“2023-10-23”版本间的差异

来自cslt Wiki
跳转至： 导航、 搜索
 
第137行： 第137行：
 
|Junhui Chen
 
|Junhui Chen
 
||
 
||
−
* overlap ASV pipline: loss decreasing normally on vox1-overlap
+
* overlap ASV pipline: loss decreasing normally on vox1-overlap[https://z1et6d3xtb.feishu.cn/docx/HLtwdr7OYoHWoJxv5qAc8mhlnLc?from=from_copylink]
 
||
 
||
 
* training on vox2-overlap
 
* training on vox2-overlap

2023年10月23日 (一) 11:10的最后版本

People This Week Next Week Task Tracking (DeadLine)
Dong Wang
  • AI-book experimental design, done.
  • Presentation for voiceprint seminar
  • Lin Wan's paper
  • SIP doc, chatp 5, part.
Lantian Li
  • GPU status [1]
  • APT-group (OverlapASV, Ob-Code)
  • DeepTrader (TG-SAC)
  • Beijing award materials
  • ICCIP and NC papers
Ying Shi
  • speaker attributed speech separation & keyword attributed speech separation
  • code for ZEUS
Zhenghai You
  • SID model [2]
  • AIdemo
Chen Chen
  • refine visual-speech data collection toolkit
  • cnvsrc evaluation toolkit for submission system
  • prepare data collection website and database
Xiaolou Li
Zehua Liu
Pengqi Li
  • Phone's discrimination in deep SID [5]
  • concat silence[6]
Wan Lin
  • NC paper
Tianhao Wang
  • ICCIP paper
Zhenyu Zhou
  • ICCIP paper
Junhui Chen
  • overlap ASV pipline: loss decreasing normally on vox1-overlap[7]
  • training on vox2-overlap
Jiaying Wang
  • Illustration of obcode[8]
Yu Zhang
  • AI-Demo[9]
  • Illustration of obcode
Wenqiang Du
Yang Wei
  • Some test on audio-visual speech separation model.
取自“http://index.cslt.org/mediawiki/index.php?title=2023-10-23&oldid=40844”