This is the supplementary materials for the submission number 3765 ("Learning Representation in Colour Conversion") to ICLR-2020.



The content of this folder: 

 - "lesion"
   This directory contains the linear transformations we computed for the lesion study.

 - "models"
   This directory contains all the trained models.

 - "notebooks"
   This directory contains six Jupiter-Notebooks to reproduce the reported results.
   Interested readers can obtain results for other images in these files.

 - "results"
   This directory contains the quantitative results for the three datasets we investigated (ImageNet, COCO, and CelebA). 

 - "src"
   The source code to train and test VQ-VAEs explored in this article.



Each subdirectory has its own readme file to explain its content in greater detail.
