keras generator实战——cifar10数据集

1、cifar10数据集

Kaggle平台上的cifar10数据集的结构为：

其中test文件夹中包含有30 0000张图片，train文件夹中包含5 0000张图片

trainLabels.csv文件内容格式如下：

sampleSubmission.csv文件内容格式如下：

2、实战——keras generator读取cifar10数据集

class_names = [
    'airplane',
    'automobile',
    'bird',
    'cat',
    'deer',
    'dog',
    'frog',
    'horse',
    'ship',
    'truck',
]

train_lables_file = './cifar10/trainLabels.csv'
test_csv_file = './cifar10/sampleSubmission.csv'
train_folder = './cifar10/train/'
test_folder = './cifar10/test/'

def parse_csv_file(filepath, folder):
    """Parses csv files into (filename(全路径), label) format"""
    results = []
    with open(filepath, 'r') as f:
        lines = f.readlines()[1:]
    for line in lines:
        image_id, label_str = line.strip('\n').split(',')
        image_full_path = os.path.join(folder, image_id + '.png')
        results.append((image_full_path, label_str))
    return results

train_labels_info = parse_csv_file(train_lables_file, train_folder)
test_csv_info = parse_csv_file(test_csv_file, test_folder)

import pprint
pprint.pprint(train_labels_info[0:5])
pprint.pprint(test_csv_info[0:5])
print(len(train_labels_info), len(test_csv_info))

[('./cifar10/train/1.png', 'frog'),
 ('./cifar10/train/2.png', 'truck'),
 ('./cifar10/train/3.png', 'truck'),
 ('./cifar10/train/4.png', 'deer'),
 ('./cifar10/train/5.png', 'automobile')]
[('./cifar10/test/1.png', 'cat'),
 ('./cifar10/test/2.png', 'cat'),
 ('./cifar10/test/3.png', 'cat'),
 ('./cifar10/test/4.png', 'cat'),
 ('./cifar10/test/5.png', 'cat')]
50000 300000

"""
在这里，一般会先把训练集切分为两部分：训练集和验证集。调整模型，当模型调整好后，
再用全部的训练集训练模型，
再在test上预测

"""
# train_df = pd.DataFrame(train_labels_info)
#切分训练集和验证集
train_df = pd.DataFrame(train_labels_info[0:45000])
valid_df = pd.DataFrame(train_labels_info[45000:])
test_df = pd.DataFrame(test_csv_info)


#给DataFrame设定列名
train_df.columns = ['filepath', 'class']
valid_df.columns = ['filepath', 'class']
test_df.columns = ['filepath', 'class']

print(train_df.head())
print(valid_df.head())
print(test_df.head())

filepath       class
0  ./cifar10/train/1.png        frog
1  ./cifar10/train/2.png       truck
2  ./cifar10/train/3.png       truck
3  ./cifar10/train/4.png        deer
4  ./cifar10/train/5.png  automobile
                    filepath       class
0  ./cifar10/train/45001.png       horse
1  ./cifar10/train/45002.png  automobile
2  ./cifar10/train/45003.png        deer
3  ./cifar10/train/45004.png  automobile
4  ./cifar10/train/45005.png    airplane
               filepath class
0  ./cifar10/test/1.png   cat
1  ./cifar10/test/2.png   cat
2  ./cifar10/test/3.png   cat
3  ./cifar10/test/4.png   cat
4  ./cifar10/test/5.png   cat

注意下面代码（dataframe）与10_monkeys数据集中的directory不同：

height = 32
width = 32
channels = 3
batch_size = 32
num_classes = 10

train_datagen = keras.preprocessing.image.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,
    fill_mode = 'nearest',
)
train_generator = train_datagen.flow_from_dataframe(
    train_df,
    directory = './', #根目录：设成当前路径
    x_col = 'filepath',
    y_col = 'class',
    classes = class_names,
    target_size = (height, width),
    batch_size = batch_size,
    seed = 7,
    shuffle = True,
    class_mode = 'sparse',
)

valid_datagen = keras.preprocessing.image.ImageDataGenerator(
    rescale = 1./255)
valid_generator = valid_datagen.flow_from_dataframe(
    valid_df,
    directory = './', #当前文件夹
    x_col = 'filepath',
    y_col = 'class',
    classes = class_names, #依据class_names将y_col转为ID
    target_size = (height, width),
    batch_size = batch_size,
    seed = 7,
    shuffle = False,
    class_mode = "sparse")

train_num = train_generator.samples
valid_num = valid_generator.samples
print(train_num, valid_num)

