Semantic Segmentation with Pytorch
image2vector: The first 47 layers of the resnet50 model of the deep learning neural network are used to extract the vector of the picture, and the output dimension is 512
