Keras 使用自己的数据分类，并使用tensorboard记录的简单实例

Keras 使用自己的数据分类，并使用tensorboard记录的简单实例

1.使用的分类图片按照不同类别保存在不同文件夹子中，并且切分好训练集和测试集，如下图显示

注意：文件名建议使用标签名

from keras.models import Sequential
import keras
from keras.preprocessing.image import ImageDataGenerator
from keras.layers import Dense, Dropout, Flatten
from keras.layers import Conv2D, MaxPooling2D

from keras.callbacks import TensorBoard
import time
import os
import tensorflow as tf
#指定GPU，限制GPU内存
os.environ["CUDA_VISIBLE_DEVICES"] = "0"
from keras.backend.tensorflow_backend import set_session
config = tf.ConfigProto()
config.gpu_options.per_process_gpu_memory_fraction = 0.7
set_session(tf.Session(config=config))


BATCHSIZE = 100
IMG_SIZE = (100,100)

#训练集，测试集文件路径
train_path = '../data/train'
test_path = '../data/test'


s_time = time.strftime("%Y%m%d%H%M%S", time.localtime())  #时间戳


#image_batch_generator

train_datagen = ImageDataGenerator(
         rescale=1./255)

test_datagen = ImageDataGenerator( rescale=1./255 )


#训练集batch生成器
train_generator = train_datagen.flow_from_directory(
        train_path,
        target_size=IMG_SIZE,
        batch_size=BATCHSIZE,
        color_mode='grayscale',
        classes=['original','tampered'],
        class_mode='categorical')

#测试集batch生成器
validation_generator = test_datagen.flow_from_directory(
        test_path,
        target_size=IMG_SIZE,
        color_mode='grayscale',
        batch_size=BATCHSIZE,
        classes=['original','tampered'],
        class_mode='categorical')



#网络结构
model = Sequential()
model.add(Conv2D(16, (3, 3), activation='relu', input_shape=(IMG_SIZE[0],IMG_SIZE[1],1)))
model.add(Conv2D(16, (3, 3), activation='relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Dropout(0.25))

model.add(Conv2D(32, (3, 3), activation='relu'))
#model.add(Conv2D(64, (3, 3), activation='relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Dropout(0.25))

model.add(Flatten())
model.add(Dense(128, activation='relu'))
model.add(Dropout(0.5))
model.add(Dense(2, activation='softmax'))

#优化器
adam = keras.optimizers.Adam(lr=0.001, beta_1=0.9, beta_2=0.999, epsilon=1e-08)
model.compile(loss='binary_crossentropy',
              optimizer=adam,
              metrics=['accuracy'])

#logs文件路径
logs_path = 'F:/zy/logs/log_%s'%(s_time)


try:
    os.makedirs(logs_path)
except:
    pass

#将loss ，acc， val_loss ,val_acc记录tensorboard
tensorboard = TensorBoard(log_dir=logs_path, histogram_freq=1,write_graph=True,write_batch_performance=True)

#模型训练
model.fit_generator(
        train_generator,
        steps_per_epoch=60,
        epochs=50,
        verbose=1,
        validation_data=validation_generator,
        validation_steps=60,
        callbacks=[tensorboard]
)