% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/create_model_lstm_cnn.R
\name{create_model_lstm_cnn_target_middle}
\alias{create_model_lstm_cnn_target_middle}
\title{Create LSTM/CNN network to predict middle part of a sequence}
\usage{
create_model_lstm_cnn_target_middle(
  maxlen = 50,
  dropout_lstm = 0,
  recurrent_dropout_lstm = 0,
  layer_lstm = 128,
  solver = "adam",
  learning_rate = 0.001,
  vocabulary_size = 4,
  bidirectional = FALSE,
  stateful = FALSE,
  batch_size = NULL,
  padding = "same",
  compile = TRUE,
  layer_dense = NULL,
  kernel_size = NULL,
  filters = NULL,
  pool_size = NULL,
  strides = NULL,
  label_input = NULL,
  zero_mask = FALSE,
  label_smoothing = 0,
  label_noise_matrix = NULL,
  last_layer_activation = "softmax",
  loss_fn = "categorical_crossentropy",
  num_output_layers = 1,
  f1_metric = FALSE,
  auc_metric = FALSE,
  bal_acc = FALSE,
  verbose = TRUE,
  batch_norm_momentum = 0.99,
  model_seed = NULL,
  mixed_precision = FALSE,
  mirrored_strategy = NULL
)
}
\arguments{
\item{maxlen}{Length of predictor sequence.}

\item{dropout_lstm}{Fraction of the units to drop for inputs.}

\item{recurrent_dropout_lstm}{Fraction of the units to drop for recurrent state.}

\item{layer_lstm}{Number of cells per network layer. Can be a scalar or vector.}

\item{solver}{Optimization method, options are \verb{"adam", "adagrad", "rmsprop"} or \code{"sgd"}.}

\item{learning_rate}{Learning rate for optimizer.}

\item{vocabulary_size}{Number of unique character in vocabulary.}

\item{bidirectional}{Use bidirectional wrapper for lstm layers.}

\item{stateful}{Boolean. Whether to use stateful LSTM layer.}

\item{batch_size}{Number of samples that are used for one network update. Only used if \code{stateful = TRUE}.}

\item{padding}{Padding of CNN layers, e.g. \verb{"same", "valid"} or \code{"causal"}.}

\item{compile}{Whether to compile the model.}

\item{layer_dense}{Vector specifying number of neurons per dense layer after last LSTM or CNN layer (if no LSTM used).}

\item{kernel_size}{Size of 1d convolutional layers. For multiple layers, assign a vector. (e.g, \code{rep(3,2)} for two layers and kernel size 3)}

\item{filters}{Number of filters. For multiple layers, assign a vector.}

\item{pool_size}{Integer, size of the max pooling windows. For multiple layers, assign a vector.}

\item{strides}{Stride values. For multiple layers, assign a vector.}

\item{label_input}{Integer or \code{NULL}. If not \code{NULL}, adds additional input layer of \code{label_input} size.}

\item{zero_mask}{Boolean, whether to apply zero masking before LSTM layer. Only used if model does not use any CNN layers.}

\item{label_smoothing}{Float in [0, 1]. If 0, no smoothing is applied. If > 0, loss between the predicted
labels and a smoothed version of the true labels, where the smoothing squeezes the labels towards 0.5.
The closer the argument is to 1 the more the labels get smoothed.}

\item{label_noise_matrix}{Matrix of label noises. Every row stands for one class and columns for percentage of labels in that class.
If first label contains 5 percent wrong labels and second label no noise, then

\code{label_noise_matrix <- matrix(c(0.95, 0.05, 0, 1), nrow = 2, byrow = TRUE )}}

\item{last_layer_activation}{Activation function of output layer(s). For example \code{"sigmoid"} or \code{"softmax"}.}

\item{loss_fn}{Either \code{"categorical_crossentropy"} or \code{"binary_crossentropy"}. If \code{label_noise_matrix} given, will use custom \code{"noisy_loss"}.}

\item{num_output_layers}{Number of output layers.}

\item{f1_metric}{Whether to add F1 metric.}

\item{auc_metric}{Whether to add AUC metric.}

\item{bal_acc}{Whether to add balanced accuracy.}

\item{verbose}{Boolean.}

\item{batch_norm_momentum}{Momentum for the moving mean and the moving variance.}

\item{model_seed}{Set seed for model parameters in tensorflow if not \code{NULL}.}

\item{mixed_precision}{Whether to use mixed precision (https://www.tensorflow.org/guide/mixed_precision).}

\item{mirrored_strategy}{Whether to use distributed mirrored strategy. If NULL, will use distributed mirrored strategy only if >1 GPU available.}
}
\value{
A keras model with two input layers. Consists of LSTN, CNN and dense layers.
}
\description{
Creates a network consisting of an arbitrary number of CNN, LSTM and dense layers.
Function creates two sub networks consisting each of (optional) CNN layers followed by an arbitrary number of LSTM layers. Afterwards the last LSTM layers
get concatenated and followed by one or more dense layers. Last layer is a dense layer.
Network tries to predict target in the middle of a sequence. If input is AACCTAAGG, input tensors should correspond to x1 = AACC, x2 = GGAA and y = T.
}
\examples{
\dontshow{if (reticulate::py_module_available("tensorflow")) (if (getRversion() >= "3.4") withAutoprint else force)(\{ # examplesIf}
create_model_lstm_cnn_target_middle(
  maxlen = 500,
  vocabulary_size = 4,
  kernel_size = c(8, 8, 8),
  filters = c(16, 32, 64),
  pool_size = c(3, 3, 3),
  layer_lstm = c(32, 64),
  layer_dense = c(128, 4),
  learning_rate = 0.001)
 
\dontshow{\}) # examplesIf}
}