The implementation of [Parsing Very High Resolution Urban Scene Images by Learning Deep ConvNets with Edge-Aware Loss] Pretrained model will be updated soon.