Woncheol Shin¹, Gyubok Lee¹, Jiyoung Lee¹, Joonseok Lee^2,3, Edward Choi¹ | Paper

¹KAIST, ²Google Research, ³Seoul National University

Recently, vector-quantized image modeling has demonstrated impressive performance on generation tasks such as text-to-image generation. However, we discover that the current image quantizers do not satisfy translation equivariance in the quantized space due to aliasing, degrading performance in the downstream text-to-image generation and image-to-text generation, even in simple experimental setups. Instead of focusing on anti-aliasing, we take a direct approach to encourage translation equivariance in the quantized space. In particular, we explore a desirable property of image quantizers, called 'Translation Equivariance in the Quantized Space' and propose a simple but effective way to achieve translation equivariance by regularizing orthogonality in the codebook embedding vectors. Using this method, we improve accuracy by +22% in text-to-image generation and +26% in image-to-text generation, outperforming the VQGAN.

conda env create -f environment.yaml
conda activate te

bash download_mnist64x64.sh

python main.py --base configs/mnist64x64_vqgan.yaml -t True --gpus 0,1 --max_epochs 40 --seed 23

To use TensorBoard,

run:

tensorboard --logdir logs --port [your_number] --bind_all

And then open your browser and go to http://localhost:[your_number]/.

Please refer to Bi-directional DALL-E.

@misc{shin2021translationequivariant,
      title={Translation-equivariant Image Quantizer for Bi-directional Image-Text Generation}, 
      author={Woncheol Shin and Gyubok Lee and Jiyoung Lee and Joonseok Lee and Edward Choi},
      year={2021},
      eprint={2112.00384},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}

The implementation of 'TE-VQGAN' and 'Bi-directional Image-Text Generator' is based on VQGAN and DALLE-pytorch.