Dear author, is the resolutionof Fuji dataset is 4032 6032 instead of 4000 6000 as you mentioned in your paper? So, how can I pack it into 9 channels when the size can not be devided by 3? When I use your code '_pack_raw()', I got the shape (b,9,1344, 2010), then how can I make it to the original shape (3, 4032, 6032)? I'm confused since 2010 can not be devided by 6032, generally , the output size should be same as or double to the input size, right?
Thank you!
Dear author, is the resolutionof Fuji dataset is 4032 6032 instead of 4000 6000 as you mentioned in your paper? So, how can I pack it into 9 channels when the size can not be devided by 3? When I use your code '_pack_raw()', I got the shape (b,9,1344, 2010), then how can I make it to the original shape (3, 4032, 6032)? I'm confused since 2010 can not be devided by 6032, generally , the output size should be same as or double to the input size, right? Thank you!