model.compile(loss=losses.categorical_crossentropy,
optimizer=optimizers.Adadelta(),
metrics=['accuracy'])
print(model.summary())
tensorboard = TensorBoard(log_dir='{0}'.format(log_dir))
# weights.{epoch:02d}-{val_loss:.2f}.hdf5 : a kind of code rule
ckpt_filepath = './{0}'.format(ckpt_dir) + '/weights.{epoch:02d}-{val_loss:.2f}.h5'
checkpointer = ModelCheckpoint(filepath=ckpt_filepath, monitor='val_loss')
history = model.fit(x_train, y_train,
batch_size=batch_size,
epochs=epochs,
verbose=1,
callbacks=[tensorboard, checkpointer],
validation_data=(x_test, y_test))
model_name = '{0}/cifar10_lenet.h5'.format(model_dir)
model.save(model_name)
del model
model = load_model(model_name)
score = model.evaluate(x_test, y_test, verbose=0)
print('Test loss:', score[0])
print('Test accuracy:', score[1])
model = load_model(filename)
initial_epoch = int(filename.split('weights.')[-1].split('-')[0])
print('restore from ckpt, initial_echo = {0}'.format(initial_epoch))
print(model.summary())
tensorboard = TensorBoard(log_dir='./logs_lenet')
# weights.{epoch:02d}-{val_loss:.2f}.hdf5 : a kind of code rule
checkpointer = ModelCheckpoint(
filepath='./ckpt_lenet/weights.{epoch:02d}-{val_loss:.2f}.h5',
monitor='val_loss')
history = model.fit(x_train,
y_train,
batch_size=batch_size,
epochs=epochs,
verbose=1,
callbacks=[tensorboard, checkpointer],
validation_data=(x_test, y_test),
initial_epoch=initial_epoch)
model.save('model/mnist_lenet.h5')
del model
model = load_model('model/mnist_lenet.h5')
score = model.evaluate(x_test, y_test, verbose=0)
print('Test loss:', score[0])
print('Test accuracy:', score[1])
same_face = []
diff_face = []
for i in range(9999):
# 如果两张图片标签相同，则将两个图片作为相同组样本
if labels[i] == labels[i + 1]:
same_face.append([images[i], images[i + 1]])
# 如果两张图片标签不同，则作为差异组样本
else:
diff_face.append([images[i], images[i + 1]])
# 转化为numpy.ndarray，便于传入keras构造的神经网络进行计算
x_train1 = np.array([f[0] for f in same_face + diff_face])
x_train2 = np.array([f[1] for f in same_face + diff_face])
y_train = np.array([1 for i in same_face] + [0 for j in diff_face])
# 如果有模型文件存在，则导入之前的模型参数继续计算
if os.path.exists("H:/face_detection/LeNet.h5"):
LeNet = load_model("H:/face_detection/LeNet.h5")
print("model loaded")
# 使用重构后的generator进行训练
LeNet.fit_generator(generator=get_train_batch([x_train1, x_train2],
y_train,
batch_size=100),
steps_per_epoch=100,
epochs=5)
# 保存模型
LeNet.save("H:/face_detection/face_detection/LeNet.h5")
