• 完成注解
• 优化训练代码
• 完成推理代码

This is a first implementation of a Colorization Diffusion Based Method

Modify the conf.yml file, set the 'mode' option to 1. Then run the main.py file specifying the path to the config file (absolute or relative) Example : python main.py --config conf.yml

I have impossibility to train and to test the model implemented due to my lack of computational power. There might be some mistakes in the code, any insight and remark is welcomed

For the validation loop in the training loop, necessity to use/find a more suitable/ an additional metric

Palette Image_to_Image Diffusion Models https://arxiv.org/pdf/2111.05826v1.pdf

Diffusion Models Beat GANs on Image Synthesis https://arxiv.org/pdf/2105.05233.pdf

The Unet Network script directly comes from the repo of this last : https://github.com/openai/guided-diffusion (with small modifications according to the Palette paper)

A colorization Dataset : https://www.kaggle.com/shravankumar9892/image-colorization ( Palette paper's researchers uses ImageNet )