如何在Keras中设置自适应学习率 [英] How to Setup Adaptive Learning Rate in Keras

问题描述

下面是我的代码:

model = Sequential([
    Dense(32, input_shape=(32,), activation = 'relu'),
    Dense(100, activation='relu'),
    Dense(65, input_shape=(65,), activation='softmax')
])

model.summary()
model.compile(SGD(lr=.1), loss='binary_crossentropy', metrics=['accuracy'])
model.fit(train_samples, train_labels, batch_size=1000, epochs=1000,shuffle = True, verbose=2)

如何设置模型的自适应学习率?

How will I set an adaptive learning rate of my model?

推荐答案

您可以使用解决方法.

对于范围(0，MaxEpoch)中的each_iteration:

For each_iteration in range(0, MaxEpoch):

1. 指定您自己的学习率函数，该函数针对每个时期输出学习率lr.然后将lr传递给your_optimiser

1. Specify your own learning rate function that outputs a learning rate lr with respect to per epoch. The lr is then passed to your_optimiser

运行model.compile(... optimizer = your_optimiser ...)

run model.compile(...optimizer=your_optimiser...)

运行model.fit(... epochs = 1 ...)

run model.fit(...epochs = 1...)

第一个纪元后，使用model.save_weights(...)

After the ONE epoch, use model.save_weights(...)

通过model.load_weights(...)加载权重以进行下一次迭代.详情请参见此处 https://keras .io/getting-started/faq/#how-can-i-save-a-keras-model

Load weights by model.load_weights(...) for next iteration. See here for details https://keras.io/getting-started/faq/#how-can-i-save-a-keras-model

实际上，#4和#5使您能够进行迁移学习

In fact, #4 and #5 enables you to do a transfer learning

这篇关于如何在Keras中设置自适应学习率的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！