Closed yiyayada closed 3 years ago
Thanks for your issue. This research is not my article. I just implement this research by pytorch. The image size of this result is 256, 256. Please refer to https://arxiv.org/abs/2005.05207. If you want to train a model you must change model architecture. this code is only for 256256 haha. I am not an author of this research.
Hello, I want to quote your article training comparison experiment, can your article train 128128 pixel data?when I train my 128128 data, the model have wrong,think there is no need to change the model?THANK YOU VERY MUCH~~