reproduction of semantic segmentation using masked autoencoder (mae) - View it on GitHub