Inflect v3 released: 34M-parameter local TTS with six voices and event tags
b111ue · reddit · 2026-11-12
The author released Inflect v3, a local TTS lineup led by Mini (34.31M parameters) with six English voices, eight vocal-event tags ([laugh], [sigh], [gasp], etc.), reference-voice tuning that turns a recording into a reusable profile (approximation, not cloning), native 24 kHz audio with optional 48 kHz enhancement, running on CPU or CUDA. Smaller variants include Micro (11.79M) and Atom (157K). Weights and inference code are Apache 2.0, available on GitHub and Hugging Face.
Related event: Inflect v3: Open-Source 34M-Parameter Local TTS Models(2 posts)→
More from Multimodal
- AssemblyAI launches Vet Mode transcription and first TTS model for voice agents — AssemblyAI · 2026-11-12
- Fable 5.2 autonomously records, directs, edits and scores a full video with near-zero human input — henloitsjoyce · 2026-11-12
- MiniMax H3-Image tested: 4MP images in 140s on Mac, but detail trails QwenImage — karminski3 · 2026-11-12
- MOSS-Image workflow: emulating real camera lens, focus and exposure for cinematic shots — IndependentBid6893 · 2026-11-12
- Game dev workflow: AI-generated sprite sheets turned into character walk GIFs — algo_diver · 2026-11-12
- Jensen Buddy AI goes semi-viral: fake selfies with Nvidia CEO Jensen Huang — omooretweets · 2026-11-12