“Asr-progress 2017.07”版本间的差异

来自cslt Wiki
跳转至: 导航搜索
(以“ ===Daily Report=== {|class="wikitable" ! Date !! Person !! start!! leave !! hours ||status |- | rowspan="4"|2017.7.3 |Yanqing Wang|| || || || |- |Yix...”为内容创建页面)
 
 
(2位用户的5个中间修订版本未显示)
第1行: 第1行:
 
 
===Daily Report===
 
===Daily Report===
  
 
{|class="wikitable"
 
{|class="wikitable"
 
! Date !! Person  !! start!! leave !! hours ||status
 
! Date !! Person  !! start!! leave !! hours ||status
 +
 +
 +
 +
 +
|-
 +
| rowspan="4"|2017.7.1
 +
|Yanqing Wang ||  11:00  || 20:00  ||  9h  ||  continue on experiments on 4 types of activation function
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||  9:00  ||  20:00  ||  11h  ||  do a meeting report on trivial events
 +
|-
 +
| Xiaofei Kang ||10:00 ||  21:00  ||  11h  ||  learn the Speech Recognition lesson from University of Edinburgh.
 +
|-
 +
 +
|-
 +
| rowspan="4"|2017.7.2
 +
|Yanqing Wang ||  11:00  ||  20:00  ||  9h    ||  continue on experiments on 4 types of activation function
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||  9:00  ||  20:00  ||    11h  ||  design a human test website
 +
|-
 +
| Xiaofei Kang ||10:00 ||  21:00  ||  11h  ||  plot a specturm of waveform with Matlab and Python.
 +
|-
 +
 
|-
 
|-
 
| rowspan="4"|2017.7.3
 
| rowspan="4"|2017.7.3
|Yanqing Wang||   ||   ||     ||   
+
|Yanqing Wang || 11:00  || 20:00  || 9h    ||  change the dimension to 1000 and retry the former exps
 
|-
 
|-
 
|Yixiang Chen ||    ||    ||      ||   
 
|Yixiang Chen ||    ||    ||      ||   
 
|-
 
|-
|Miao Zhang  ||   ||   ||     ||   
+
|Miao Zhang  || 9:00  || 20:00  ||   11h  ||  design a human test website
 
|-
 
|-
 
| Xiaofei Kang ||    ||    ||      ||   
 
| Xiaofei Kang ||    ||    ||      ||   
 
|-
 
|-
 +
  
 
|-
 
|-
第21行: 第47行:
 
|Yixiang Chen ||    ||    ||      ||   
 
|Yixiang Chen ||    ||    ||      ||   
 
|-
 
|-
|Miao Zhang  ||   ||   ||     ||   
+
|Miao Zhang  || 9:00  || 20:00  ||   11h  ||  design a human test website
 
|-
 
|-
| Xiaofei Kang ||   ||   ||     ||  
+
| Xiaofei Kang ||10:00 || 21:00  || 11h  || prepare my biweekly report.
 
|-
 
|-
  
 +
|-
 +
| rowspan="4"|2017.7.5
 +
|Yanqing Wang ||    ||    ||      || 
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||  9:00  ||  20:00  ||  11h  ||  design a human test website
 +
|-
 +
| Xiaofei Kang ||10:00 ||  21:00  ||  11h  || continue to prepare my biweekly report.
 +
|-
  
 
|-
 
|-
| rowspan="4"|2017.7.31
+
| rowspan="4"|2017.7.6
 +
|Yanqing Wang || 11:00  || 20:00  ||  9h    ||  change the dimension to 1000 and retry the former exps
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||  9:00  ||  20:00  ||  11h  ||  design a human test website
 +
|-
 +
| Xiaofei Kang ||10:00 ||  21:00  ||  11h  || continue to prepare my biweekly report.
 +
|-
 +
 
 +
|-
 +
| rowspan="4"|2017.7.7
 +
|Yanqing Wang ||  11:00  ||  20:00  ||  9h  ||  continue former exps, read source code of Kaldi
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||  9:00  ||  20:00  ||  11h  ||  finish the human test website
 +
|-
 +
| Xiaofei Kang ||10:00 ||  21:00  ||  11h  || continue to prepare my biweekly report.
 +
|-
 +
 
 +
|-
 +
| rowspan="4"|2017.7.8
 +
|Yanqing Wang ||  11:00  ||  20:00  ||  9h  ||  continue former exps, read source code of Kaldi
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||    ||    ||      || 
 +
|-
 +
| Xiaofei Kang ||  ||    ||    ||
 +
|-
 +
 
