in Coding on Pytorch
librosa.filters.mel는 기본 옵션으로 htk=False, norm="slaney"로 설정되어있다.torchaudio.transforms.MelScale는 기본 옵션으로 norm=None, mel_scale="htk"로 설정되어있다.torchaudio의 옵션을 norm="slaney", mel_scale="slaney"로 바꿔주면 librosa 전처리와 같은 Mel-spectrogram을 만들 수 있다.in Seminar on Text-to-Speech
CLAPSpeech 논문 요약
Zhenhui Ye, Rongjie Huang, Yi Ren, Ziyue Jiang, Jinglin Liu, Jinzheng He, Xiang Yin, Zhou Zhao
Accepted by ACL2023 (Main Conference)
[Paper][Demo]
in Seminar on Text-to-Speech
P-Flow 논문 요약
Sungwon Kim, Kevin J Shih, Rohan Badlani, Joao Felipe Santos, Evelina Bhakturina, Mikyas Dest, Rafael Valle, Sungroh Yoon, Bryan Catanzaro
“P-Flow: A Fast and Data-Efficient Zero-Shot TTS through Speech Prompting”
Accepted by NeurIPS2023
[Paper][Demo]
in Post on Post
wakeonlan 설치sudo apt-get install wakeonlan
in Seminar on Text-to-Speech
Matcha-TTS 논문 요약