This is a lightweight (200 loc) implementation of the VQ-VAE Neural Discrete representation learning.
By default it trains on cifar10
python vq-vae-img.py
Edit hyperparams, paths in source code to train on ImageNet
MIT
| Name | Name | Last commit date | ||
|---|---|---|---|---|
This is a lightweight (200 loc) implementation of the VQ-VAE Neural Discrete representation learning.
By default it trains on cifar10
python vq-vae-img.py
Edit hyperparams, paths in source code to train on ImageNet
MIT