AI BriefingNVIDIA Releases NemotronLabs RE-USE, a 9.6M-Parameter Multilingual Speech Enhancement Model
·2026.11.03 18:44
Key point
The model supports input sampling rates from 8 to 48 kHz and is designed to restore speech quality while preserving fidelity across diverse degradations.
Details
- NVIDIA released NemotronLabs RE-USE, a 9.6-million-parameter multilingual universal speech enhancement model.
- The model restores speech quality across real-world degradations including additive noise, reverberation, clipping, bandwidth limitation, codec artifacts, packet loss, and low-quality mics.
- It supports multiple input sampling rates: 8, 16, 22.05, 24, 32, 44.1, and 48 kHz.
- The model is language-agnostic and ready for commercial use.
- A separate variant, NVIDIA-NemotronLabs-RE-USE-laughter, is available for better laughter preservation.
- For overlapping speech, the model is trained to preserve the main speaker while suppressing others.
- A Gradio Interactive Demo is available for direct testing.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.