Training an autoencoder via conjugate gradients with a L2 reconstruction

$39

Description

Test script for training an autoencoder as if it is sparse coding with no latent variables
via conjugate gradients with a L2 reconstruction
error term and abs(tanh) nonlinearity for regulatization only. Cost is thus:
C = 1/2*||W^T*W*x – x|| + lambda*|tanh(W*x)|

Denoising using autoencoders in TensorFlow

Autoencoders of raw images in TensorFlow

 

 

Reviews

There are no reviews yet.

Be the first to review “Training an autoencoder via conjugate gradients with a L2 reconstruction”
SKU: P2018F222_neuralNets2 Category: