回调函数是一组在训练的特定阶段被调用的函数集，你可以使用回调函数来观察训练过程中网络内部的状态和统计信息。通过传递回调函数列表到模型的.fit()中，即可在给定的训练阶段调用该函数集中的函数。
【Tips】虽然我们称之为回调“函数”，但事实上Keras的回调函数是一个类，回调函数只是习惯性称呼
Callback
例子No.1：官网示例
保存训练过程中最好的模型
from __future__ import print_function
import numpy as np
from keras.datasets import mnist
from keras.models import Sequential
from keras.layers.core import Dense, Activation
from keras.optimizers import SGD
from keras.utils import np_utils
from keras.callbacks import ModelCheckpoint
from keras.callbacks import TensorBoard
import matplotlib.pyplot as plt
np.random.seed(1671) # for reproducibility
# network and training
NB_EPOCH = 20
BATCH_SIZE = 128
VERBOSE = 1
NB_CLASSES = 10 # number of outputs = number of digits
OPTIMIZER = SGD() # optimizer, explained later in this chapter
N_HIDDEN = 128
VALIDATION_SPLIT=0.2 # how much TRAIN is reserved for VALIDATION
# data: shuffled and split between train and test sets
(X_train, y_train), (X_test, y_test) = mnist.load_data()
#X_train is 60000 rows of 28x28 values --> reshaped in 60000 x 784
RESHAPED = 784
#
X_train = X_train.reshape(60000, RESHAPED)
X_test = X_test.reshape(10000, RESHAPED)
X_train = X_train.astype('float32')
X_test = X_test.astype('float32')
# normalize
X_train /= 255
X_test /= 255
print(X_train.shape[0], 'train samples')
print(X_test.shape[0], 'test samples')
# convert class vectors to binary class matrices
Y_train = np_utils.to_categorical(y_train, NB_CLASSES)
Y_test = np_utils.to_categorical(y_test, NB_CLASSES)
# M_HIDDEN hidden layers
# 10 outputs
# final stage is softmax
model = Sequential()
model.add(Dense(N_HIDDEN, input_shape=(RESHAPED,)))
model.add(Activation('relu'))
model.add(Dense(N_HIDDEN))
model.add(Activation('relu'))
model.add(Dense(NB_CLASSES))
model.add(Activation('softmax'))
model.summary()
#model.load_weights("weight_bst.hdf5")
model.compile(loss='categorical_crossentropy',
optimizer=OPTIMIZER,
metrics=['accuracy'])
#保存.fit()过程中最好的模型
checkpoint=ModelCheckpoint("/weights.hdf5",verbose=1,save_best_only=True)
history = model.fit(X_train, Y_train,
batch_size=BATCH_SIZE, epochs=5,
verbose=VERBOSE, validation_split=VALIDATION_SPLIT,
callbacks=[checkpoint])
例子No.2：保存每一个改进CheckPoint模型
应用Checkpoint时，应在每次训练中观察到改进时输出模型权重。
Checkpoint设置成当验证数据集的分类精度提高时保存网络权重(monitor=’val_acc’ and mode=’max’)。权重存储在一个包含评价的文件中(weights-improvement – { val_acc = .2f } .hdf5)。
from __future__ import print_function
import numpy as np
from keras.datasets import mnist
from keras.models import Sequential
from keras.layers.core import Dense, Activation
from keras.optimizers import SGD
from keras.utils import np_utils
from keras.callbacks import ModelCheckpoint
from keras.callbacks import TensorBoard
import matplotlib.pyplot as plt
np.random.seed(1671) # for reproducibility
# network and training
NB_EPOCH = 20
BATCH_SIZE = 128
VERBOSE = 1
NB_CLASSES = 10 # number of outputs = number of digits
OPTIMIZER = SGD() # optimizer, explained later in this chapter
N_HIDDEN = 128
VALIDATION_SPLIT=0.2 # how much TRAIN is reserved for VALIDATION
# data: shuffled and split between train and test sets
(X_train, y_train), (X_test, y_test) = mnist.load_data()
#X_train is 60000 rows of 28x28 values --> reshaped in 60000 x 784
RESHAPED = 784
#
X_train = X_train.reshape(60000, RESHAPED)
X_test = X_test.reshape(10000, RESHAPED)
X_train = X_train.astype('float32')