 +
|-
 +
| rowspan="4"|2017.7.9
 +
|Yanqing Wang ||  11:00  ||  20:00  ||  9h  ||  continue former exps, read source code of Kaldi
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||    ||    ||      || 
 +
|-
 +
| Xiaofei Kang ||  ||    ||    ||
 +
|-
 +
 
 +
|-
 +
| rowspan="4"|2017.7.10
 
|Yanqing Wang ||    ||    ||      ||   
 
|Yanqing Wang ||    ||    ||      ||   
 
|-
 
|-
 
|Yixiang Chen ||    ||    ||      ||   
 
|Yixiang Chen ||    ||    ||      ||   
 
|-
 
|-
|Miao Zhang  ||    ||    ||      ||   
+
|Miao Zhang  ||  9:00  ||  20:00  ||  11h  ||  Read the paper of Paralinguistics
 +
|-
 +
| Xiaofei Kang ||10:00 ||  21:00  ||  11h  ||  prepare a PPT for biweekly report
 +
|-
 +
 
 +
|-
 +
| rowspan="4"|2017.7.11
 +
|Yanqing Wang ||  11:00  ||  20:00  ||  9h  ||  start to change the source code of Kaldi in order to implement retraining the nnet
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||  9:00  ||  20:00  ||  11h  ||  Read the paper of Paralinguistics
 +
|-
 +
| Xiaofei Kang || 10:00|| 21:00 ||  11h  ||  do a biweekly report about STFT Spectrogram
 +
|-
 +
 
 +
|-
 +
| rowspan="4"|2017.7.12
 +
|Yanqing Wang ||    ||    ||      || 
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||  9:00  ||  20:00  ||  11h  ||  Read the paper of Paralinguistics
 +
|-
 +
| Xiaofei Kang || 10:00|| 21:00 ||  11h  ||  learn the srfft transform from kaldi.
 +
|-
 +
 
 +
|-
 +
| rowspan="4"|2017.7.13
 +
|Yanqing Wang ||  14:00  || 20:00  ||  6h    ||  start to change the source code of Kaldi in order to implement retraining the nnet
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||  9:00  ||  20:00  ||  11h  ||  Make a plan for recording
 +
|-
 +
| Xiaofei Kang || 10:00|| 21:00 ||  11h  ||  pick out the relationship between kaldi and python about spectrum transform.
 +
|-
 +
 
 +
|-
 +
| rowspan="4"|2017.7.14
 +
|Yanqing Wang ||    ||    ||      || 
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||  9:00  ||  20:00  ||  11h  ||  Read material from Teacher Li
 +
|-
 +
| Xiaofei Kang ||10:00 ||  21:00  ||  11h  || learn the nnet3 theory from tanghui
 +
|-
 +
 
 +
|-
 +
| rowspan="4"|2017.7.15
 +
|Yanqing Wang ||  11:00  ||  20:00  ||  9h  ||  change the source code of Kaldi in order to implement retraining the nnet
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||    ||    ||      || 
 +
|-
 +
| Xiaofei Kang ||10:00 ||  21:00  ||  11h  || do a speech conversion experiment with Python.
 +
|-
 +
 
 +
|-
 +
| rowspan="4"|2017.7.16
 +
|Yanqing Wang ||    ||    ||      || 
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||    ||    ||      || 
 +
|-
 +
| Xiaofei Kang ||10:00 ||  21:00  ||  11h  ||  prepare a PPT for biweekly supplementary report
 +
|-
 +
 
 +
 
 +
|-
 +
| rowspan="4"|2017.7.17
 +
|Yanqing Wang ||  11:00  ||  20:00  ||  9h  || change the source code of Kaldi
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||  9:00  ||  20:00  ||  11h  ||  check the book of deep learning
 +
|-
 +
| Xiaofei Kang ||10:00 ||  21:00  ||  11h  ||  do a biweekly supplementary report about speech conversion.
 +
|-
 +
 
 +
|-
 +
| rowspan="4"|2017.7.18
 +
|Yanqing Wang ||  11:00  ||  20:00  ||  9h  || change the source code of Kaldi
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||  9:00  ||  20:00  ||  11h  ||  check the book of deep learning
 +
|-
 +
| Xiaofei Kang ||10:00 ||  21:00  ||  11h  ||  learn the papers about speaker recognition
 +
|-
 +
 
