Masked Autoencoders Enable Efficient Knowledge Distillers | Read Paper on Bytez