“Zhiyong Zhang”版本间的差异

来自cslt Wiki
跳转至: 导航搜索
Task schedule
第1行: 第1行:
==Task schedule==
+
==Task schedules==
  
 +
===Summary===
 +
    --------------------------------------------------------------------------------------------------------
 +
    Priority | Tasks name                    |      Status          |    Notions
 +
    --------------------------------------------------------------------------------------------------------   
 +
        1    | Bi-Softmax                    | ■■■■■■■□□□ | 1400h am training and problem fixing
 +
    --------------------------------------------------------------------------------------------------------
 +
        2    | RNN+DAE                      | □□□□□□□□□□ |
 +
    --------------------------------------------------------------------------------------------------------
 +
 +
==Speech Recognition==
 +
===Multi-lingual Am training===
 +
====Bi-Softmax====
 +
* Using two distinct softmax for English and Chinese data.
 +
* Testing on 100h-Ch+100h-En, better performance observed.
 +
* Now testing the source code on 1400h_8k data, but stange decoding results got.
 +
  Need to further investigate.
  
 
==To Do==
 
==To Do==
  
 
===Papers To Read ===
 
===Papers To Read ===

2015年1月7日 (三) 00:57的版本

Task schedules

Summary

   --------------------------------------------------------------------------------------------------------
    Priority | Tasks name                    |      Status          |     Notions
   --------------------------------------------------------------------------------------------------------    
        1    | Bi-Softmax                    | ■■■■■■■□□□ | 1400h am training and problem fixing
   --------------------------------------------------------------------------------------------------------
        2    | RNN+DAE                       | □□□□□□□□□□ |
   --------------------------------------------------------------------------------------------------------

Speech Recognition

Multi-lingual Am training

Bi-Softmax

  • Using two distinct softmax for English and Chinese data.
  • Testing on 100h-Ch+100h-En, better performance observed.
  • Now testing the source code on 1400h_8k data, but stange decoding results got.
 Need to further investigate.

To Do

Papers To Read