“NLP Status Report 2016-11-07”版本间的差异

来自cslt Wiki
跳转至: 导航搜索
 
(2位用户的4个中间修订版本未显示)
第4行: 第4行:
 
| rowspan="4"|2016/10/31
 
| rowspan="4"|2016/10/31
 
|Yang Feng ||
 
|Yang Feng ||
1. ran rnng+MN with offline lstm states, but failed, as the actions in oracle and lstm states conflicted.
+
* ran rnng+MN with offline lstm states, but failed, as the actions in oracle and lstm states conflicted.
2. wrote the code to generate lstm states online and ran some experiments. [[http://cslt.riit.tsinghua.edu.cn/mediawiki/images/f/f8/Progress_of_RNNG_with_memory_network.pdf report]]
+
* wrote the code to generate lstm states online and ran some experiments. [[http://cslt.riit.tsinghua.edu.cn/mediawiki/images/f/f8/Progress_of_RNNG_with_memory_network.pdf report]]
3. transferred the code to Shiyue.
+
* transferred the code to Shiyue.
 
||
 
||
1. optimize rnng+MN;
+
* optimize rnng+MN;
2. do the work of sequence-to-sequence generation;
+
* do the work of sequence-to-sequence generation;
3. Huilan work.
+
* Huilan work.
 
|-
 
|-
 
|Jiyuan Zhang ||
 
|Jiyuan Zhang ||
第17行: 第17行:
 
*ran memory model to test whether to achieve the expected
 
*ran memory model to test whether to achieve the expected
 
*results of model of poor effect [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/6/6c/OriginalModel.pdf OriginalModel] [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/0/02/MemoryModel.pdf memoryModel]
 
*results of model of poor effect [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/6/6c/OriginalModel.pdf OriginalModel] [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/0/02/MemoryModel.pdf memoryModel]
*results of model of different memory(The results are almost the same)
+
*results of models of different memory(The results are almost the same)
 +
*results of models of orignal and memory(training same specific matrixes) [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/0/02/MemoryModel.pdf memoryModel]  [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/7/7b/OriginalModel_specificmatrixes.pdf originalModel]
 
||  
 
||  
 
*improve the memory model
 
*improve the memory model
 
*document
 
*document
 +
*make time table
 
|-
 
|-
 
|Andi Zhang ||
 
|Andi Zhang ||
第34行: 第36行:
 
* wrote an experiment report [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/2/2f/RNNG%2Bmm%E5%AE%9E%E9%AA%8C%E6%8A%A5%E5%91%8A.pdf]
 
* wrote an experiment report [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/2/2f/RNNG%2Bmm%E5%AE%9E%E9%AA%8C%E6%8A%A5%E5%91%8A.pdf]
 
||
 
||
 +
* see in time table , the 2016/11/7 work plan [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/Asr-language-processing-research-rnng-mn#Time_Table]
 
|-
 
|-
 
|}
 
|}

2016年11月7日 (一) 10:58的最后版本

Date People Last Week This Week
2016/10/31 Yang Feng
  • ran rnng+MN with offline lstm states, but failed, as the actions in oracle and lstm states conflicted.
  • wrote the code to generate lstm states online and ran some experiments. [report]
  • transferred the code to Shiyue.
  • optimize rnng+MN;
  • do the work of sequence-to-sequence generation;
  • Huilan work.
Jiyuan Zhang
  • finished the memory model
  • solved the bugs about rythme
  • ran memory model to test whether to achieve the expected
  • results of model of poor effect OriginalModel memoryModel
  • results of models of different memory(The results are almost the same)
  • results of models of orignal and memory(training same specific matrixes) memoryModel originalModel
  • improve the memory model
  • document
  • make time table
Andi Zhang
  • tried to run the code on gpu but failed, queried on github
  • make the final time table
  • successfully run the code on gpu
  • document on Memn2n(if have time)
Shiyue Zhang
  • finished the planned work in time table [1]
  • wrote an experiment report [2]
  • see in time table , the 2016/11/7 work plan [3]