X_test = X_test.astype('float32')
# normalize
X_train /= 255
X_test /= 255
print(X_train.shape[0], 'train samples')
print(X_test.shape[0], 'test samples')
# convert class vectors to binary class matrices
Y_train = np_utils.to_categorical(y_train, NB_CLASSES)
Y_test = np_utils.to_categorical(y_test, NB_CLASSES)
# M_HIDDEN hidden layers
# 10 outputs
# final stage is softmax
model = Sequential()
model.add(Dense(N_HIDDEN, input_shape=(RESHAPED,)))
model.add(Activation('relu'))
model.add(Dense(N_HIDDEN))
model.add(Activation('relu'))
model.add(Dense(NB_CLASSES))
model.add(Activation('softmax'))
model.summary()
#model.load_weights("weight_bst.hdf5")
model.compile(loss='categorical_crossentropy',
optimizer=OPTIMIZER,
metrics=['accuracy'])
#保存.fit()训练过程中，精确度提高的模型
filepath="weights-improvement-{epoch:02d}-{val_acc:.2f}.hdf5"
checkpoint=ModelCheckpoint(filepath,monitor='val_acc',verbose=1,save_best_only=True,mode='max')
history = model.fit(X_train, Y_train,
batch_size=BATCH_SIZE, epochs=5,
verbose=VERBOSE, validation_split=VALIDATION_SPLIT,
callbacks=[checkpoint])
程序运行结果：
在epoch=5次训练中，val_acc均提高，所以每次都保存了一个新的模型
例子No.3：保存最佳的Checkpoint模型
如果验证精度提高的话，一个更简单的Checkpoint策略是将模型权重保存到相同的文件中。
这可以使用上述相同的代码轻松完成，并将输出文件名更改为固定（不包括评价或次数的信息）。
在这种情况下，只有当验证数据集上的模型的分类精度提高到到目前为止最好的时候，才会将模型权重写入文件“weights.best.hdf5”。
from __future__ import print_function
import numpy as np
from keras.datasets import mnist
from keras.models import Sequential
from keras.layers.core import Dense, Activation
from keras.optimizers import SGD
from keras.utils import np_utils
from keras.callbacks import ModelCheckpoint
from keras.callbacks import TensorBoard
import matplotlib.pyplot as plt
np.random.seed(1671) # for reproducibility
# network and training
NB_EPOCH = 20
BATCH_SIZE = 128
VERBOSE = 1
NB_CLASSES = 10 # number of outputs = number of digits
OPTIMIZER = SGD() # optimizer, explained later in this chapter
N_HIDDEN = 128
VALIDATION_SPLIT=0.2 # how much TRAIN is reserved for VALIDATION
# data: shuffled and split between train and test sets
(X_train, y_train), (X_test, y_test) = mnist.load_data()
#X_train is 60000 rows of 28x28 values --> reshaped in 60000 x 784
RESHAPED = 784
#
X_train = X_train.reshape(60000, RESHAPED)
X_test = X_test.reshape(10000, RESHAPED)
X_train = X_train.astype('float32')
X_test = X_test.astype('float32')
# normalize
X_train /= 255
X_test /= 255
print(X_train.shape[0], 'train samples')
print(X_test.shape[0], 'test samples')
# convert class vectors to binary class matrices
Y_train = np_utils.to_categorical(y_train, NB_CLASSES)
Y_test = np_utils.to_categorical(y_test, NB_CLASSES)
# M_HIDDEN hidden layers
# 10 outputs
# final stage is softmax
model = Sequential()
model.add(Dense(N_HIDDEN, input_shape=(RESHAPED,)))
model.add(Activation('relu'))
model.add(Dense(N_HIDDEN))
model.add(Activation('relu'))
model.add(Dense(NB_CLASSES))
model.add(Activation('softmax'))
model.summary()
#model.load_weights("weight_bst.hdf5")
model.compile(loss='categorical_crossentropy',
optimizer=OPTIMIZER,
metrics=['accuracy'])
#保存训练过程中最好的模型
filepath="weight_bst.hdf5"
checkpoint=ModelCheckpoint(filepath,monitor='val_acc',verbose=1,save_best_only=True,mode='max')
history = model.fit(X_train, Y_train,
batch_size=BATCH_SIZE, epochs=5,
verbose=VERBOSE, validation_split=VALIDATION_SPLIT,
callbacks=[checkpoint)])
