Mengyuan Zhao 2015-12-28

来自cslt Wiki
跳转至： 导航、 搜索

RNN AM

  • AM training on big dataset --mengyuan
  • Further test nnet3-ctc training on 4000h 8k & 10000h 16k dataset. Tune decoding configures (phone-lm-weight, acwt, blank-scale), but ctc is still about 5 percent worse than standard nnet3.
取自“http://index.cslt.org/mediawiki/index.php?title=Mengyuan_Zhao_2015-12-28&oldid=18339”