Home

Awesome

VQ-VAE (Neural Discrete Representation Learning) Tensorflow

Intro

This repository implements the paper, Neural Discrete Representation Learning (VQ-VAE) in Tensorflow.

:warning: This is not an official implementation, and might have some glitch (,or a major defect).

Requirements

Updated Result: ImageNet

Updated Result: Sampling with PixelCNN

Results

All training is done with Quadro M4000 GPU. Training MNIST only takes less than 10 minutes.

Training

It will download required datasets on the directory ./datasets/{mnist,cifar10} by itself. Hence, just run the code will do the trick.

Run train

Change the hyperparameters accordingly as you want. Please check at the bottom of each script.

Evaluation

I provide the model and the code for generating (,or reconstructing) images in the form of Jupyter notebook. Run jupyter notebook server, then run it to see more results with provided models.

If you want to test NLL, then run test() function on cifar.py by uncomment the line. You can find it at the bottom of the file.

TODO

Contributions are welcome!

Thoughts and Help request

Acknowledgement