“2021-11-15”版本间的差异

来自cslt Wiki
跳转至: 导航搜索
第95行: 第95行:
 
|Weida Liang
 
|Weida Liang
 
||   
 
||   
*  
+
* Renew the project website and upload relevant speech sequences
 +
* Draw various loss curves
 +
* Prepare test data for content-relevant tests
 
||
 
||
*  
+
* Quantitave test
 +
* Prepare test data for speaker separation tests
 
||
 
||
 
*   
 
*   
|-
+
|-  
  
 
|-
 
|-

2021年11月15日 (一) 11:10的版本

People This Week Next Week Task Tracking (DeadLine)
Dong Wang
  • Spoof paper almost done
  • Spoof paper cleaning
  • Hard trials paper
Yunqi Cai
  • Completed THS2021 data preprocessing, image feature extraction, and baseline
Lantian Li
  • Complete hard trials paper v1.
  • Go on preparing my defence.
Ying Shi
  • make a comparison between fncmd with new methods
  • find some cross-modality methods(Cross modality attention)
  • Implement Cross modality attention
Haoran Sun
  • speaker and robustness test for CycleFlow-spk
  • some improvement
  • some other exploration for the model
Chen Chen
Pengqi Li
Qingyang Zhu
Weida Liang
  • Renew the project website and upload relevant speech sequences
  • Draw various loss curves
  • Prepare test data for content-relevant tests
  • Quantitave test
  • Prepare test data for speaker separation tests
Zixi Yan
  • Modify wav2vec-u gan network code and output intermediate results
  • Training wav2vec2 model
Sirui Li
  • Compare the training process of TIMIT and Tibetan
  • Fine-tune the Tibetan wav2vec model
  • Prepare the thesis opening report
Haoyu Jiang
Ruihai Hou
Renmiao Chen