igunosalabs/Unose-Voice-v1.0

Unose Voice v1.0 (Production Release)

Unose Voice v1.0 is an edge-optimized neural text-to-speech acoustic model developed by Igunosa Labs as part of the Unose speech suite.

Model Details

• Architecture: VITS (Variational Inference with adversarial learning for end-to-end Text-to-Speech)
• Base Acoustic Checkpoint: Native male en_US-ryan-medium adaptation (500 epochs)
• Graph Footprint: 60.5 MB standalone ONNX computational graph
• Inference Latency: ~50ms on commodity CPU / ARM edge hardware

Ecosystem Links

• Official GitHub Codebase: https://github.com/Igunosa-Labs/unose-voice
• STT Suite (Base 74M): igunosalabs/Unose-Huen-Base
• STT Suite (Small 244M): igunosalabs/Unose-Huen-Small

Inference (Piper CLI)

echo "Patrick, this is the final release of Unose Voice version 1.0." |   piper -m Unose-Voice-v1.0.onnx -f output.wav

License & Attribution

Developed by Igunosa Labs. Released under the MIT License.