Scaling and bias codes for modeling speaker-adaptive DNN-based speech synthesis systems | Read Paper on Bytez