https://www.zhihu.com/question/66283266/answer/240344515網絡
第一句ide
We adapted the VGG-16 network (Simonyan & Zisserman, 2015) for dense prediction and removed the last two pooling and striding layers.
第二句ci
Specifically, each of these pooling and striding layers was removed and convolutions in all subsequent layers were dilated by a factor of 2 for each pooling layer that was ablated.
第三句rem
Most significantly, we remove the last two pooling and striding layers entirely, whereas Long et al. kept them and Chen et al. replaced striding by dilation but kept the pooling layers.