|
|
| 第193行： |
第193行： |
| | |Yang Wei | | |Yang Wei |
| | || | | || |
| − | * | + | * Text enroll KWS model training without CTC loss. |
| | || | | || |
| | * | | * |
| People |
This Week |
Next Week |
Task Tracking (DeadLine)
|
| Dong Wang
|
|
|
|
| Lantian Li
|
- Polish high-school AI handbook
|
|
|
| Ying Shi
|
- Analyze and summarize the previous experiments
- Project things
|
- continue on overlap asr task
- submit HUADANIAN challenge proposal
|
|
| Zhenghai You
|
|
|
|
| Junming Yuan
|
|
|
|
| Xiaolou Li
|
- Finish CN-AV-HuBERT code modify, already test on small data
- Simplified the post process of collected data
- CVS3 audit for December
|
- prepare training data for AVHuBERT, cp it to GA
- Some function demand on process server need to be solve
- Data Audit
|
|
| Zehua Liu
|
- Organize my previous Experiment Result[1]
- Add correct loss, but code encounters some bugs, so i fix it and train again(still training)
|
|
|
| Pengqi Li
|
- IS25 Proposal almost done(Double check and stat result)
|
|
|
| Wan Lin
|
- NS: multi-enroll+margin bce: 1.638% -> 1.521% (5w spk)
|
|
|
| Tianhao Wang
|
- Try using attention instead of FiLM:
- valid loss: query emb cross-attn: -8.065 vs. query feature cross-attn: -9.127 vs. FiLM: -10.054
- read a paper show that FiLM is better than attention in sound sep task
- Huawei project testing
|
|
|
| Xiaoxue Luo
|
- final exam
- wrote the remaining course papers
|
|
|
| Zhenyu Zhou
|
|
|
|
| Junhui Chen
|
- finished all course homework and final exams
|
|
|
| Jiaying Wang
|
|
|
|
| Yu Zhang
|
- Finish Memory mechanism and Policy fusion part, start run and do further debug
|
|
|
| Wenqiang Du
|
|
|
|
| Yang Wei
|
- Text enroll KWS model training without CTC loss.
|
|
|
| Turi
|
- Midterm defense
- Assessment exam for application
- Experiment on improving Oromo ASR[2]
|
|
|
| Yue Gu
|
- report my recent work
- read paper about functionally invariant path, which was reported by Prof. Wang.
|
|
|
| Qi Qu
|
- KWS:
- pre-prod routine work for zh48+yichang5 models.
- android demo backed with gRPC services deployed on peacock01.
- QAT: exploration with little progress.
|
|
|
