output is same with input

[noseDewdrop]

your code is helpful, but i met some problems, when i training, i found that the mask only having 0 and 1 can not cover the detail image, so my training results make that input images are same with output images, i see that in tensorboard, and when i use the result model to test, the test results is same of that, i am not understanding about that, can you give me some explain?