 +
|-
 +
| rowspan="4"|2017.7.19
 +
|Yanqing Wang ||  11:00  ||  20:00  ||  9h  || after changing source code, compile Kaldi and redo the former exps
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||  9:00  ||  20:00  ||  11h    ||  work out the recording plan with instruction from Teacher Li
 +
|-
 +
| Xiaofei Kang ||10:00 ||  21:00  ||  11h  ||  repare the data set of Speaker Recognition : pick out whisper
 +
|-
 +
 
 +
|-
 +
| rowspan="4"|2017.7.20
 +
|Yanqing Wang ||  11:00  ||  20:00  ||  9h  || redo the former exps
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||  9:00  ||  20:00  ||  11h  ||  learnt kaldi and did experiments
 +
|-
 +
| Xiaofei Kang ||10:00 ||  21:00  ||  11h  ||    Learn the the nnet3 model
 +
|-
 +
 
 +
|-
 +
| rowspan="4"|2017.7.21
 +
|Yanqing Wang ||  11:00  ||  20:00  ||  9h  || redo the former exps and test the conclusions
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||  9:00  ||  20:00  ||  11h  ||  joined a meeting in Chinese Academy of Social Sciences
 +
|-
 +
| Xiaofei Kang ||10:00 ||  21:00  ||  11h  ||  run the nnet3 experiment
 +
|-
 +
 
 +
|-
 +
| rowspan="4"|2017.7.22
 +
|Yanqing Wang ||    ||    ||      || 
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||    ||    ||      ||   
 
|-
 
|-
 
| Xiaofei Kang ||    ||    ||      ||   
 
| Xiaofei Kang ||    ||    ||      ||   
 
|-
 
|-
|}
 
  
 +
|-
 +
| rowspan="4"|2017.7.23
 +
|Yanqing Wang ||    ||    ||      || 
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||    ||    ||      || 
 +
|-
 +
| Xiaofei Kang ||    ||    ||      || 
 +
|-
  
 +
|-
 +
| rowspan="4"|2017.7.24
 +
|Yanqing Wang ||  11:00  ||  20:00  ||  9h  || debug and find the wrong place of exps
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||  9:00  ||  20:00  ||  11h  ||  test performances on 12-style database
 +
|-
 +
| Xiaofei Kang ||10:00 ||  21:00  ||  11h  ||  read papers about speaker recognition
 +
|-
  
 +
|-
 +
| rowspan="4"|2017.7.25
 +
|Yanqing Wang ||  11:00  ||  20:00  ||  9h  || redo the former exps and test the conclusions
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||  9:00  ||  20:00  ||  11h  ||  test performances on 12-style database
 +
|-
 +
| Xiaofei Kang ||10:00 ||  21:00  ||  11h  || write a script to cut the vad of the wavefiles
 +
|-
  
 +
|-
 +
| rowspan="4"|2017.7.26
 +
|Yanqing Wang ||  11:00  ||  20:00  ||  9h  || redo the former exps
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||  9:00  ||  20:00  ||  11h  ||  test performances on 12-style database
 +
|-
 +
| Xiaofei Kang ||10:00 ||  21:00  ||  11h  || write a script to convert 8bit to 16bit.
 +
|-
  
 +
|-
 +
| rowspan="4"|2017.7.27
 +
|Yanqing Wang ||  11:00  ||  20:00  ||  9h  || redo the former exps
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||  9:00  ||  20:00  ||  11h    ||  test performances on 12-style database
 +
|-
 +
| Xiaofei Kang ||10:00 ||  21:00  ||  11h  ||  Finish the Speaker Recognition experiment:mouth with candy
 +
|-
  
===Time Off Table===
 
{| class="wikitable"
 
! Name      !! Days off
 
 
|-
 
|-
|Yanqing Wang||  
+
| rowspan="4"|2017.7.28
 +
|Yanqing Wang || 11:00  ||  20:00  ||  9h  || conclude the exps and start to write a report
 
|-
 
|-
|Yixiang Chen||   
+
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||  9:00  ||  20:00  ||  11h  ||  test performances on 12-style database
 +
|-
 +
| Xiaofei Kang ||10:00 ||  21:00  ||  11h  || Finish the Speaker Recognition experiment:normal chat
 +
|-
 +
 
