whatmodel.sh / tts / FastSpeech + HiFi-GAN (xcmyz, Grad-TTS study)

FastSpeech + HiFi-GAN (xcmyz, Grad-TTS study)

+ contribute
Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

FastSpeech + HiFi-GAN (xcmyz, Grad-TTS study) is a closed-weights text-to-speech model from Grad-TTS authors / Huawei Noah, released May 2021.

Catalogue identity notes

Unofficial xcmyz FastSpeech implementation used by the Grad-TTS authors, coupled with HiFi-GAN; not the original Microsoft FastSpeech checkpoint. Exact trained weights not verified.

26.4±20.4comparison only: no current run option verifiedfrom 1 benchmark result across 1 source* thinly tested, treat the rank as provisional

Score within the Grad-TTS 2021 LJ Speech human MOS comparison set only.

$ how it compares

Scored text-to-speech models released closest to FastSpeech + HiFi-GAN (xcmyz, Grad-TTS study), out of 148 with a score.

#modelscorewhereRAMreleased
81Azure Ameha Neural Microsoft50±16comparison only2021-10
82Azure Mekdes Neural Microsoft50±16comparison only2021-10
124VITS (VCTK multi-speaker) Kakao Enterprise (Jaehyeon Kim et al.)34±5local~0.2GB2021-06
135Glow-TTS + HiFi-GAN (VCTK, VITS study) VITS study (public implementations)28±6comparison only2021-06
142Tacotron 2 + HiFi-GAN (VCTK, VITS study) VITS study (public implementations)19±6comparison only2021-06

see all text-to-speech models · find the ones that fit your machine

$ what it's good at

other tests1 not in a task · 66% of its score
testresultvs the others in its rundifficultycountswhy
Grad-TTS 2021 LJ Speech human MOS3.68 / 5 ✎█·····█···██ beats 0/3easy · avg 4.17 / 5▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 66%at its level

$ where to run it

No verified API or download recorded yet · add one

Weights and documentation: github.com/xcmyz ↗

$ how the score was worked out · Grad-TTS: original ICML 2021… 66%

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

sourceresultsshare of the scorelicence
Grad-TTS: original ICML 2021 listening study1▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 66%CC-BY-4.0
prior · pull toward the middle▇▇▇▇▇▇▇▇▇▇ 34%
weighted average= 26.4

$ improve this page

Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated · sources and licences · json