keras_cudnn_cnn_lstm: keras_cudnn_cnn_lstm

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

Description

Word embedding + 1D pooled convolution + lstm layer + GPU

Usage

keras_cudnn_cnn_lstm(input_dim, embed_dim = 128, seq_len = 50,
  filter_size = 5, n_filters = 100, pool_size = 4, lstm_dim = 64,
  bidirectional = F, dropout = 0.2, output_dim = 1,
  output_fun = "sigmoid")

Arguments

input_dim	Number of unique vocabluary/tokens
embed_dim	Number of word vectors
seq_len	Length of the input sequences
filter_size	the window size (kernel_size)
n_filters	the number of convolutional filters
pool_size	pooling dimension (filters)
lstm_dim	Number of lstm neurons (default 32)
bidirectional	default is F
output_dim	Number of neurons of the output layer
output_fun	Output activation function
lstm_drop	default is 2

Value

keras model

systats/deeplyr documentation built on Oct. 4, 2020, 7:59 p.m.