Found 45000 validated image filenames belonging to 10 classes.
Found 5000 validated image filenames belonging to 10 classes.
45000 5000

for i in range(2):
    x, y = train_generator.next()
    print(x.shape, y.shape)
    print(y)

(32, 32, 32, 3) (32,)
[2. 1. 4. 4. 4. 4. 6. 5. 2. 8. 4. 6. 6. 3. 7. 1. 7. 2. 8. 8. 3. 0. 5. 3.
 9. 1. 4. 5. 6. 7. 9. 2.]
(32, 32, 32, 3) (32,)
[0. 7. 2. 7. 5. 5. 7. 0. 5. 4. 9. 7. 6. 3. 0. 4. 4. 4. 6. 3. 5. 4. 6. 6.
 4. 1. 8. 2. 4. 4. 3. 0.]

3、实现——模型训练与预测

网络结构：

Model: "sequential"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
conv2d (Conv2D)              (None, 32, 32, 128)       3584      
_________________________________________________________________
batch_normalization (BatchNo (None, 32, 32, 128)       512       
_________________________________________________________________
conv2d_1 (Conv2D)            (None, 32, 32, 128)       147584    
_________________________________________________________________
batch_normalization_1 (Batch (None, 32, 32, 128)       512       
_________________________________________________________________
max_pooling2d (MaxPooling2D) (None, 16, 16, 128)       0         
_________________________________________________________________
conv2d_2 (Conv2D)            (None, 16, 16, 256)       295168    
_________________________________________________________________
batch_normalization_2 (Batch (None, 16, 16, 256)       1024      
_________________________________________________________________
conv2d_3 (Conv2D)            (None, 16, 16, 256)       590080    
_________________________________________________________________
batch_normalization_3 (Batch (None, 16, 16, 256)       1024      
_________________________________________________________________
max_pooling2d_1 (MaxPooling2 (None, 8, 8, 256)         0         
_________________________________________________________________
conv2d_4 (Conv2D)            (None, 8, 8, 512)         1180160   
_________________________________________________________________
batch_normalization_4 (Batch (None, 8, 8, 512)         2048      
_________________________________________________________________
conv2d_5 (Conv2D)            (None, 8, 8, 512)         2359808   
_________________________________________________________________
batch_normalization_5 (Batch (None, 8, 8, 512)         2048      
_________________________________________________________________
max_pooling2d_2 (MaxPooling2 (None, 4, 4, 512)         0         
_________________________________________________________________
flatten (Flatten)            (None, 8192)              0         
_________________________________________________________________
dense (Dense)                (None, 512)               4194816   
_________________________________________________________________
dense_1 (Dense)              (None, 10)                5130      
=================================================================
Total params: 8,783,498
Trainable params: 8,779,914
Non-trainable params: 3,584
_________________________________________________________________

epochs = 5
history = model.fit_generator(train_generator,
                              steps_per_epoch = train_num // batch_size,
                              epochs = epochs,
                              validation_data = valid_generator,
                              validation_steps = valid_num // batch_size)

#用训练好的模型在测试集上进行测试
test_datagen = keras.preprocessing.image.ImageDataGenerator(
    rescale = 1./255)
test_generator = test_datagen.flow_from_dataframe(
    test_df,
    directory = './',
    x_col = 'filepath',
    y_col = 'class',
    classes = class_names,
    target_size = (height, width),
    batch_size = batch_size,
    seed = 7,
    shuffle = False,
    class_mode = "sparse")
test_num = test_generator.samples
print(test_num) 

Found 300000 validated image filenames belonging to 10 classes.
300000

for i in range(2):
    x, y = test_generator.next()
    print(x.shape, y.shape) #(32, 32, 32, 3) (32,)
    print(y)

(32, 32, 32, 3) (32,)
[3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3.
 3. 3. 3. 3. 3. 3. 3. 3.]
(32, 32, 32, 3) (32,)
[3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3. 3.
 3. 3. 3. 3. 3. 3. 3. 3.]

