반응형
[Paper 리뷰] OZSpeech: One-Step Zero-Shot Speech Synthesis with Learned-Prior-Conditioned Flow Matching
OZSpeech: One-Step Zero-Shot Speech Synthesis with Learned-Prior-Conditioned Flow MatchingWaveform, spectrogram과 같은 기존의 speech representation은 speech attribute를 overlooking 하고 high computational cost를 가짐OZSpeechOne-step sampling과 learned prior를 condition으로 사용하여 sampling step 수를 reduceToken format의 disentangled, factorized component를 활용하여 speech attributre를 modeling논문 (ACL 2025) : Paper Link1. In..
Paper/TTS
2025. 6. 30. 17:03
반응형