728x90
반응형

참조한 자료들

 

영어자료(외국)

https://www.youtube.com/watch?v=hS42xD3O55E Unsupervised Speech Recognition(Wav2Vec-U) 2021.5.23

https://www.youtube.com/watch?v=XkUVOijzAt8 Wav2Vec: Unsupervised pre-training for speech recognition, 2019.7.6

https://www.youtube.com/watch?v=aUSXvoWfy3w wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations. 2020.06.25

https://www.youtube.com/watch?v=C4UQWJcp7w4 [CVPR 2020 Tutorial] Talk #5 Self-supervised Learning by Licheng Yu, Yen-Chun Chen and Linjie Li, 2020.06.17

https://github.com/ShigekiKarita/espnet-semi-supervised , 2018-2019, InterSpeech 2018, PyTorch...

https://www.youtube.com/watch?v=8Kpowre6yyk wav2vec 2.0 | Lecture 76 (Part 3) | Applied Deep Learning, 2021.05.07

 

국내자료(한국)

https://www.youtube.com/watch?v=Z1lSukzyA0E [Paper Review] Semi-Supervised Learning in Auto Speech Recognition, KU, 2021.07.07

http://dsba.korea.ac.kr/review/?mod=document&uid=1408 2020 NIPS 후기, DSBA, Korea University. 2020.12.31

https://github.com/kakaobrain/pororo/issues/54 kakaobrain, wav2vec 2.0 한국어 실험환경....

https://ichi.pro/ko/eumhyang-deiteo-sajeon-gyoyuggwa-gat-eun-eon-eo-model-270503402471882 음향데이터 사전 교육과 같은 언어모델

 

LM Pretained Model == Contextual Word-Embeddeing...

Word-Embedding :: 유사한 의미를 가진 단어가 유사한 표현을 갖도록하는 일종의 표현,

비지도 / 자기 감독 표현 학습

 

 

Wav2Vec 활용분야 :: 음성인식, 오디오 분할, 이상한 음향 감지

 

Fairseq에는 wav2vec, vq-wav2vec, wav2vec 2.0의 예시적인 구현이 있습니다.

 

728x90
반응형

+ Recent posts