xxliang

Results 1 comments of xxliang

> Hi Vivian, > > It seems that you directly resized the original images and labels to 384x384 for network training. > In our experiments, the model was trained on...