yuanming-hu / exposure

Learning infinite-resolution image processing with GAN and RL from unpaired image datasets, using a differentiable photo editing model.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

How to generate the training data from the very begining

vissac opened this issue · comments

Hi yuanming, I'm try to generate the training data from fivek origin data and I'm come into some problems, you said
In the Collections list, select collection Inputs/Input with Daylight WhiteBalance minus 1.5.
but I'm confused at minus 1.5, what does minus 1.5 mean? Minus tone? tint? or exposure?

Thanks for the question. I think that's exposure, as they appear darker. I'm just following some previous works to select this collection.

Thanks for the question. I think that's exposure, as they appear darker. I'm just following some previous works to select this collection.

Got it~ Thank you. And I also download the data after your process, but I cannot reproduce the data in image_raw.npy to its origin dng. How can I connect the data in npy with origin dng. And in meta_raw.pkl I only find the name of the files, is it necessary for recover the origin data from npy?

Sorry for the delayed reply. The image_raw.py file is the pre-processed dataset for fast data loading during training. It is not necessary to recover dngs from that file.