In this Deep Learning Tutorial we learn how Autoencoders work and how we can implement them in PyTorch.
An autoencoder is not used for supervised learning. We will no longer try to predict something about our input. Instead, an autoencoder is considered a generative model: It learns a distributed representation of our training data, and can even be used to generate new instances of the training data.
An autoencoder model contains two components:
- An encoder that takes an image as input, and outputs a low-dimensional embedding (representation) of the image.
- A decoder that takes the low-dimensional embedding, and reconstructs the image.