Inflect-Micro-v2: complete voice in 9.36M parameters

Inflect-Micro-v2

Complete local text-to-waveform speech synthesis under 10M parameters.

Fixed-voice English TTS with deterministic seeds, long-text handling, and CPU or CUDA inference.

Overview

Inflect v2 is an independently developed and funded text-to-speech project. It uses one public API across two sizes: Micro prioritizes quality below 10M parameters; Nano prioritizes footprint below 4M.

• 9,356,513 deployable parameters
• 37.53 MB FP32 weight size
• 24 kHz mono output

Evaluation Metrics

No single metric captures TTS quality. Inflect v2 reports human preference, predicted naturalness, multi-ASR intelligibility, complete footprint, and runtime separately:

• Community preference: 66.2% (21 wins, 10 losses, 3 ties)
• UTMOS22 naturalness score: 4.395 (95% CI: 4.381–4.408)
• Two-ASR semantic WER: 3.99% (Qwen3-ASR: 2.52%, Nemotron 3.5: 5.45%, Whisper large-v3: 2.73%)
• Throughput (4-thread CPU): 6.28× real-time (Steady-state RTF: 0.1593)

Family Comparison

| Feature | Inflect-Nano-v2 | Inflect-Micro-v2 | |---|---|---| | Complete parameters | 3,966,721 | 9,356,513 | | FP32 weights | 15.97 MB | 37.53 MB | | Positioning | Smallest practical footprint | Strongest Inflect v2 quality | | Throughput (4-thread CPU) | 10.72× real-time | 6.28× real-time | | 24 kHz waveform decoder | Included | Included |

Quickstart

python -m pip install --upgrade huggingface_hub
hf download owensong/Inflect-Micro-v2 --local-dir Inflect-Micro-v2
cd Inflect-Micro-v2
python -m pip install -r requirements.txt

from inference import InflectTTS

tts = InflectTTS(".", device="cpu")
tts.save(
    "A small voice can still have something meaningful to say.",
    "sample.wav",
    speed=1.0,
    variation=0.667,
    seed=7,
)