Skip to main content
EURASIP Journal on Audio, Speech, and Music Processing
Fig. 1 | EURASIP Journal on Audio, Speech, and Music Processing

Fig. 1

From: Wise teachers train better DNN acoustic models

Fig. 1

Pair of teacher and student networks. Predictions from the teacher network are used as training targets for an easier-to-evaluate student network using a large amount of unlabeled data. Teacher and student networks may have different input representations, sizes, and architectures

Back to article page