yangyiliu21 / ntire2023_itbdehaze Goto Github PK
View Code? Open in Web Editor NEWOfficial PyTorch implementation of the Winner Award solution of NTIRE 2023 NonHomogeneous Dehazing Challenge (CVPR Workshop 2023).
Official PyTorch implementation of the Winner Award solution of NTIRE 2023 NonHomogeneous Dehazing Challenge (CVPR Workshop 2023).
I want to train only on the 2020 data on a device with 24GB GPU MEM。Is there a recommended configuration?
Hello,I want to konw how to compute the gamma values such as '# for 2020 GT images, gamma values should be: R(1.07), G(1.17), B(1.05)'?
Hi,
I’m trying to train the model on NH-Haze20 dataset with the given code, and I had followed the settings showed in the paper which is a little bit different from train.py.
However, the testing result (PSNR=20.36, SSIM=0.697) is not as good as the performance shows in the paper (PSNR=21.44, SSIM=0.71).
And the following are the settings I’ve used, including splitting the dataset into 45 training data, 5 validation data and 5 testing data, setting training epoch=6500, batch size=8, randomly crop the images to 256*256, the learning rate of G is setting to 1e−4 initially and decreased to 1e−6 with a cosine annealing strategy, and doing the same data augmentation strategies showed in section4.2.
I’d like to confirm whether there is any setting that I misunderstood, leading to the difference between our results?
Thank you so much!!
Hi, I'm a little confused about the Swin Transformer V2 part. How did you get the ImageNet pretrained SwinTransformerV2 weights? Is it downloaded from elsewhere or trained by yourself? Moreover, I am expecting to know the specific training details and how you determined the number of Swin Transformer blocks? Thank you! Best wishes.
The problem of pretrained model which is named 'best.pkl' can not be correctly put into the model MyEnsembleNet. The problem is induced because of the incorrcet key and value. How could you test or train the model with the problem? Please offer a sound and complete solution as good as you can. Much thanks!
Hello, thank you for your outstanding work, could you please share the code about the part of your work that uses gamma correction to enlarge the dataset? At the same time, if you have NH-HAZE2020, NH-HAZE2021 and NH-HAZE2023 original datasets, is it convenient to share? I would appreciate it so much!
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.