tensorflow.contrib.rnn

The `tensorflow.contrib.rnn` module contains implementations of various recurrent neural network (RNN) cells and functions that can be used for building RNNs in TensorFlow. Some of the commonly used classes in this module are:

1. `BasicRNNCell`: Implements a basic RNN cell with a single hidden layer.
2. `LSTMCell`: Implements a Long Short-Term Memory (LSTM) cell.
3. `GRUCell`: Implements a Gated Recurrent Unit (GRU) cell.
4. `MultiRNNCell`: Implements a multi-layer RNN cell by stacking multiple RNN cells together.
5. `static_rnn`: Builds an RNN from a sequence of inputs using a specified RNN cell.

These classes and functions can be used to build various types of RNNs, such as language models, sequence-to-sequence models, and time series models.