본문 바로가기 메뉴 바로가기

Let IT Begin

프로필사진
  • 글쓰기
  • 관리
  • 태그
  • 방명록
  • RSS

Let IT Begin

검색하기 폼
  • 분류 전체보기 (566) N
    • Music (17)
      • Review (11)
      • Guide (6)
    • 결산 (38)
      • 2020's (8)
      • 2010's (10)
      • 2000's (10)
      • 1990's (10)
    • Paper (477) N
      • TTS (166)
      • SVS (24)
      • ASR (10)
      • Vocoder (69)
      • Conversion (58)
      • Neural Codec (39)
      • Language Model (30) N
      • Representation (44) N
      • Verification (6)
      • Separation (6)
      • Signal Processing (6)
      • NAS (12)
      • ETC (7)
    • Algorithm (34)
      • Basic (34)
  • 방명록

FastSpeech2 (1)
반응형
[Paper 리뷰] FastSpeech2: Fast and High-Quality End-to-End Text to Speech

FastSpeech2: Fast and High-Quality End-to-End Text to Speech FastSpeech와 같은 non-autoregressive Text-to-Speech (TTS) 모델은 빠르게 음성합성이 가능함 FastSpeech는 duration prediction과 knowledge distillation을 위해 autoregressive teacher 모델에 의존적임 Teacher-student distillation 과정이 복잡하고 시간 소모적임 Teacher 모델에서 추출한 duration이 정확하지 않고 target mel-spectrogram의 단순함으로 인해 정보 손실이 발생함 FastSpeech 2 Teacher의 단순화된 output 대신 ground-tru..

Paper/TTS 2023. 7. 21. 12:23
반응형
이전 1 다음
이전 다음
최근에 올라온 글
최근에 달린 댓글
«   2025/11   »
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
Total
Today
Yesterday

Blog is powered by Tistory / Designed by Tistory

티스토리툴바