The `tensorflow.contrib.rnn.MultiRNNCell.zero_state` function in Python's Tensorflow library returns the initial state of a multi-layer RNN cell. This function initializes the hidden state and cell state of each layer of the RNN cell to zero. It takes the number of layers and the data type as input parameters and returns the initial state with the specified shape and data type. This function is commonly used for initializing the state of recurrent neural networks before processing a sequence of data.
Python MultiRNNCell.zero_state - 33 examples found. These are the top rated real world Python examples of tensorflow.contrib.rnn.MultiRNNCell.zero_state extracted from open source projects. You can rate examples to help us improve the quality of examples.