python - Keras 无法使用回调来存储检查点

Question

train_data = ImageDataGenerator(rescale=1./255)
test_data = ImageDataGenerator(rescale=1./255)
train_set = train_data.flow_from_directory('train_set',
                                           target_size=(28,28),
                                           batch_size=16,
                                           class_mode='categorical')
test_set = train_data.flow_from_directory('test_set',
                                           target_size=(28,28),
                                           batch_size = 16,
                                           class_mode='categorical')
checkpointer = ModelCheckpoint(filepath="best_weight.hdf5",
                               monitor='val_acc',
                               verbose=1,
                               save_best_only=True)
history = classifier.fit_generator(train_set,
                                   steps_per_epoch=210,
                                   epochs=5,
                                   callbacks=[checkpointer],
                                   validation_data=test_set,
                                   validation_steps=90)

classifier.load_weights('best_weights.hdf5')

当我尝试加载最佳权重时，它显示没有这样的文件或目录。我应该如何做出改变？非常感谢你们!

Answer 1

您的文件名有拼写错误；您保存在

filepath="best_weight.hdf5"

当您尝试加载'best_weights.hdf5'时。

删除文件名中最后的s，即:

classifier.load_weights('best_weight.hdf5')