“Sinovoice-2014-01-13”版本间的差异

来自cslt Wiki
跳转至: 导航搜索
6000 hour 16k trainin
 
(相同用户的3个中间修订版本未显示)
第12行: 第12行:
  
 
==Corpora==
 
==Corpora==
* How many extra data were obtained?
+
* 60 hour data were cut this week
 
+
* Just send out to vendors for labeling
 +
* Waiting for out-source platform construction
 +
* We assume 60 hour data per week in future
  
 
==470 hour 8k training==
 
==470 hour 8k training==
第31行: 第33行:
 
==6000 hour 16k training==
 
==6000 hour 16k training==
  
* Audio files done. File with incorrect sampling rates were removed
+
* Audio files ready. Files with incorrect sampling rates were removed
* Lexicon and LM were done
+
* Lexicon and LM were ready
 
* Making MFCC features
 
* Making MFCC features
 +
* Initial model (6 iterations etc) can be delivered before the spring holiday
  
 
=DNN Decoder=
 
=DNN Decoder=
 
* Initial trail of DNN decoder based on the Sinovoice code was failed, largely due to FST compiler
 
* Initial trail of DNN decoder based on the Sinovoice code was failed, largely due to FST compiler
 
* Change the strategy to an integrated approach: use the sinovoice system to control connections, and use Kaldi base for asr engine
 
* Change the strategy to an integrated approach: use the sinovoice system to control connections, and use Kaldi base for asr engine
 +
* Xiaoming will do some investigation on the Sinovoice FST compiler, while Liu Chao will focus on the Kaldi-based decoder

2014年1月13日 (一) 07:11的最后版本

Project management

  • Xiaoming and Xiao Na were added into the mail list
  • Potential Huawei conference-transcribing project was discussed

DNN training

Environment setting

  • New disk space (3T) was created and mounted at /nfs/disk1
  • Jobs with 100 threads work fine on the cluster

Corpora

  • 60 hour data were cut this week
  • Just send out to vendors for labeling
  • Waiting for out-source platform construction
  • We assume 60 hour data per week in future

470 hour 8k training

  • CE training done
  • MPE training partially done
Model CE MPE1 MPE2 MPE3 MPE4
4k states 23.27/22.85 21.35/18.87 21.18/18.76 21.07/18.54
8k states 22.16/22.22 - 20.36/17.94 -

6000 hour 16k training

  • Audio files ready. Files with incorrect sampling rates were removed
  • Lexicon and LM were ready
  • Making MFCC features
  • Initial model (6 iterations etc) can be delivered before the spring holiday

DNN Decoder

  • Initial trail of DNN decoder based on the Sinovoice code was failed, largely due to FST compiler
  • Change the strategy to an integrated approach: use the sinovoice system to control connections, and use Kaldi base for asr engine
  • Xiaoming will do some investigation on the Sinovoice FST compiler, while Liu Chao will focus on the Kaldi-based decoder