#预测：test_predict.shape:（30 0000, 10）
test_predict = model.predict_generator(test_generator,
                                       workers = 5,#并行度
                                       use_multiprocessing = False) #多进程做并行化

#使用最大概率的索引来获得它对应的类别预测
test_predict_class_indices = np.argmax(test_predict, axis = 1)#获得30 0000个整数,对应最大值的位置

test_predict_class = [class_names[index] 
                      for index in test_predict_class_indices]

#写到submission文件中去
def generate_submissions(filename, predict_class):
    with open(filename, 'w') as f:
        f.write('id,label\n')
        for i in range(len(predict_class)):
            f.write('%d,%s\n' % (i+1, predict_class[i]))

output_file = "./cifar10/submission.csv"
generate_submissions(output_file, test_predict_class)

附代码：

%matplotlib inline
import matplotlib as mpl
import matplotlib.pyplot as plt
import numpy as np
import os
import pandas as pd
import sklearn
import sys
import tensorflow as tf
import time

from tensorflow import keras

print(tf.__version__)
print(sys.version_info)
for module in mpl, np, pd, sklearn, tf, keras:
    print(module.__name__, module.__version__)

physical_devices = tf.config.experimental.list_physical_devices('GPU')
assert len(physical_devices) > 0, "Not enough GPU hardware devices available"
tf.config.experimental.set_memory_growth(physical_devices[0], True)

class_names = [
    'airplane',
    'automobile',
    'bird',
    'cat',
    'deer',
    'dog',
    'frog',
    'horse',
    'ship',
    'truck',
]

train_lables_file = './cifar10/trainLabels.csv'
test_csv_file = './cifar10/sampleSubmission.csv'
train_folder = './cifar10/train/'
test_folder = './cifar10/test/'

def parse_csv_file(filepath, folder):
    """Parses csv files into (filename(全路径), label) format"""
    results = []
    with open(filepath, 'r') as f:
        lines = f.readlines()[1:]
    for line in lines:
        image_id, label_str = line.strip('\n').split(',')
        image_full_path = os.path.join(folder, image_id + '.png')
        results.append((image_full_path, label_str))
    return results

train_labels_info = parse_csv_file(train_lables_file, train_folder)
test_csv_info = parse_csv_file(test_csv_file, test_folder)

import pprint
pprint.pprint(train_labels_info[0:5])
pprint.pprint(test_csv_info[0:5])
print(len(train_labels_info), len(test_csv_info))

"""
在这里，一般会先把训练集切分为两部分：训练集和验证集。调整模型，当模型调整好后，
再用全部的训练集训练模型，
再在test上预测

"""
# train_df = pd.DataFrame(train_labels_info)
#切分训练集和验证集
train_df = pd.DataFrame(train_labels_info[0:45000])
valid_df = pd.DataFrame(train_labels_info[45000:])
test_df = pd.DataFrame(test_csv_info)

print(train_df.head())
print(valid_df.head())
print(test_df.head())
print(train_df.shape)
print(valid_df.shape)
print(test_df.shape)

#给DataFrame设定列名
train_df.columns = ['filepath', 'class']
valid_df.columns = ['filepath', 'class']
test_df.columns = ['filepath', 'class']

print(train_df.head())
print(valid_df.head())
print(test_df.head())

height = 32
width = 32
channels = 3
batch_size = 32
num_classes = 10

train_datagen = keras.preprocessing.image.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,
    fill_mode = 'nearest',
)
train_generator = train_datagen.flow_from_dataframe(
    train_df,
    directory = './', #根目录：设成当前路径
    x_col = 'filepath',
    y_col = 'class',
    classes = class_names,
    target_size = (height, width),
    batch_size = batch_size,
    seed = 7,
    shuffle = True,
    class_mode = 'sparse',
)

valid_datagen = keras.preprocessing.image.ImageDataGenerator(
    rescale = 1./255)
valid_generator = valid_datagen.flow_from_dataframe(
    valid_df,
    directory = './', #当前文件夹
    x_col = 'filepath',
    y_col = 'class',
    classes = class_names, #依据class_names将y_col转为ID
    target_size = (height, width),
    batch_size = batch_size,
    seed = 7,
    shuffle = False,
    class_mode = "sparse")

train_num = train_generator.samples
valid_num = valid_generator.samples
print(train_num, valid_num)

