Cannot reproduce results for the Chikusei dataset

[mtntruong]

Greetings, thank you for sharing your wonderful work. I have re-run the codes and successfully reproduced the results for the Pavia and Botswana datasets. Unfortunately, for the Chikusei dataset, I can only achieve ~35dB instead of 43dB as reported in the paper. I would be very grateful if you can verify the settings in the config files as they may effect the performance. It's worth noting that, for the Pavia and Botswana datasets, I used a single GPU for training. However, for the Chikusei dataset, I used two GPUs as the images are bigger, they cannot fit in a single GPU.
Thank you very much.

[Nileshmaharjan]

@wgcban

Hi there! I tried to replicate the results for the Chikusei dataset and also could only achieve a PSNR of around 35 dB instead of the 43 dB reported in the paper. It would be helpful if you could share the config files as this might impact the performance. In my case, I was only using one GPU.

Thank you very much.