Simple Pytorch Implementation of Residual Quantizer and RQ-VAE

RQ-VAE (Residual-Quantized Variational Autoencoder) is a modification of VQ-VAE (Vector Quantized-Variational Autoencoder).

RQ-VAE is proposed in this paper. Its official implementation is found in the link.

Here, we provide a very simple implementation of RQ-VAE and residual quantization therein.

The implmentation here is laregly based on the colab implementation of VQ-VAE.

Requirements

• Pytorch 1.8.2