 +
|-
 +
| rowspan="4"|2017.7.29
 +
|Yanqing Wang ||    ||    ||      || 
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||    ||    ||      || 
 +
|-
 +
| Xiaofei Kang ||    ||    ||      || 
 +
|-
 +
 
 +
|-
 +
| rowspan="4"|2017.7.30
 +
|Yanqing Wang ||    ||    ||      || 
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 +
|-
 +
|Miao Zhang  ||    ||    ||      || 
 +
|-
 +
| Xiaofei Kang ||    ||    ||      || 
 +
|-
 +
 
 +
 
 +
|-
 +
| rowspan="4"|2017.7.31
 +
|Yanqing Wang ||  11:00  ||  20:00  ||  9h  || learn to use LaTeX
 +
|-
 +
|Yixiang Chen ||    ||    ||      || 
 
|-
 
|-
|Miao Zhang||   
+
|Miao Zhang   ||  9:00  ||  20:00  ||      ||  optimize the vad parameter to improve the performance
 
|-
 
|-
|Xiaofei Kang ||
+
| Xiaofei Kang ||10:00 ||  21:00  ||  11h  || read papers about i-vector and d-vector
 
|-
 
|-
 
|}
 
|}

2017年8月24日 (四) 05:27的最后版本

Daily Report

Date Person start leave hours status



2017.7.1 Yanqing Wang 11:00 20:00 9h continue on experiments on 4 types of activation function
Yixiang Chen
Miao Zhang 9:00 20:00 11h do a meeting report on trivial events
Xiaofei Kang 10:00 21:00 11h learn the Speech Recognition lesson from University of Edinburgh.
2017.7.2 Yanqing Wang 11:00 20:00 9h continue on experiments on 4 types of activation function
Yixiang Chen
Miao Zhang 9:00 20:00 11h design a human test website
Xiaofei Kang 10:00 21:00 11h plot a specturm of waveform with Matlab and Python.
2017.7.3 Yanqing Wang 11:00 20:00 9h change the dimension to 1000 and retry the former exps
Yixiang Chen
Miao Zhang 9:00 20:00 11h design a human test website
Xiaofei Kang
2017.7.4 Yanqing Wang
Yixiang Chen
Miao Zhang 9:00 20:00 11h design a human test website
Xiaofei Kang 10:00 21:00 11h prepare my biweekly report.
2017.7.5 Yanqing Wang
Yixiang Chen
Miao Zhang 9:00 20:00 11h design a human test website
Xiaofei Kang 10:00 21:00 11h continue to prepare my biweekly report.
2017.7.6 Yanqing Wang 11:00 20:00 9h change the dimension to 1000 and retry the former exps
Yixiang Chen
Miao Zhang 9:00 20:00 11h design a human test website
Xiaofei Kang 10:00 21:00 11h continue to prepare my biweekly report.
2017.7.7 Yanqing Wang 11:00 20:00 9h continue former exps, read source code of Kaldi
Yixiang Chen
Miao Zhang 9:00 20:00 11h finish the human test website
Xiaofei Kang 10:00 21:00 11h continue to prepare my biweekly report.
2017.7.8 Yanqing Wang 11:00 20:00 9h continue former exps, read source code of Kaldi
Yixiang Chen
Miao Zhang
Xiaofei Kang
2017.7.9 Yanqing Wang 11:00 20:00 9h continue former exps, read source code of Kaldi
Yixiang Chen
Miao Zhang
Xiaofei Kang
2017.7.10 Yanqing Wang
Yixiang Chen
Miao Zhang 9:00 20:00 11h Read the paper of Paralinguistics
Xiaofei Kang 10:00 21:00 11h prepare a PPT for biweekly report
2017.7.11 Yanqing Wang 11:00 20:00 9h start to change the source code of Kaldi in order to implement retraining the nnet
Yixiang Chen
Miao Zhang 9:00 20:00 11h Read the paper of Paralinguistics
Xiaofei Kang 10:00 21:00 11h do a biweekly report about STFT Spectrogram
2017.7.12 Yanqing Wang
Yixiang Chen
Miao Zhang 9:00 20:00 11h Read the paper of Paralinguistics
Xiaofei Kang 10:00 21:00 11h learn the srfft transform from kaldi.
2017.7.13 Yanqing Wang 14:00 20:00 6h start to change the source code of Kaldi in order to implement retraining the nnet
Yixiang Chen
