backprop_rnn

backpropagate the error in a model object of type rnn

Implementation of a Recurrent Neural Network architectures in native R, including Long Short-Term Memory (Hochreiter and Schmidhuber, <doi:10.1162/neco.1997.9.8.1735>), Gated Recurrent Unit (Chung et al., <arXiv:1412.3555>) and vanilla RNN.

Bastiaan Quast

Recurrent Neural Network

backprop_rnn function

<dl><dt>model</dt>
<dd>the output model object</dd>
<dt>a</dt>
<dd>the input of this learning batch</dd>
<dt>c</dt>
<dd>the output of this learning batch</dd>
<dt>j</dt>
<dd>the indexes of the sample in the current batch</dd>
<dt>...</dt>
<dd>argument to be passed to method</dd></dl>

Arguments

backprop_rnn — backprop_rnn

<dl>

<dt>model</dt>
<dd>the output model object</dd>


<dt>a</dt>
<dd>the input of this learning batch</dd>


<dt>c</dt>
<dd>the output of this learning batch</dd>


<dt>j</dt>
<dd>the indexes of the sample in the current batch</dd>


<dt>...</dt>
<dd>argument to be passed to method</dd>

</dl>

backprop_rnn: backprop_rnn

Description

Usage

Value

Arguments