Found 1 document(s)
Page 1 of 1
-
Self-supervised speech models commonly consist of a set of convolutional neural network (CNN) layers used as a local feature extractor, followed by transformer layers. The CNN layers are computationally demanding, and ar...