“2024-04-01”版本间的差异

来自cslt Wiki
跳转至: 导航搜索
 
(10位用户的11个中间修订版本未显示)
第37行: 第37行:
 
* Finish Chinese Overlap ASR Model training
 
* Finish Chinese Overlap ASR Model training
 
* Detect wake-up words from Continuous speech still in the training process
 
* Detect wake-up words from Continuous speech still in the training process
 +
* [https://z1et6d3xtb.feishu.cn/wiki/OfyEwaqjkiCRmGk1Cl0cZKGEnIT?from=from_copylink group work ]
 
||
 
||
 
* Finish wake-up Model
 
* Finish wake-up Model
第49行: 第50行:
 
|Zhenghai You
 
|Zhenghai You
 
||  
 
||  
*  
+
* Huawei project[https://z1et6d3xtb.feishu.cn/wiki/VOF4wORTWiw7tZkD3M4cZFkInOe]
 +
* Minimal loss without cohort
 
||
 
||
 
*  
 
*  
第70行: 第72行:
 
|Chen Chen
 
|Chen Chen
 
||  
 
||  
*  
+
* Finish thesis v1
 +
* group work [https://z1et6d3xtb.feishu.cn/docx/YMiLdBy4Yo59ITxmoYDcCfdMnBf?from=from_copylink]
 +
||
 +
* refine thesis
 
||
 
||
 
*  
 
*  
||
 
 
 
|-
 
|-
  
第81行: 第84行:
 
|Xiaolou Li
 
|Xiaolou Li
 
||  
 
||  
*  
+
* reproduce VSR structures
 
||
 
||
 
*  
 
*  
第103行: 第106行:
 
|Pengqi Li
 
|Pengqi Li
 
||   
 
||   
*  
+
* summary[https://z1et6d3xtb.feishu.cn/docx/Mt46dqE8UoBnuBxu7q8cPletnzb] of speech processing XAI for NSFC
 +
** finished structure and almost content
 
||
 
||
*
+
* polish, reference, tables
 
||
 
||
 
*   
 
*   
第114行: 第118行:
 
|Wan Lin
 
|Wan Lin
 
||  
 
||  
*  
+
* Neural Scoring
 +
** Adjust the training and testing phase of mutil-speaker tasks
 +
** Design of the new architecture
 +
* Graduation paper
 
||
 
||
 
*
 
*
第125行: 第132行:
 
|Tianhao Wang
 
|Tianhao Wang
 
||  
 
||  
*  
+
* EA-ASP (from SJTU) reproducing
 +
** data preparation
 +
** EA-ASP implement
 
||
 
||
 
*  
 
*  
第136行: 第145行:
 
|Zhenyu Zhou
 
|Zhenyu Zhou
 
||  
 
||  
*
+
*Baseline design[https://z1et6d3xtb.feishu.cn/docx/BywjdkGvNou12sxQ4dAcxYa9noh]
 
||
 
||
 
*
 
*
第147行: 第156行:
 
|Junhui Chen
 
|Junhui Chen
 
||
 
||
*  
+
* reproduce frequency-channel attention for Neural Scoring
 
||
 
||
 
*
 
*
第158行: 第167行:
 
|Jiaying Wang
 
|Jiaying Wang
 
||  
 
||  
*  
+
* Huawei Project
 +
* icassp 2024 paper selection
 
||
 
||
*  
+
* speakerbeam + ResNet [-10,4]
 
||
 
||
 
*   
 
*   
第169行: 第179行:
 
|Yu Zhang
 
|Yu Zhang
 
||
 
||
*  
+
* SAC training
 +
* R2-SAC adapted database build (2020 - 2023, benchmark and stocks)
 
||
 
||
 
*
 
*

2024年4月1日 (一) 11:02的最后版本

People This Week Next Week Task Tracking (DeadLine)
Dong Wang
  • NeuralMag paper finalization
  • Review on TALIP
  • Design for AI primary education
Lantian Li
  • GPU status [1]
  • Submit three delivery packages (KTV-TSE, Bullying-D, SpeechActivity-D).
  • ASIP-BUPT (NeuralScoring [with MixT/MFA], CohortTSE)
  • AI course series
  • New machine (rabbit04)
Ying Shi
  • got sick and was unable to complete most of my work as expected
  • Finish Chinese Overlap ASR Model training
  • Detect wake-up words from Continuous speech still in the training process
  • group work
  • Finish wake-up Model
  • deliver ASR model and wake-up model
  • Finish draft of SPL paper....
Zhenghai You
  • Huawei project[2]
  • Minimal loss without cohort
Junming Yuan
  • Experimental report on "learn not to listen" v3[3]
Chen Chen
  • Finish thesis v1
  • group work [4]
  • refine thesis
Xiaolou Li
  • reproduce VSR structures
Zehua Liu
  • try many distillation ways[5]
Pengqi Li
  • summary[6] of speech processing XAI for NSFC
    • finished structure and almost content
  • polish, reference, tables
Wan Lin
  • Neural Scoring
    • Adjust the training and testing phase of mutil-speaker tasks
    • Design of the new architecture
  • Graduation paper
Tianhao Wang
  • EA-ASP (from SJTU) reproducing
    • data preparation
    • EA-ASP implement
Zhenyu Zhou
  • Baseline design[7]
Junhui Chen
  • reproduce frequency-channel attention for Neural Scoring
Jiaying Wang
  • Huawei Project
  • icassp 2024 paper selection
  • speakerbeam + ResNet [-10,4]
Yu Zhang
  • SAC training
  • R2-SAC adapted database build (2020 - 2023, benchmark and stocks)
Wenqiang Du
  • Try to find the reasons why the offline testing and online running of the model have different effects
    • almost fix [8]
Yang Wei
  • Train ASR model for children mispronunciation detection and diagnosis task.
Lily
  • Data Analysis
Turi
  • Text data collection
    • Pdf documents
    • News articles from official news website
  • Course work