Miao Zhang 9:00 20:00 11h Make a plan for recording
Xiaofei Kang 10:00 21:00 11h pick out the relationship between kaldi and python about spectrum transform.
2017.7.14 Yanqing Wang
Yixiang Chen
Miao Zhang 9:00 20:00 11h Read material from Teacher Li
Xiaofei Kang 10:00 21:00 11h learn the nnet3 theory from tanghui
2017.7.15 Yanqing Wang 11:00 20:00 9h change the source code of Kaldi in order to implement retraining the nnet
Yixiang Chen
Miao Zhang
Xiaofei Kang 10:00 21:00 11h do a speech conversion experiment with Python.
2017.7.16 Yanqing Wang
Yixiang Chen
Miao Zhang
Xiaofei Kang 10:00 21:00 11h prepare a PPT for biweekly supplementary report
2017.7.17 Yanqing Wang 11:00 20:00 9h change the source code of Kaldi
Yixiang Chen
Miao Zhang 9:00 20:00 11h check the book of deep learning
Xiaofei Kang 10:00 21:00 11h do a biweekly supplementary report about speech conversion.
2017.7.18 Yanqing Wang 11:00 20:00 9h change the source code of Kaldi
Yixiang Chen
Miao Zhang 9:00 20:00 11h check the book of deep learning
Xiaofei Kang 10:00 21:00 11h learn the papers about speaker recognition
2017.7.19 Yanqing Wang 11:00 20:00 9h after changing source code, compile Kaldi and redo the former exps
Yixiang Chen
Miao Zhang 9:00 20:00 11h work out the recording plan with instruction from Teacher Li
Xiaofei Kang 10:00 21:00 11h repare the data set of Speaker Recognition : pick out whisper
2017.7.20 Yanqing Wang 11:00 20:00 9h redo the former exps
Yixiang Chen
Miao Zhang 9:00 20:00 11h learnt kaldi and did experiments
Xiaofei Kang 10:00 21:00 11h Learn the the nnet3 model
2017.7.21 Yanqing Wang 11:00 20:00 9h redo the former exps and test the conclusions
Yixiang Chen
Miao Zhang 9:00 20:00 11h joined a meeting in Chinese Academy of Social Sciences
Xiaofei Kang 10:00 21:00 11h run the nnet3 experiment
2017.7.22 Yanqing Wang
Yixiang Chen
Miao Zhang
Xiaofei Kang
2017.7.23 Yanqing Wang
Yixiang Chen
Miao Zhang
Xiaofei Kang
2017.7.24 Yanqing Wang 11:00 20:00 9h debug and find the wrong place of exps
Yixiang Chen
Miao Zhang 9:00 20:00 11h test performances on 12-style database
Xiaofei Kang 10:00 21:00 11h read papers about speaker recognition
2017.7.25 Yanqing Wang 11:00 20:00 9h redo the former exps and test the conclusions
Yixiang Chen
Miao Zhang 9:00 20:00 11h test performances on 12-style database
Xiaofei Kang 10:00 21:00 11h write a script to cut the vad of the wavefiles
2017.7.26 Yanqing Wang 11:00 20:00 9h redo the former exps
Yixiang Chen
Miao Zhang 9:00 20:00 11h test performances on 12-style database
Xiaofei Kang 10:00 21:00 11h write a script to convert 8bit to 16bit.
2017.7.27 Yanqing Wang 11:00 20:00 9h redo the former exps
Yixiang Chen
Miao Zhang 9:00 20:00 11h test performances on 12-style database
Xiaofei Kang 10:00 21:00 11h Finish the Speaker Recognition experiment:mouth with candy
2017.7.28 Yanqing Wang 11:00 20:00 9h conclude the exps and start to write a report
Yixiang Chen
Miao Zhang 9:00 20:00 11h test performances on 12-style database
Xiaofei Kang 10:00 21:00 11h Finish the Speaker Recognition experiment:normal chat
2017.7.29 Yanqing Wang
Yixiang Chen
Miao Zhang
Xiaofei Kang
2017.7.30 Yanqing Wang
Yixiang Chen
Miao Zhang
Xiaofei Kang
2017.7.31 Yanqing Wang 11:00 20:00 9h learn to use LaTeX
Yixiang Chen
Miao Zhang 9:00 20:00 optimize the vad parameter to improve the performance
Xiaofei Kang 10:00 21:00 11h read papers about i-vector and d-vector