#还没有做数据增强？

for i in range(2):
    x, y = train_generator.next()
    print(x.shape, y.shape)
    print(y)
model = keras.models.Sequential([
    keras.layers.Conv2D(filters=128, kernel_size=3, padding='same',
                        activation='relu', 
                        input_shape=[width, height, channels]),
    keras.layers.BatchNormalization(),
    keras.layers.Conv2D(filters=128, kernel_size=3, padding='same',
                        activation='relu'),
    keras.layers.BatchNormalization(),
    keras.layers.MaxPool2D(pool_size=2),
    
    keras.layers.Conv2D(filters=256, kernel_size=3, padding='same',
                        activation='relu'),
    keras.layers.BatchNormalization(),
    keras.layers.Conv2D(filters=256, kernel_size=3, padding='same',
                        activation='relu'),
    keras.layers.BatchNormalization(),
    keras.layers.MaxPool2D(pool_size=2),
    keras.layers.Conv2D(filters=512, kernel_size=3, padding='same',
                        activation='relu'),
    keras.layers.BatchNormalization(),
    keras.layers.Conv2D(filters=512, kernel_size=3, padding='same',
                        activation='relu'),
    keras.layers.BatchNormalization(),
    keras.layers.MaxPool2D(pool_size=2),
    keras.layers.Flatten(),
    keras.layers.Dense(512, activation='relu'),
    keras.layers.Dense(num_classes, activation='softmax'),
])

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

epochs = 5
history = model.fit_generator(train_generator,
                              steps_per_epoch = train_num // batch_size,
                              epochs = epochs,
                              validation_data = valid_generator,
                              validation_steps = valid_num // batch_size)

def plot_learning_curves(history, label, epcohs, min_value, max_value):
    data = {}
    data[label] = history.history[label]
    data['val_'+label] = history.history['val_'+label]
    pd.DataFrame(data).plot(figsize=(8, 5))
    plt.grid(True)
    plt.axis([0, epochs, min_value, max_value])
    plt.show()
    
plot_learning_curves(history, 'accuracy', epochs, 0, 1)
plot_learning_curves(history, 'loss', epochs, 0, 2)

#用训练好的模型在测试集上进行测试
test_datagen = keras.preprocessing.image.ImageDataGenerator(
    rescale = 1./255)
test_generator = test_datagen.flow_from_dataframe(
    test_df,
    directory = './',
    x_col = 'filepath',
    y_col = 'class',
    classes = class_names,
    target_size = (height, width),
    batch_size = batch_size,
    seed = 7,
    shuffle = False,
    class_mode = "sparse")
test_num = test_generator.samples
print(test_num) 

for i in range(2):
    x, y = test_generator.next()
    print(x.shape, y.shape) #(32, 32, 32, 3) (32,)
    print(y)

#预测
test_predict = model.predict_generator(test_generator,
                                       workers = 5,#并行度
                                       use_multiprocessing = False) #多进程做并行化

#使用最大概率的索引来获得它对应的类别预测
test_predict_class_indices = np.argmax(test_predict, axis = 1)#获得30 0000个整数,对应最大值的位置

test_predict_class = [class_names[index] 
                      for index in test_predict_class_indices]

#写到submission文件中去
def generate_submissions(filename, predict_class):
    with open(filename, 'w') as f:
        f.write('id,label\n')
        for i in range(len(predict_class)):
            f.write('%d,%s\n' % (i+1, predict_class[i]))

output_file = "./cifar10/submission.csv"
generate_submissions(output_file, test_predict_class)