“M2asr-delivery-phaseI”版本间的差异

来自cslt Wiki
跳转至: 导航搜索
Mongolian
Uyghur
第13行: 第13行:
 
[http://cslt.riit.tsinghua.edu.cn/mediawiki/images/7/71/Uyghur.train_trans.txt train set transcription]
 
[http://cslt.riit.tsinghua.edu.cn/mediawiki/images/7/71/Uyghur.train_trans.txt train set transcription]
  
[http://cslt.riit.tsinghua.edu.cn/mediawiki/images/5/57/Uyghur.test_trans.txt test set transcription]
+
[http://cslt.riit.tsinghua.edu.cn/mediawiki/images/7/70/Uy.trans.test.new.txt test set transcription]
 
+
In the first phase, we release 100h speech audio of M2ASR-UYGH-PRIO for free. LICENSE is needed.
+
  
 +
In the first phase, we release 100h speech audio of M2ASR-UYGH-PRIO for free. LICENSE is needed.
  
 
==Kazakh==
 
==Kazakh==

2017年12月21日 (四) 07:57的版本

Language resources

Uyghur

M2ASR-UYGH-PRIO Speech Database (XJU)

phone set

lexicon

train set transcription

test set transcription

In the first phase, we release 100h speech audio of M2ASR-UYGH-PRIO for free. LICENSE is needed.

Kazakh

M2ASR-KAZAK-PRIO Speech Database (XJU)

phone set

lexicon

train transcription

test transcription

In the first phase, we release 50h speech audio of M2ASR-KAZAK-PRIO for free. LICENSE is needed.


Tibetan

M2ASR-TIBETAN-PRIO Speech Database (NMU)

phone set

lexicon

train transcription

test transcription

In the first phase, we release 10h speech audio of M2ASR-TIBETAN-PRIO for free. LICENSE is needed.

Note: We use the unicode of tibetan to represent tibetan letters in the transcriptions at present.

Mongolian

M2ASR-MONGOLIAN-PRIO Speech Database (NMU)

lexicon

Kirgiz

Coming soon.


Models

Systems

Platform

License

Contact Dr. Dong Wang (wangdong99@mails.tsinghua.edu.cn)