir, 'cats')
os.mkdir(validation_cats_dir)
# Directory with our validation dog pictures
validation_dogs_dir = os.path.join(validation_dir, 'dogs')
os.mkdir(validation_dogs_dir)
# Directory with our validation cat pictures
test_cats_dir = os.path.join(test_dir, 'cats')
os.mkdir(test_cats_dir)
# Directory with our validation dog pictures
test_dogs_dir = os.path.join(test_dir, 'dogs')
os.mkdir(test_dogs_dir)
# Copy first 1000 cat images to train_cats_dir
fnames = ['cat.{}.jpg'.format(i) for i in range(10)]
for fname in fnames:
src = os.path.join(original_dataset_dir, fname)
dst = os.path.join(train_cats_dir, fname)
shutil.copyfile(src, dst)
# Copy next 500 cat images to validation_cats_dir
fnames = ['cat.{}.jpg'.format(i) for i in range(10, 20)]
for fname in fnames:
src = os.path.join(original_dataset_dir, fname)
dst = os.path.join(validation_cats_dir, fname)
shutil.copyfile(src, dst)
# Copy next 500 cat images to test_cats_dir
fnames = ['cat.{}.jpg'.format(i) for i in range(20, 30)]
for fname in fnames:
src = os.path.join(original_dataset_dir, fname)
dst = os.path.join(test_cats_dir, fname)
shutil.copyfile(src, dst)
# Copy first 1000 dog images to train_dogs_dir
fnames = ['dog.{}.jpg'.format(i) for i in range(10)]
for fname in fnames:
src = os.path.join(original_dataset_dir, fname)
dst = os.path.join(train_dogs_dir, fname)
shutil.copyfile(src, dst)
# Copy next 500 dog images to validation_dogs_dir
fnames = ['dog.{}.jpg'.format(i) for i in range(10, 20)]
for fname in fnames:
src = os.path.join(original_dataset_dir, fname)
dst = os.path.join(validation_dogs_dir, fname)
shutil.copyfile(src, dst)
# Copy next 500 dog images to test_dogs_dir
fnames = ['dog.{}.jpg'.format(i) for i in range(20, 30)]
for fname in fnames:
src = os.path.join(original_dataset_dir, fname)
dst = os.path.join(test_dogs_dir, fname)
shutil.copyfile(src, dst)
将创建一个包含三个子集的新数据集：
一个训练集，每个类别有1000个样本；
一个验证集，每个类别有500个样本；
一个测试集，每个类别有500个样本。
print('total training cat images:', len(os.listdir(train_cats_dir)))
total training cat images: 10
print('total training dog images:', len(os.listdir(train_dogs_dir)))
total training dog images: 10
print('total validation cat images:', len(os.listdir(validation_cats_dir)))
total validation cat images: 10
print('total validation dog images:', len(os.listdir(validation_dogs_dir)))
total validation dog images: 10
print('total test cat images:', len(os.listdir(test_cats_dir)))
total test cat images: 10
print('total test dog images:', len(os.listdir(test_dogs_dir)))
total test dog images: 10
使用生成器将模型拟合到数据。
在训练后始终保存模型。
在训练期间 在训练和验证数据上绘制模型的损失和准确性。
from keras import layers
from keras import models
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'))
model.add(layers.Dense(1, activation='sigmoid'))
model.summary()
Model: "sequential_1"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
conv2d_1 (Conv2D) (None, 148, 148, 32) 896
_________________________________________________________________
max_pooling2d_1 (MaxPooling2 (None, 74, 74, 32) 0
_________________________________________________________________
conv2d_2 (Conv2D) (None, 72, 72, 64) 18496
_________________________________________________________________
max_pooling2d_2 (MaxPooling2 (None, 36, 36, 64) 0
_________________________________________________________________
conv2d_3 (Conv2D) (None, 34, 34, 128) 73856
_________________________________________________________________
max_pooling2d_3 (MaxPooling2 (None, 17, 17, 128) 0
_________________________________________________________________
conv2d_4 (Conv2D) (None, 15, 15, 128) 147584
_________________________________________________________________
max_pooling2d_4 (MaxPooling2 (None, 7, 7, 128) 0
_________________________________________________________________
flatten_1 (Flatten) (None, 6272) 0
_________________________________________________________________
dense_1 (Dense) (None, 512) 3211776
_________________________________________________________________
dense_2 (Dense) (None, 1) 513
=================================================================
Total params: 3,453,121
Trainable params: 3,453,121
Non-trainable params: 0
_________________________________________________________________
from keras import optimizers
model.compile(loss='binary_crossentropy',
optimizer=optimizers.RMSprop(lr=1e-4),
metrics=['acc'])
数据预处理，Keras拥有实用程序来自动完成这些步骤：
1.读取图片文件。
2.将JPEG内容解码为RBG像素网格。
3.将它们转换为浮点张量。
4.将像素值（0到255之间）重新缩放为[0，1]间隔（如您所知，神经网络更喜欢处理较小的输入值）。
from keras.preprocessing.image import ImageDataGenerator
# All images will be rescaled by 1./255
train_datagen = ImageDataGenerator(rescale=1./255)
test_datagen = ImageDataGenerator(rescale=1./255)
train_generator = train_datagen.flow_from_directory(
# This is the target directory
train_dir,
# All images will be resized to 150x150
target_size=(150, 150),
batch_size=20,
# Since we use binary_crossentropy loss, we need binary labels
class_mode='binary')
validation_generator = test_datagen.flow_from_directory(
validation_dir,
target_size=(150, 150),
batch_size=20,
class_mode='binary')
Found 20 images belonging to 2 classes.
