README.md

metadata

title: EMA Lightning
emoji: ⚡
colorFrom: blue
colorTo: indigo
sdk: gradio
sdk_version: 6.29.1
app_file: app.py
short_description: Turkish TTS from an 8.6M-param flow-matching model
python_version: '3.12'
startup_duration_timeout: 30m

Turkish text-to-speech with EMA Lightning — a 5.6M-parameter DiT acoustic model plus a 3M-parameter vocoder: 8.6M parameters in total, about 34 MB, Apache-2.0. It reads Turkish characters directly, expands digits, dates, times and currency to spoken Turkish, and synthesizes 48 kHz speech in four ODE steps. 0.92% WER on Freya-TR-Eval, the lowest of every system measured, including ElevenLabs v4, Gemini 3.8 and the 2.38B-parameter Trendyol-TTS.

The Space runs the authors' own ema-lightning pip package against the checkpoint at canberkkkkkk/ema-lightning. One fixed voice, no cloning. Example sentences are the authors' own showcase texts from the model card.

• Model card
• GitHub
• Text frontend: normalizer-tr