“2019-12-02”版本间的差异

来自cslt Wiki
跳转至： 导航、 搜索
 
（6位用户的8个中间修订版本未显示）
第19行： 第19行：
 
|Yunqi Cai
 
|Yunqi Cai
 
||  
 
||  
−
*  
+
* trained LDA-flow with guass constrain on mean-vector
 
||
 
||
−
*  
+
* nonlinear LDA-flow：appending some couple layers (VP) before the linear layer
 
||
 
||
 
*   
 
*   
第55行： 第55行：
 
* review the code about flow and glow
 
* review the code about flow and glow
 
||
 
||
−
* Experiment about [link http://192.168.0.51:5555/cgi-bin/cvss/cvss_request.pl?account=shiying&step=view_request&cvssid=725]
+
* Experiment about cvss 725
 
||
 
||
 
*   
 
*   
第77行： 第77行：
 
* Completed the technical summary document of cmd module of CeaOS platform
 
* Completed the technical summary document of cmd module of CeaOS platform
 
||
 
||
−
*  
+
* Start to write module code
 +
* Plan data collection
 
||
 
||
 
*   
 
*   
第96行： 第97行：
 
|Yue Fan
 
|Yue Fan
 
||  
 
||  
−
*  
+
* Environment configuration and data preprocessing
 +
* CN-Celeb egs
 
||
 
||
−
*  
+
* Experiences on flow
 
||
 
||
 
*   
 
*   
第107行： 第109行：
 
|Jiawen Kang
 
|Jiawen Kang
 
||  
 
||  
−
* Look into Flow applications;
+
* Look into Flow applications & flow-based models;
 
*Hands on experiences about speech conversion;
 
*Hands on experiences about speech conversion;
 
*Graduate program application & Research proposal.
 
*Graduate program application & Research proposal.
第133行： 第135行：
 
|Sitong Cheng
 
|Sitong Cheng
 
||  
 
||  
−
*  
+
* data preparation of ERJ
 +
* script of data augmentation with musan and thchs30
 
||
 
||
−
*  
+
* run glow of train set with noise to find the destination which noise is independence
 
||
 
||
 
*   
 
*   
第144行： 第147行：
 
|Zhixin Liu
 
|Zhixin Liu
 
||  
 
||  
−
*  
+
* Get familiar with ERJ
 +
* data preparation of ERJ
 
||
 
||
−
*  
+
* train data using glow
 
||
 
||
 
*   
 
*   

2019年12月2日 (一) 07:53的最后版本

People This Week Next Week Task Tracking (DeadLine)
Dong Wang
  • Go through papers on ICASSP 2019 and Interspeech 2019
  • MelGan investigation
  • MelGan training
  • Interspeech 2019 and NIPS 2019 paper checking
Yunqi Cai
  • trained LDA-flow with guass constrain on mean-vector
  • nonlinear LDA-flow：appending some couple layers (VP) before the linear layer
Zhiyuan Tang
  • Trained a basic glow with timit.
  • Re-train the model with labels.
Lantian Li
  • CN-Celeb egs
  • Flow review and Hire-Flow design
  • Hire-Flow
Ying Shi
  • review the code about flow and glow
  • Experiment about cvss 725
Wenqiang Du
  • Prepare for the experiment in my graduation thesis(CNN+TDNN+Attention）
  • Model parameter adjustment
Yue Liu
  • Run through the CeaOS source code
  • Completed the technical summary document of cmd module of CeaOS platform
  • Start to write module code
  • Plan data collection
Haoran Sun
Yue Fan
  • Environment configuration and data preprocessing
  • CN-Celeb egs
  • Experiences on flow
Jiawen Kang
  • Look into Flow applications & flow-based models;
  • Hands on experiences about speech conversion;
  • Graduate program application & Research proposal.
  • Continue experiences about speech conversion;
  • Research proposal.
Ruiqi Liu
  • did experiments and read papers on ‘speech factorization’ and 'Glow'.
  • Learn some basics and pytorch.
  • Determine my research direction.
Sitong Cheng
  • data preparation of ERJ
  • script of data augmentation with musan and thchs30
  • run glow of train set with noise to find the destination which noise is independence
Zhixin Liu
  • Get familiar with ERJ
  • data preparation of ERJ
  • train data using glow
取自“http://cslt.org/mediawiki/index.php?title=2019-12-02&oldid=34392”