“2025-12-15”版本间的差异

来自cslt Wiki
跳转至： 导航、 搜索
 
（17位用户的19个中间修订版本未显示）
第28行： 第28行：
 
|Ying Shi
 
|Ying Shi
 
||
 
||
−
*  
+
* thesis
 +
* Multi-talker ASR Demo
 
||
 
||
 
*  
 
*  
第39行： 第40行：
 
|Zhenghai You
 
|Zhenghai You
 
||
 
||
−
*  
+
* Support work for Huawei SS project (2/3/4 zh mix)
 +
* Complete the training of base model and iterative model, and Hubert-base extractor is training
 
||
 
||
 
*
 
*
第50行： 第52行：
 
|Junming Yuan
 
|Junming Yuan
 
||
 
||
−
*  
+
* Current experimental progress of MT-HuBERT[https://z1et6d3xtb.feishu.cn/docx/AzyhdbCjyooVKOxLFV3cfY2MnuZ]
 +
* MT-HuBERT paper EN v0.4(4/7)
 
||
 
||
 
*
 
*
第61行： 第64行：
 
|Xiaolou Li
 
|Xiaolou Li
 
||
 
||
−
*  
+
* PhD Application
 
||
 
||
 
*  
 
*  
第72行： 第75行：
 
|Zehua Liu
 
|Zehua Liu
 
||
 
||
−
*
+
*Add Experiment on my graduation thesis
 +
*job hunting
 
||
 
||
 
*
 
*
第83行： 第87行：
 
|Pengqi Li
 
|Pengqi Li
 
||
 
||
−
*
+
* Transfer CAMs methods from the classification task to speaker verification task.
 
||
 
||
−
*
+
* Qualitatively and quantitatively measure the reliability of these methods. (FGGB, SSAB, Corr-RISE, SFAM, CAMs)
 
||
 
||
 
*
 
*
第94行： 第98行：
 
|Tianhao Wang
 
|Tianhao Wang
 
||
 
||
−
*  
+
* refine the ChainSep paper and adjust it to overleaf
 
||
 
||
 
*
 
*
第105行： 第109行：
 
|Xiaoxue Luo
 
|Xiaoxue Luo
 
||
 
||
−
*  
+
* Write code to combine SepFormer and EDA method, the model is in training
 +
* read papers
 
||
 
||
 
*
 
*
第116行： 第121行：
 
|Junhui Chen
 
|Junhui Chen
 
||
 
||
−
*
+
* HUAWEI gradio demo (4/5)
 +
* Integrate the current metric collection code into GPTSwarm (80%)
 
||
 
||
 
*
 
*
第127行： 第133行：
 
|Jiaying Wang
 
|Jiaying Wang
 
||
 
||
−
*
+
* content experiments test finish
 +
* loudness experiments testing
 +
* spk model code preparing
 
||
 
||
 
*
 
*
第138行： 第146行：
 
|Yu Zhang
 
|Yu Zhang
 
||
 
||
−
*  
+
* GPU Util: https://z1et6d3xtb.feishu.cn/wiki/XX4NwX3tJiBDcgkMi0hcFUtInHh?from=from_copylink
 +
* LLM: Latent learning simple reverse exp code (90%, lack for metric analysis code and Oracle Retrieval Code)
 
||
 
||
 
*
 
*
第149行： 第158行：
 
|Wenqiang Du
 
|Wenqiang Du
 
||
 
||
−
*  
+
* Rebuttal the reviewers' comments about high school AI course's draft(80%)
 +
* Calendar done
 
||
 
||
 
*
 
*
第160行： 第170行：
 
|Yang Wei
 
|Yang Wei
 
||
 
||
−
*  
+
* Prepare librispeech MD experiment git repository.
 +
* Reproduce MD experiment on TIMIT[https://z1et6d3xtb.feishu.cn/docx/Qmvld9Fw1omSrwxI14ccMHZVnYd?from=from_copylink].
 
||
 
||
 
*
 
*
第171行： 第182行：
 
|Yue Gu
 
|Yue Gu
 
||
 
||
−
*  
+
* write my Phd thesis
 
||
 
||
 
*
 
*
第182行： 第193行：
 
|Lily
 
|Lily
 
||
 
||
−
*  
+
* Check AI handbook (初二下册，高一下册)
 +
* AIGE Annual Forum（December 19th）
 
||
 
||
 
*
 
*
第193行： 第205行：
 
|Bochao Hu
 
|Bochao Hu
 
||
 
||
−
*  
+
* Phoneme to sentence
 +
** 0% error, WER: 2.1% | 5% error, WER: 11.0% | 10% error, WER: 32.4% | 20% error, WER: 66.0% | 30% error, WER: 89.3%
 +
* Video to Phoneme: Change the model architecture to baseline. First stage(<4s) is in training
 +
||
 +
*
 +
||
 +
*
 +
|-
 +
 
 +
 
 +
|-
 +
|Hongcheng Zhang
 +
||
 +
* construct code for ASU-LLM（10%）
 
||
 
||
 
*
 
*

2025年12月15日 (一) 11:39的最后版本

People This Week Next Week Task Tracking (DeadLine)
Dong Wang
Lantian Li
Ying Shi
  • thesis
  • Multi-talker ASR Demo
Zhenghai You
  • Support work for Huawei SS project (2/3/4 zh mix)
  • Complete the training of base model and iterative model, and Hubert-base extractor is training
Junming Yuan
  • Current experimental progress of MT-HuBERT[1]
  • MT-HuBERT paper EN v0.4(4/7)
Xiaolou Li
  • PhD Application
Zehua Liu
  • Add Experiment on my graduation thesis
  • job hunting
Pengqi Li
  • Transfer CAMs methods from the classification task to speaker verification task.
  • Qualitatively and quantitatively measure the reliability of these methods. (FGGB, SSAB, Corr-RISE, SFAM, CAMs)
Tianhao Wang
  • refine the ChainSep paper and adjust it to overleaf
Xiaoxue Luo
  • Write code to combine SepFormer and EDA method, the model is in training
  • read papers
Junhui Chen
  • HUAWEI gradio demo (4/5)
  • Integrate the current metric collection code into GPTSwarm (80%)
Jiaying Wang
  • content experiments test finish
  • loudness experiments testing
  • spk model code preparing
Yu Zhang
Wenqiang Du
  • Rebuttal the reviewers' comments about high school AI course's draft(80%)
  • Calendar done
Yang Wei
  • Prepare librispeech MD experiment git repository.
  • Reproduce MD experiment on TIMIT[2].
Yue Gu
  • write my Phd thesis
Lily
  • Check AI handbook (初二下册，高一下册)
  • AIGE Annual Forum（December 19th）
Bochao Hu
  • Phoneme to sentence
    • 0% error, WER: 2.1% | 5% error, WER: 11.0% | 10% error, WER: 32.4% | 20% error, WER: 66.0% | 30% error, WER: 89.3%
  • Video to Phoneme: Change the model architecture to baseline. First stage(<4s) is in training
Hongcheng Zhang
  • construct code for ASU-LLM（10%）
取自“http://www.cslt.org/mediawiki/index.php?title=2025-12-15&oldid=43707”