gluonts.mx.block.rnn module

class gluonts.mx.block.rnn.RNN(mode: str, num_hidden: int, num_layers: int, bidirectional: bool = False, **kwargs)[source]

Bases: mxnet.gluon.block.HybridBlock

Defines an RNN block.

Parameters

  • mode – type of the RNN. Can be either: rnn_relu (RNN with relu activation), rnn_tanh, (RNN with tanh activation), lstm or gru.

  • num_hidden – number of units per hidden layer.

  • num_layers – number of hidden layers.

  • bidirectional – toggle use of bi-directional RNN as encoder.

hybrid_forward(F, inputs: Union[mxnet.ndarray.ndarray.NDArray, mxnet.symbol.symbol.Symbol]) → Union[mxnet.ndarray.ndarray.NDArray, mxnet.symbol.symbol.Symbol][source]
Parameters

  • F – A module that can either refer to the Symbol API or the NDArray API in MXNet.

  • inputs – input tensor with shape (batch_size, num_timesteps, num_dimensions)

Returns

rnn output with shape (batch_size, num_timesteps, num_dimensions)

Return type

Tensor