'''
score=model.evaluate(X_test,Y_test,verbose=VERBOSE)
print(score)
'''
程序运行结果：
这是一个在你的实验中需要经常用到的方便的Checkpoint策略。它将确保你的最佳模型被保存，以便稍后使用。它避免了输入代码来手动跟踪，并在训练时序列化最佳模型。
例子No.4:加载之前保存过的Checkpoint模型
Checkpoint只包括模型权重。它假定你了解网络结构。这也可以序列化成JSON或YAML格式。
在下面的示例中，模型结构是已知的，并且最好的权重从先前的实验中加载，然后存储在weights.best.hdf5文件的工作目录中。
那么将该模型用于对整个数据集进行预测。
from __future__ import print_function
import numpy as np
from keras.datasets import mnist
from keras.models import Sequential
from keras.layers.core import Dense, Activation
from keras.optimizers import SGD
from keras.utils import np_utils
from keras.callbacks import ModelCheckpoint
from keras.callbacks import TensorBoard
import matplotlib.pyplot as plt
np.random.seed(1671) # for reproducibility
# network and training
NB_EPOCH = 20
BATCH_SIZE = 128
VERBOSE = 1
NB_CLASSES = 10 # number of outputs = number of digits
OPTIMIZER = SGD() # optimizer, explained later in this chapter
N_HIDDEN = 128
VALIDATION_SPLIT=0.2 # how much TRAIN is reserved for VALIDATION
# data: shuffled and split between train and test sets
(X_train, y_train), (X_test, y_test) = mnist.load_data()
#X_train is 60000 rows of 28x28 values --> reshaped in 60000 x 784
RESHAPED = 784
#
X_train = X_train.reshape(60000, RESHAPED)
X_test = X_test.reshape(10000, RESHAPED)
X_train = X_train.astype('float32')
X_test = X_test.astype('float32')
# normalize
X_train /= 255
X_test /= 255
print(X_train.shape[0], 'train samples')
print(X_test.shape[0], 'test samples')
# convert class vectors to binary class matrices
Y_train = np_utils.to_categorical(y_train, NB_CLASSES)
Y_test = np_utils.to_categorical(y_test, NB_CLASSES)
# M_HIDDEN hidden layers
# 10 outputs
# final stage is softmax
model = Sequential()
model.add(Dense(N_HIDDEN, input_shape=(RESHAPED,)))
model.add(Activation('relu'))
model.add(Dense(N_HIDDEN))
model.add(Activation('relu'))
model.add(Dense(NB_CLASSES))
model.add(Activation('softmax'))
model.summary()
#加载模型
model.load_weights("weight_bst.hdf5")
model.compile(loss='categorical_crossentropy',
optimizer=OPTIMIZER,
metrics=['accuracy'])
score=model.evaluate(X_test,Y_test,verbose=VERBOSE)
print(score)
例子：LearningRateScheduler调整学习率
from keras.callbacks import LearningRateScheduler
def scheduler(epoch):
if epoch % 100 == 0 and epoch != 0:
lr = K.get_value(model.optimizer.lr)
K.set_value(model.optimizer.lr, lr * 0.1)
print("lr changed to {}".format(lr * 0.1))
return K.get_value(model.optimizer.lr)
reduce_lr = LearningRateScheduler(scheduler)
model.fit(train_x, train_y, batch_size=32, epochs=5, callbacks=[reduce_lr])
例子：ReduceLROnPlateau自动减少学习率
from keras.callbacks import ReduceLROnPlateau
reduce_lr = ReduceLROnPlateau(monitor='val_loss', factor=0.2,patience=5, min_lr=0.001)
model.fit(X_train, Y_train, callbacks=[reduce_lr])
例子：CSVLoggerb保存训练结果
csv_logger = CSVLogger('training.log')
model.fit(X_train, Y_train, callbacks=[csv_logger])
例子：轻量级自定义回调函数LambdaCallback
# Print the batch number at the beginning of every batch.
batch_print_callback = LambdaCallback(
on_batch_begin=lambda batch,logs: print(batch))
# Plot the loss after every epoch.
import numpy as np
import matplotlib.pyplot as plt
plot_loss_callback = LambdaCallback(
on_epoch_end=lambda epoch, logs: plt.plot(np.arange(epoch),
logs['loss']))
# Terminate some processes after having finished model training.
processes = ...