Found 20 images belonging to 2 classes.
for data_batch, labels_batch in train_generator:
print('data batch shape:', data_batch.shape)
print('labels batch shape:', labels_batch.shape)
break
data batch shape: (20, 150, 150, 3)
labels batch shape: (20,)
history = model.fit_generator(
train_generator,
steps_per_epoch=80,
epochs=30,
validation_data=validation_generator,
validation_steps=50)
Epoch 1/30
80/80 [==============================] - 48s 605ms/step - loss: 0.2500 - acc: 0.9200 - val_loss: 0.6217 - val_acc: 0.6000
Epoch 2/30
80/80 [==============================] - 45s 560ms/step - loss: 0.0012 - acc: 1.0000 - val_loss: 1.0331 - val_acc: 0.5500
.........
Epoch 30/30
80/80 [==============================] - 40s 501ms/step - loss: 1.0213e-08 - acc: 1.0000 - val_loss: 2.7475 - val_acc: 0.5500
model.save('cats_and_dogs_small_1.h5')
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(len(acc))
plt.plot(epochs, acc, 'bo', label='Training acc')
plt.plot(epochs, val_acc, 'b', label='Validation acc')
plt.title('Training and validation 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 validation loss')
plt.legend()
plt.show()
4.使用数据增强
过度拟合是由于要学习的样本太少而导致的，通过许多随机变换来“增强”样本，在训练时，我们的模型永远不会看到相同的图片两次。
过拟合：
过拟合是指为了得到一致假设而使假设变得过度严格。避免过拟合是分类器设计中的一个核心任务。通常采用增大数据量和测试样本集的方法对分类器性能进行评价。
数据增强：
你的数据太少？
别怕，我们有数据增强技术。翻转，旋转，缩放比例，剪裁，移位，高斯噪声，颜色增强，条件型生成对抗网络。
datagen = ImageDataGenerator(
rotation_range=40,
width_shift_range=0.2,
height_shift_range=0.2,
shear_range=0.2,
zoom_range=0.2,
horizontal_flip=True,
fill_mode='nearest')
在Keras中，这可以通过配置要对ImageDataGenerator 实例读取的图像执行的随机转换来完成。
# This is module with image preprocessing utilities
from keras.preprocessing import image
fnames = [os.path.join(train_cats_dir, fname) for fname in os.listdir(train_cats_dir)]
# We pick one image to "augment"
img_path = fnames[3]
# Read the image and resize it
img = image.load_img(img_path, target_size=(150, 150))
# Convert it to a Numpy array with shape (150, 150, 3)
x = image.img_to_array(img)
# Reshape it to (1, 150, 150, 3)
x = x.reshape((1,) + x.shape)
# The .flow() command below generates batches of randomly transformed images.
# It will loop indefinitely, so we need to `break` the loop at some point!
i = 0
for batch in datagen.flow(x, batch_size=1):
plt.figure(i)
imgplot = plt.imshow(image.array_to_img(batch[0]))
i += 1
if i % 4 == 0:
break
plt.show()
在紧密连接的分类器之前为模型添加一个Dropout层：
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.Dropout(0.5))
model.add(layers.Dense(512, activation='relu'))
model.add(layers.Dense(1, activation='sigmoid'))
model.compile(loss='binary_crossentropy',
optimizer=optimizers.RMSprop(lr=1e-4),
metrics=['acc'])
train_datagen = ImageDataGenerator(
rescale=1./255,
rotation_range=40,
width_shift_range=0.2,
height_shift_range=0.2,
shear_range=0.2,
zoom_range=0.2,
horizontal_flip=True,)
# Note that the validation data should not be augmented!
test_datagen = ImageDataGenerator(rescale=1./255)
train_generator = train_datagen.flow_from_directory(
# This is the target directory
train_dir,
# All images will be resized to 150x150
target_size=(150, 150),
batch_size=32,
# Since we use binary_crossentropy loss, we need binary labels
class_mode='binary')
validation_generator = test_datagen.flow_from_directory(
validation_dir,
target_size=(150, 150),
batch_size=32,
class_mode='binary')
history = model.fit_generator(
train_generator,
steps_per_epoch=50,
epochs=80,
validation_data=validation_generator,
validation_steps=50)
Found 20 images belonging to 2 classes.
Found 20 images belonging to 2 classes.
Epoch 1/80
50/50 [==============================] - 29s 573ms/step - loss: 0.6646 - acc: 0.6150 - val_loss: 0.8158 - val_acc: 0.5500
Epoch 2/80
50/50 [==============================] - 29s 587ms/step - loss: 0.5737 - acc: 0.6990 - val_loss: 0.9219 - val_acc: 0.3500
................
Epoch 80/80
50/50 [==============================] - 28s 559ms/step - loss: 0.0023 - acc: 0.9990 - val_loss: 13.6142 - val_acc: 0.4000
保存模型-在convnet可视化部分中使用它。
model.save('cats_and_dogs_small_2.h5')
acc = history.history['acc']
val_acc = history.history['val_acc']
loss = history.history['loss']
val_loss = history.history['val_loss']
epochs = range(len(acc))
plt.plot(epochs, acc, 'bo', label='Training acc')
plt.plot(epochs, val_acc, 'b', label='Validation acc')
plt.title('Training and validation 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 validation loss')
plt.legend()
plt.show()
209
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
