search

yinboc / liif

Learning Continuous Image Representation with Local Implicit Image Function, in CVPR 2021 (Oral)
https://yinboc.github.io/liif/
BSD 3-Clause "New" or "Revised" License
1.28k stars 145 forks source link

about the code #43

Open xuejiancai opened 3 years ago

xuejiancai commented 3 years ago

Hello yinbo, I have a question about the meaning of data_norm in train configuration file.I guess inp,gt mean input,ground-truth respectively.But what is the meaning of sub:[0.5] and div:[0.5]? data_norm: inp: {sub: [0.5], div: [0.5]} gt: {sub: [0.5], div: [0.5]} Could you help me?Thank you!

Surayuth commented 3 years ago

Those values are used to normalize the inputs, which are ground truths(gt) and low-resolution images(inp), of the model(feature extracter)​. As you can see, inp: {sub:[0.5], div: [0.5]} means the low-resolution images (which are actually tensors in this case) will be normalized as inp = (inp - sub)/div.