cleanup_callback = LambdaCallback(
on_train_end=lambda logs: [
p.terminate() for p in processes if p.is_alive()])
model.fit(...,
callbacks=[batch_print_callback,
plot_loss_callback,
cleanup_callback])
例子：查看TensorBoard监控
from __future__ import print_function
import numpy as np
from keras.datasets import mnist
from keras.models import Sequential
from keras.layers.core import Dense, Activation
from keras.optimizers import SGD
from keras.utils import np_utils
from keras.callbacks import ModelCheckpoint
from keras.callbacks import TensorBoard
import matplotlib.pyplot as plt
np.random.seed(1671) # for reproducibility
# network and training
NB_EPOCH = 20
BATCH_SIZE = 128
VERBOSE = 1
NB_CLASSES = 10 # number of outputs = number of digits
OPTIMIZER = SGD() # optimizer, explained later in this chapter
N_HIDDEN = 128
VALIDATION_SPLIT=0.2 # how much TRAIN is reserved for VALIDATION
# data: shuffled and split between train and test sets
(X_train, y_train), (X_test, y_test) = mnist.load_data()
#X_train is 60000 rows of 28x28 values --> reshaped in 60000 x 784
RESHAPED = 784
#
X_train = X_train.reshape(60000, RESHAPED)
X_test = X_test.reshape(10000, RESHAPED)
X_train = X_train.astype('float32')
X_test = X_test.astype('float32')
# normalize
X_train /= 255
X_test /= 255
print(X_train.shape[0], 'train samples')
print(X_test.shape[0], 'test samples')
# convert class vectors to binary class matrices
Y_train = np_utils.to_categorical(y_train, NB_CLASSES)
Y_test = np_utils.to_categorical(y_test, NB_CLASSES)
# M_HIDDEN hidden layers
# 10 outputs
# final stage is softmax
model = Sequential()
model.add(Dense(N_HIDDEN, input_shape=(RESHAPED,)))
model.add(Activation('relu'))
model.add(Dense(N_HIDDEN))
model.add(Activation('relu'))
model.add(Dense(NB_CLASSES))
model.add(Activation('softmax'))
model.summary()
model.compile(loss='categorical_crossentropy',
optimizer=OPTIMIZER,
metrics=['accuracy'])
history = model.fit(X_train, Y_train,
batch_size=BATCH_SIZE, epochs=5,
verbose=VERBOSE, validation_split=VALIDATION_SPLIT,
callbacks=[TensorBoard(log_dir="./log")])
score=model.evaluate(X_test,Y_test,verbose=VERBOSE)
print(score)
在命令行输入命令：
运行结果：
编写自己的回调函数
我们可以通过继承keras.callbacks.Callback编写自己的回调函数，回调函数通过类成员self.model访问访问，该成员是模型的一个引用。
这里是一个简单的保存每个batch的loss的回调函数
class LossHistory(keras.callbacks.Callback):
def on_train_begin(self, logs={}):
self.losses = []
def on_batch_end(self, batch, logs={}):
self.losses.append(logs.get('loss'))
例子：记录损失函数的历史数据
class LossHistory(keras.callbacks.Callback):
def on_train_begin(self, logs={}):
self.losses = []
def on_batch_end(self, batch, logs={}):
self.losses.append(logs.get('loss'))
model = Sequential()
model.add(Dense(10, input_dim=784, kernel_initializer='uniform'))
model.add(Activation('softmax'))
model.compile(loss='categorical_crossentropy', optimizer='rmsprop')
history = LossHistory()
model.fit(X_train, Y_train, batch_size=128, epochs=20, verbose=0, callbacks=[history])
print history.losses
# outputs
'''
[0.66047596406559383, 0.3547245744908703, ..., 0.25953155204159617, 0.25901699725311789]
参考：https://blog.csdn.net/weixin_38517705/article/details/82971147
扫一扫
258
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
