“2021-05-24”版本间的差异

来自cslt Wiki
跳转至: 导航搜索
 
(4位用户的4个中间修订版本未显示)
第6行: 第6行:
 
|Dong Wang
 
|Dong Wang
 
||  
 
||  
*
+
* Unsupervised LSTF
 +
* VAE LSTF
 
||
 
||
*
+
* More complicated LSTF decoder
 
||
 
||
 
*   
 
*   
第28行: 第29行:
 
|Lantian Li
 
|Lantian Li
 
||  
 
||  
*  
+
* Review CN-Celeb collection pipeline.
*  
+
* Finish CN-Celeb rebuttal except Cllr.
 +
* Prepare Cross-device delivery.
 
||
 
||
*  
+
* Submit CN-Celeb rebuttal.
 +
* Complete Cross-device delivery.
 
||
 
||
 
*   
 
*   
第67行: 第70行:
 
||  
 
||  
 
* Organize experimental codes.
 
* Organize experimental codes.
* Sth about operation of the official account.
+
* Sth about operation and propaganda of the official account.
 
||
 
||
*  
+
* Continue the experiment part.
 
||
 
||
 
*   
 
*   
第102行: 第105行:
 
|Tiankai Zhi
 
|Tiankai Zhi
 
||  
 
||  
*  
+
* Infrared lips and speech
||
+
||  
 
*  
 
*  
 
||
 
||

2021年5月24日 (一) 05:59的最后版本

People This Week Next Week Task Tracking (DeadLine)
Dong Wang
  • Unsupervised LSTF
  • VAE LSTF
  • More complicated LSTF decoder
Yunqi Cai
  • thermal to RGB image investigation
Lantian Li
  • Review CN-Celeb collection pipeline.
  • Finish CN-Celeb rebuttal except Cllr.
  • Prepare Cross-device delivery.
  • Submit CN-Celeb rebuttal.
  • Complete Cross-device delivery.
Ying Shi
  • webrtc retrain
  • test webrtc retrain model
Haoran Sun
  • reproduction of speechflow
  • proposal
  • report for proposal
  • continuous speaker embedding of speechflow
  • investigation of LSTF
Jiao Han
  • Organize experimental codes.
  • Sth about operation and propaganda of the official account.
  • Continue the experiment part.
Yang Zhang
  • train some new models for future experiments and upload to thu cloud
  • train more models with data augment and other loss functions
Di Wang
  • Carry out the second evaluation activities.
  • Learn ASR egs.
  • Data analysis.
  • Learn m2asr.
Tiankai Zhi
  • Infrared lips and speech
Chen Chen
  • the speech enhancement report
  • learn recent speech enhancement studies such as WaveRNN/CRN/WaveNet
Jingxin Shen
  • reproduction of pix2pix model
  • test of the cycyle_GAN model