如何在 tensorflow 中编写自定义的 LSTM?

How to write a customized LSTM in tensorflow?

我正在尝试重新实现这篇论文 Semantically Conditioned LSTM-based Natural Language Generation for Spoken Dialogue Systems,他们在其中向 LSTM 单元添加了一个门并更改了状态的计算方式。

我如何在 tensorflow 中执行此操作?我需要添加新的 OP 吗?

tf.nn.rnn() and tf.nn.dynamic_rnn() functions accept an argument cell of type tf.nn.rnn_cell.RNNCell. For example you can take a look at the implementation of tf.nn.rnn_cell.BasicLSTMCell (in particular the BasicLSTMCell.__call__() method),这可能是您自定义 LSTM 的良好起点。