EndtoEnd Speech and Language Processing
端到端的语音处理系统，Recently,encoder-decoderneuralnetworks
haveshownimpressiveperformance
onmanysequence-relatedtasks.
Thearchitecturecommonlyusesanattentional
mechanismwhichallowsthemodel
tolearnalignmentsbetweenthesource
andthetargetsequence.Mostattentional
mechanismsusedtoday
下载地址
用户评论
