Gradium open-sources Phonon: sub-50M on-device TTS models under Apache 2.0
thisdudelikesAI · x · 2026-10-21
Gradium open-sourced Phonon, a suite of TTS models under 50M parameters claiming a new standard for naturalness and multilingual accuracy, with broad inference support; Phonon-31m ships under Apache 2.0.
The pitch is on-device speech: generation runs locally so input text never leaves the device, avoiding per-request cloud fees. A French-language game demo showcases the quality.
Related event: Gradium Open-Sources Phonon: Sub-50M On-Device TTS with 1.7% WER(6 posts)→
More from Multimodal
- Full AI Virtual Influencer Character Sheet for 'JataJett': Turnarounds, Expressions and Action Modes in One Page — CurieuxExplorer · 2026-10-21
- AI-Generated Series 'The Sage' Hits Episode 8, Made Entirely with Invideo Agent — Kavanthekid · 2026-10-21
- Flova workflow turns one spooky idea into a consistent cinematic AI horror film — SimplyAnnisa · 2026-10-21
- Testing Nano Banana 2.1 for Fashion: Fabric Textures From Suede to Fur — aziz4ai · 2026-10-21
- Seedance 2.5 demo shows a lighthouse balancing against gravity, delightfully wrong physics — CurieuxExplorer · 2026-10-21
- AI design looks painfully same-y; Vyb generates layered editable designs from real brand assets — cneuralnetwork · 2026-10-21