“Zhiyong Zhang - 130415”版本间的差异

来自cslt Wiki
跳转至: 导航搜索
(以内容“Weekly Report 1. Debug the SLH-based VAD procedure(Full waveform number of my SLH autocorrelation cure is less than the paper report, confusing thing). 2. Read some cor...”创建新页面)
 
第1行: 第1行:
 +
 +
 
Weekly Report
 
Weekly Report
 +
 
1. Debug the SLH-based VAD procedure(Full waveform number of my SLH autocorrelation cure is less than the paper report, confusing thing).
 
1. Debug the SLH-based VAD procedure(Full waveform number of my SLH autocorrelation cure is less than the paper report, confusing thing).
 +
 
2. Read some corpus about the speech recogniton tools KALDI.  
 
2. Read some corpus about the speech recogniton tools KALDI.  
 +
  
 
Target of next week
 
Target of next week
 +
 
1. Test the dataset from Tencent using HTK's HMM model and the HMM model converted from HTK pattern to KALDI pattern respectively.  
 
1. Test the dataset from Tencent using HTK's HMM model and the HMM model converted from HTK pattern to KALDI pattern respectively.  
 +
 
2. Read technology papers about mllr fmllr map mmi fmmi mpe lda hlda etc.
 
2. Read technology papers about mllr fmllr map mmi fmmi mpe lda hlda etc.

2013年4月15日 (一) 12:45的版本


Weekly Report

1. Debug the SLH-based VAD procedure(Full waveform number of my SLH autocorrelation cure is less than the paper report, confusing thing).

2. Read some corpus about the speech recogniton tools KALDI.


Target of next week

1. Test the dataset from Tencent using HTK's HMM model and the HMM model converted from HTK pattern to KALDI pattern respectively.

2. Read technology papers about mllr fmllr map mmi fmmi mpe lda hlda etc.