Extract Keras models from an autoencoder wrapper

# S3 method for ruta_autoencoder
to_keras(learner, encoder_end = "encoding",
  decoder_start = "encoding", weights_file = NULL)

# S3 method for ruta_autoencoder_variational
to_keras(learner, ...)

Arguments

learner

Object of class "ruta_autoencoder". Needs to have a member input_shape indicating the number of attributes of the input data

encoder_end

Name of the Keras layer where the encoder ends

decoder_start

Name of the Keras layer where the decoder starts

weights_file

The name of a hdf5 weights file in order to load from a trained model

...

Additional parameters for to_keras.ruta_autoencoder

Value

A list with several Keras models:

  • autoencoder: model from the input layer to the output layer

  • encoder: model from the input layer to the encoding layer

  • decoder: model from the encoding layer to the output layer

See also