This repository contains the implementation of deep learning models working as autoencoders.
The code is provided in iPython Jupyter Notebook for visuality and simplicity. As first insights the datasets used are the MNIST (handrwitten digit) dataset and Fashion-MNIST dataset since they are easy to handle and greyscale images. Currently this repository contains the implementation of:
- Standard Neural Network Autoencoder with 1 Hidden Layer Neural Network
- Convolutional Neural Network Autoencoder with 3 Conv Layers in the encoder and 3 Conv Layers in the decoder
- Denoising Convolutional Neural Network Autoencoder with 3 Conv Layers in the encoder and 3 Conv Layers in the decoder