whatmodel.sh / tts / EMA Lightning

EMA Lightning

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

EMA Lightning is an open-weights text-to-speech model from EMA Lightning (canberkkkkkk), released October 2026. Its published size is 0.0086B parameters. Licence: apache-2.0.

Catalogue identity notes

EMA Lightning Turkish TTS (5.6M-parameter DiT acoustic model + 3M vocoder; canberkkkkkk/ema-lightning, pinned 7a6ba1a, Apache-2.0; pip package ema-lightning). The Antalia re-run's harness/systems/run_ema.py runs the same Hugging Face release (card text: 'EMA Lightning 1.0.1').

Model information sources

Release date ↗ · Weights licence ↗

8.6M total (5.6M DiT + 3M vocoder) as stated on the card.

40.2±2.9#86 of 118 text-to-speech modelsfrom 2 benchmark results across 2 sources* thinly tested, treat the rank as provisional

Score within the Vapi Humanness · Bradley–Terry rating, 2026-09-25 snapshot comparison set only.

$ how it compares

The text-to-speech models ranked either side of EMA Lightning, out of 148 with a score.

#modelscorewhereRAMreleased
107OpenAudio S1 Mini Fish Audio41±5api price unknown +local~1.7GB2025-06
108Fish Speech 1.5 Fish Audio41±3local~1.5GB2024-12
109EMA Lightning EMA Lightning (canberkkkkkk)40±3local~0GB2026-10
110Antalia-2 Mini Patientdesk.ai40±4comparison only2026-10
111HierSpeech++ v1.1 Korea University (Sang-Hoon Lee et al.)39±3local~2.1GB2023-11

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

$ what it's good at

other tests1 not in a task · 97% of its score
testresultvs the others in its rundifficultycountswhy
Freya-TR-Eval Turkish TTS intelligibility (WER, faster-whisper large-v3, 8 kHz band) +11.0% ✎▄··▄··▄···▂█ beats 10/10avg 0.3%▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 97%several runs

$ where to run it

Other ways to run it

  1. Run it as published · needs ~0.04 GB · from huggingface.co/canberkkkkkk
    pip install ema-lightning && python -c "from ema_lightning import EMA; EMA().say('Merhaba, size nasıl yardımcı olabilirim?', path='merhaba.wav')"

Weights and documentation: huggingface.co/canberkkkkkk ↗

$ how the score was worked out · Antalia-2 Mini model card (cl… 53%, EMA Lightning model card (can… 43%

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
Antalia-2 Mini model card (cloud0day3/antalia-mini), public Freya-TR-Eval re-run table1▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 53%Apache-2.0
EMA Lightning model card (canberkkkkkk/ema-lightning), Benchmarks: Freya-TR-Eval · lab-reported1▇▇▇▇▇▇▇▇▇▇▇▇▇ 43%Apache-2.0
prior · pull toward the middle▇ 3%
weighted average= 40.2

$ 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