1.8GB open TTS model Tontaube V1.1 clones voices and narrates audiobooks on CPU alone
solyarisoftware · x · 2026-11-16
Tontaube V1.1 is an open-weights 2.9B TTS model built on Qwen3; its 4-bit GPTQ build is just 1.8GB and can clone voices and narrate full audiobooks on a Ryzen CPU, no GPU needed.
Key details:
- Voice cloning from 5–60s of audio; control over rate, pitch, pauses, intonation
- Ryzen 9 5900X: 1x real time; RTX 3090: 6–7x; RTX 5090: 13x single request, 21–23x batched
- Loads in 4GB VRAM (peak 6.5GB); ships with its own inference server and AVX2 CPU kernels; Linux/Windows via WSL2, community license
More from Multimodal
- Meta & NYU paper: vision prior unnecessary, dedicated capacity placed early is what matters — burny_tech · 2026-11-16
- Fable 5.2 renders photorealistic 3D scene: 1M tokens, 5hrs render — Dr_Singularity · 2026-11-16
- Fable 5.2 turns a Starship launch into an entirely JavaScript painting — brandon_galang · 2026-11-16
- Sony unveils Walrus, its most capable SpeechLLM with SOTA recognition, translation and understanding — mittu1204 · 2026-11-16
- Fable 5.2 tries mosaic art style, lands a bridge with 14 unique spans — repligate · 2026-11-16
- An endless cinematic loop: Arabian warrior and a jinn made with Minimax H3 and ACE-Step — magratheya_64 · 2026-11-16