Inflect-Micro-v2: complete voice in 9.36M parameters
huggingface.co
July 26, 2026
11 min read
58/100
Summary
Inflect-Micro-v2 is a text-to-speech synthesis model with under 10 million parameters, offering fixed-voice English TTS, deterministic seeds, long-text handling, and support for CPU or CUDA inference. The developer aims to expand the project to include additional languages, voices, and stability improvements if it gains traction.
Key Takeaways
- Inflect-Micro-v2 is a text-to-speech model that synthesizes speech with under 10 million parameters and offers long-text handling with deterministic outputs.
- The model achieved a 66.2% preference rate in a community study, indicating a favorable reception compared to other TTS systems.
- Inflect-Micro-v2 reports a UTMOS22 score of 4.395, reflecting its predicted naturalness while maintaining a compact footprint of 37.53 MB.
- The model demonstrates low word error rates (2.52% for Qwen3-ASR and 5.45% for Nemotron 3.5), indicating high intelligibility on unseen text.
Coverage & discussions
Hacker News · huggingface.co
Community discussions
Historical count: 4 · 4 saved in sample · Historical sample
Awaiting analysis
Related Articles
Pure C, CPU-only inference with Mistral Voxtral Realtime 4B speech to text model
Feb 10, 2026
Nvidia PersonaPlex 7B on Apple Silicon: Full-Duplex Speech-to-Speech in Swift
Mar 5, 2026
How we made a text-to-speech model respond in sub-50 ms
Aug 21, 2026
Microsoft VibeVoice: Open-Source Frontier Voice AI
Apr 28, 2026
Qwen/Qwen3.8-2.4T-A95B
Aug 12, 2026