Junlin Han

> > > > Hello, Yes, the result can be bad when performing underwater image enhancement tasks. Please try this CWR pre-trained on HICRD, fine-tuned on UIEB: https://drive.google.com/file/d/1MCYZcAjKAf3y_rSNes_Fs3tm7L0U_77e/view?usp=sharing > >...

> > > > > > Hello, Yes, the result can be bad when performing underwater image enhancement tasks. Please try this CWR pre-trained on HICRD, fine-tuned on UIEB: https://drive.google.com/file/d/1MCYZcAjKAf3y_rSNes_Fs3tm7L0U_77e/view?usp=sharing...

Hi, More instructions here: Download link: https://data.csiro.au/collections/collection/CIcsiro:49488 To download the dataset, you need to: 1: Click the download link. 2: Click Download, select Download all files via WebDAV in select...

Hello ddeunga, Evaluation codes are not included in this repo, you might check this [link](https://github.com/JunlinHan/BID/tree/main/metrics) for Matlab PSNR/SSIM .

> Thank you so much for your care. But I can't use MATLAB, so could you tell me the Python code? Check [skimage](https://scikit-image.org/docs/stable/api/skimage.metrics.html), btw results from skimage might be different...

您好，我们没有测试更大的，用1主要是follow了之前这种gan网络的设置。

Yes, it is generated on-the-fly. When it trains for longer epochs (and sees lots of samples) the performance should be similar.

Hello @khawar-islam , Unstable results between multiple runs are quite common. Our reported results in the paper should be the mean of 3 individual trials (for CIFAR datasets). We also...

Hello, Try adding `--preprocess none`!

> I have another problem. How to generate images using only the G_A generator during the testing phase. Thanks! You have to change some code in ./models/dcl_model.py, a simple solution...