“2019-01-09”版本间的差异

来自cslt Wiki
跳转至: 导航搜索
 
(8位用户的12个中间修订版本未显示)
第9行: 第9行:
 
|Yibo Liu
 
|Yibo Liu
 
||  
 
||  
*  
+
* Improved Iambics generation.
 +
* Learned about GAN and VAE.
 
||
 
||
*  
+
* Try to implement VAE in the existing model.
 +
* Try to implement BERT.
 +
* Improve the quality of Iambics.
 
||
 
||
 
*   
 
*   
第22行: 第25行:
 
|Xiuqi Jiang
 
|Xiuqi Jiang
 
||  
 
||  
*
+
* Went deeper into current model and discussed about the possibility of VAE
 +
* Song iambics can now be generated and different styles of tunes are available
 
||
 
||
*  
+
* Polish up the iambics parts
 
||
 
||
 
*   
 
*   
第35行: 第39行:
 
|Jiayao Wu
 
|Jiayao Wu
 
||  
 
||  
* finished the KWS chapter for Speechbook[http://cslt.riit.tsinghua.edu.cn/mediawiki/images/c/c5/Kws.pdf kws]
+
* finished the speech book -- [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/c/c5/Kws.pdf kws]
 
* doing experiment of node -pruning on WSJ chain model
 
* doing experiment of node -pruning on WSJ chain model
 
||
 
||
第49行: 第53行:
 
|Zhaodi Qi
 
|Zhaodi Qi
 
||  
 
||  
*  
+
* Run LID (about Chinese English and Japanese) and Control configs_size within 20M,But it still running.
 +
* Data segmentation completed, and Waiting for test.
 
||
 
||
*  
+
* Finish the speech book
 +
* Completion the tdnn system performance comparison of test sets of different lengths
 
||
 
||
 
*   
 
*   
第62行: 第68行:
 
|Jiawei Yu
 
|Jiawei Yu
 
||  
 
||  
*  
+
* Done for max-margin recipe.
 +
* write the speech book.
 +
* learn to use Tensorflow for attention experiment.
 
||
 
||
*  
+
* finish the emotion recognition speech book.
 +
* Keep learning the Tensorflow and move max-margin experiment to this platform.
 
||
 
||
 
*   
 
*   
第73行: 第82行:
 
|Yunqi Cai
 
|Yunqi Cai
 
||  
 
||  
*
+
*Run wsj data; Read paper about RNN-LM
 +
||
 +
*Use RNN-LM to rescore the wsj data
 
||
 
||
 
*
 
*
||
 
*
 
 
|-
 
|-
  
第84行: 第93行:
 
|Dan He
 
|Dan He
 
||  
 
||  
*  
+
* Compared the inference time, but the results is not well.
 +
*After TT decomposing the two full-connection layers, it is found that the testing accuracy is very low.
 
||
 
||
*
+
*According to some problems found in the experiment, continue to do comparative experiments and analyze the reasons.
 
||
 
||
 
*
 
*
第98行: 第108行:
 
||  
 
||  
 
* Wrote a brief [https://github.com/zyzisyz/VPR-wx-client document]  
 
* Wrote a brief [https://github.com/zyzisyz/VPR-wx-client document]  
* submit the source code to [https://gitlab.com/zyzisyz/nebula-listen GitLab].
+
* submited the source code to [https://gitlab.com/zyzisyz/nebula-listen GitLab].
 
||
 
||
 
* revise for my school subjects and prepare for the final-term examinations.  
 
* revise for my school subjects and prepare for the final-term examinations.  

2019年1月9日 (三) 04:32的最后版本

People Last Week This Week Task Tracking (DeadLine)
Yibo Liu
  • Improved Iambics generation.
  • Learned about GAN and VAE.
  • Try to implement VAE in the existing model.
  • Try to implement BERT.
  • Improve the quality of Iambics.
Xiuqi Jiang
  • Went deeper into current model and discussed about the possibility of VAE
  • Song iambics can now be generated and different styles of tunes are available
  • Polish up the iambics parts
Jiayao Wu
  • finished the speech book -- kws
  • doing experiment of node -pruning on WSJ chain model
  • keep on node-pruning research
Zhaodi Qi
  • Run LID (about Chinese English and Japanese) and Control configs_size within 20M,But it still running.
  • Data segmentation completed, and Waiting for test.
  • Finish the speech book
  • Completion the tdnn system performance comparison of test sets of different lengths
Jiawei Yu
  • Done for max-margin recipe.
  • write the speech book.
  • learn to use Tensorflow for attention experiment.
  • finish the emotion recognition speech book.
  • Keep learning the Tensorflow and move max-margin experiment to this platform.
Yunqi Cai
  • Run wsj data; Read paper about RNN-LM
  • Use RNN-LM to rescore the wsj data
Dan He
  • Compared the inference time, but the results is not well.
  • After TT decomposing the two full-connection layers, it is found that the testing accuracy is very low.
  • According to some problems found in the experiment, continue to do comparative experiments and analyze the reasons.
Yang Zhang
  • revise for my school subjects and prepare for the final-term examinations.