This project started out as a simple test to implement Variational Autoencoders using PyTorch Lightning but evolved into a 4 part blog/tutorial on TowardsDataScience. Check them out here:
Part 1: Mathematical Foundations and Implementation
Part 2: Supercharge with PyTorch Lightning
Part 3: Convolutional VAE, Inheritance and Unit Testing
Part 4: Deployment
Hosted on Heroku
https://vae-playground.herokuapp.com/
You can train the VAE models by using train.py
and editing the config.yaml
file.
Hyperparameters to change are:
- model_type
- alpha
- hidden_dim
- dataset
docker pull reoneo/vae-playground
docker run -p 8501:8501 -e PORT=8501 reoneo/vae-playground
docker build . -t <tag>
docker run -p 8501:8501 -e PORT=8501 <tag>
This will host the app on localhost:8501