python - TensorFlow: simple recurrent neural network -

i've built neural networks tensorflow, basic mlps , convolutional neural networks. want move on recurrent neural networks. however, i'm not experienced in natural language processing. therefore tensorflow nlp tutorials rnns not easy read me (and not interesting, too).

basically want start off simple, not lstm.

how 1 build simple recurrent neural network, elman network, in tensorflow?

i able find gru- or lstm rnn examples tensorflow, nlp. know of simple recurrent neural network tutorials or examples tensorflow?

this figure shows basic elman network, called srn (simple recurrent network):

one option use built-in rnncell located in tensorflow/python/ops/rnn_cell.py.

if don't want can make own rnn. rnn train using back-propagation through time. try unrolling network fixed number of steps, e.g. consider input sequences of length ten. can write loop in python of matrix multiplications each step of network. each time can take output previous step , concatenate input step. not many lines of code working.