Unifying Variational Autoencoder (VAE) implementations in Pytorch (NeurIPS 2022)
-
Updated
Jul 31, 2024 - Python
Unifying Variational Autoencoder (VAE) implementations in Pytorch (NeurIPS 2022)
Reimplementation of Wasserstein Auto Encoder (WAE) with Wasserstein GAN based penalty D_Z in Tensorflow
Music generation with Wasserstein Autoencoders
This research project intent is to review and demonstrate a comparability among recent auto-encoder methods by utilizing single architecture and resolution. Each method will be ranked based on selective performance measure in modeling healthy brain and the sensitivity towards domain shift.
Wasserstein Autoencoders implementation in Tensorflow 2
This is the code implementation of our research paper in AICAI '19 Dubai.
Add a description, image, and links to the wasserstein-autoencoder topic page so that developers can more easily learn about it.
To associate your repository with the wasserstein-autoencoder topic, visit your repo's landing page and select "manage topics."