poli.objective_repository.gfp_cbas.make_vae#

Functions

build_vae(latent_dim[, n_tokens, ...])

Returns a compiled VAE model

identity_loss(y_true, y_pred)

Returns the predictions

summed_categorical_crossentropy(y_true, y_pred)

Negative log likelihood of categorical distribution

Classes

BaseVAE(input_shape, latent_dim, *args, **kwargs)

Base class for Variational Autoencoders implemented in Keras

SimpleVAE(input_shape, latent_dim[, flatten])

Basic VAE where the encoder and decoder can be constructed from lists of layers