wujiekd opened this issue 2 years ago · 1 comments
Hello, during the test, we will input GT, and then generate images to compare with GT?
That's correct. You evaluate the model at a reconstruction task.