海词手机词典
  • The approach uses tri-phone models as recognition sub-word unit in the acoustics and speech layer, and uses statistic language model in syntax layer which is based on bi-gram and tri-gram.

    播放读音 播放读音