keras_bi_gru: keras_bi_gru
In systats/deeplyr: Pretrained keras models for predicting ideology from tweets.

Word embedding + spatial dropout + (pooled) gated recurrent unit

1
2
3

keras_bi_gru(input_dim, embed_dim = 128, seq_len = 50, gru_dim = 64,
  gru_drop = 0.2, bidirectional = F, output_fun = "sigmoid",
  output_dim = 1)

`input_dim`	Number of unique vocabluary/tokens
`embed_dim`	Number of word vectors
`seq_len`	Length of the input sequences
`gru_dim`	Number of recurrent neurons (default 64)
`gru_drop`	Recurrent dropout ratio
`output_fun`	Output activation function
`output_dim`	Number of neurons of the output layer