Kim, Kong, Son, Yoon (Kakao) · 2020 · normalizing flow
Parallel TTS with a normalizing flow and monotonic alignment search — the same Kakao lab that later shipped VITS. Fast, diverse, no attention failures. The clip is the authors’ LJSpeech demo (temperature 0.667) next to Tacotron 2 on that page.
The flow model VITS grew out of.
Vendor / paper demo
Glow-TTS (LJSpeech, T=0.667) · Official Glow-TTS project-page LJSpeech sample lj_01_ours_6.wav (WAV converted to MP3).
But Mrs. Solomons could not resist the temptation to dabble in stolen goods, and she was found shipping watches of the wrong category to New York.