일 | 월 | 화 | 수 | 목 | 금 | 토 |
---|---|---|---|---|---|---|
1 | ||||||
2 | 3 | 4 | 5 | 6 | 7 | 8 |
9 | 10 | 11 | 12 | 13 | 14 | 15 |
16 | 17 | 18 | 19 | 20 | 21 | 22 |
23 | 24 | 25 | 26 | 27 | 28 | 29 |
30 | 31 |
- 캐나다월세
- Docker
- DTW
- Normalizing flow
- Flow
- 캐나다 TD 한국인 예약
- 캐나다 은행계좌 개설
- 캐나다콘도렌트
- 딥러닝
- agnoster폰트꺠짐
- Generative model
- 캐나다 TD 한국인 직원 예약
- iterm2꾸미기
- pytorch
- GenerativeModel
- 콘도렌트
- 터미널꾸미기
- 캐나다 TD 예약
- 프라이탁
- 캐나다 TDBAnk
- EATS
- iterm2환경설정
- 프라이탁 존버
- agnoster폰트깨짐
- iterm2자동완성
- MachineLearning
- agnoster폰트
- GAN
- 캐나다 은행 계좌 개설
- 머신러닝
- Today
- Total
목록머신러닝 (2)
TechNOTE

요약 conditional information을 GAN의 discriminator에 합치는 것을 projection based 로 하는 것을 제안했다. 기존에 input conditional information을 concat 하는 방식과는 다르고, 이 방법으로 class conditional image generation task에서 SOTA 달성했다고 한다. https://github.com/pfnet-research/sngan_projection pfnet-research/sngan_projection GANs with spectral normalization and projection discriminator - pfnet-research/sngan_projection github.com 1. ..

딥마인드에서 음성 합성 관련해서 새 논문이 나왔다! 바로 END-TO-END ADVERSARIAL TEXT-TO-SPEECH 인데.. 사실 End to End 라는 말보다 (말의 정의가 너무 애매하다) 1 stage speech synthsis 라고 하는게 더 맞지 않나 싶다. 여튼 이번 뉴립스에는 떨어지고 ICLR 에 다시 낸 것 같다. deepmind.com/research/publications/End-to-End-Adversarial-Text-to-Speech End-to-End Adversarial Text-to-Speech Modern text-to-speech synthesis pipelines typically involve multiple processing stages, each of ..