问题描述
我想将层归一化应用到使用 tf.keras 的循环神经网络.在 TensorFlow 2.0 中，tf.layers.experimental 中有一个 LayerNormalization 类，但不清楚如何在 之类的循环层中使用它LSTM，在每个时间步(因为它被设计用于使用).我应该创建一个自定义单元格，还是有更简单的方法?
I would like to apply layer normalization to a recurrent neural network using tf.keras. In TensorFlow 2.0, there is a LayerNormalization class in tf.layers.experimental, but it's unclear how to use it within a recurrent layer like LSTM, at each time step (as it was designed to be used). Should I create a custom cell, or is there a simpler way?
例如，在创建LSTM层时，在每个时间步应用dropout就像设置recurrent_dropout参数一样简单，但没有recurrent_layer_normalization代码> 参数.
For example, applying dropout at each time step is as easy as setting the recurrent_dropout argument when creating an LSTM layer, but there is no recurrent_layer_normalization argument.
推荐答案
在 tensorflow 插件中，有一个开箱即用的预构建 LayerNormLSTMCell.
In tensorflow addons, there's a pre-built LayerNormLSTMCell out of the box.
有关详细信息，请参阅此文档.您可能必须先安装 tensorflow-addons，然后才能导入此单元格.
See this doc for more details. You may have to install tensorflow-addons before you can import this cell.
pip install tensorflow-addons
这篇关于如何使用 tf.keras 在 RNN 中应用层归一化?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！