AntixK/PyTorch-VAE Alternatives

Choose it when you want a broader generative-modelling library that also includes VAEs alongside flows/autoregressive models in one API.

Easy generative modeling in PyTorch

Use it if you want a single unified VAE benchmarking API rather than seed's per-model scripts, though it is stale.

Unifying Variational Autoencoder (VAE) implementations in Pytorch (NeurIPS 2022)

Use it if you want VAE training combined with disentanglement evaluation, accepting older research code.

Experiments for understanding disentanglement in VAE latent representations

Use it for a standalone categorical Gumbel-Softmax VAE demo in a notebook instead of the full suite.

categorical variational autoencoder using the Gumbel-Softmax estimator

Use it when you only need a basic PyTorch VQ-VAE implementation, not the whole collection; it is unmaintained.

A pytorch implementation of the vector quantized variational autoencoder (https://arxiv.org/abs/1711.00937)

Use it for another standalone VQ-VAE example in Python; no license and unmaintained.

Vector Quantized VAEs - PyTorch Implementation

Use it for a bare PyTorch beta-VAE implementation without the extra models in seed; no recent maintenance.

Pytorch implementation of β-VAE

Use it if you specifically need the original Caffe teaching code for a vanilla VAE instead of a modern PyTorch suite.

Caffe code to accompany my Tutorial on Variational Autoencoders

Use it when all you need is a single simple PyTorch VAE implementation rather than multiple variants.

A Variational Autoencoder (VAE) implemented in PyTorch

Use it when you want a minimal CNN-VAE demo in a notebook instead of the collection; Jupyter-only and unmaintained.

A CNN Variational Autoencoder (CNN-VAE) implemented in PyTorch

Use it for a single beta-TCVAE implementation and its disentanglement experiments, not the varied VAE suite.

code for "Isolating Sources of Disentanglement in Variational Autoencoders".

Use it only if you are on a legacy TensorFlow 1 stack and need a simple MNIST VAE.

Tensorflow implementation of variational auto-encoder for MNIST

Use it when you want a compact, current PyTorch VAE/CVAE example on MNIST instead of the CelebA multi-model codebase.

Variational Autoencoder and Conditional Variational Autoencoder on MNIST in PyTorch

Use it if you want older TensorFlow code with VAE plus GAN/DRAW examples rather than a VAE-only PyTorch collection.

A collection of generative methods implemented with TensorFlow (Deep Convolutional Generative Adversarial Networks (DCGAN), Variational Autoencoder (VAE) and DRAW: A Recurrent Neural Network For Image Generation).

Use it for a clean single VAE/IAF implementation in PyTorch/TensorFlow instead of the many-model seed.

Variational autoencoder implemented in tensorflow and pytorch (including inverse autoregressive flow)

Use it if you work in legacy Torch/Lua and want autoencoder implementations that include VAEs.

Torch implementations of various types of autoencoders