NVIDIA open-sources Nemotron 3 Diarization to fix voice agents' speaker-blindness
andimarafioti · x · 2026-09-23
Voice agents still can't tell who's speaking to them — a gap hidden by phone-call demos. NVIDIA just open-sourced Nemotron 3 Diarization, a model that reliably tracks speakers in live conversations under a commercial-friendly license.
- Tested with one-second speech chunks; quality is solid enough for voice agents
- Demoed on a Reachy Mini robot with speech-to-speech on DGX Spark: it notices a new voice, asks for a name, and remembers it
- Day-zero integration with Transformers
Related event: NVIDIA Open-Sources Nemotron 3 Diarization, Tops Benchmark(16 posts)→
More from Embodied
- Cube-Solving Robot Demo Is Its Best Case: ~90 Moves per Solve on Average — haydendevs · 2026-11-11
- ETH researcher vibe-codes SuperMimic humanoid reproduction in 1h, LAFAN passes in 1.5 GPU-hours — ChongZzZhang · 2026-11-11
- Open-source companion robot Lamp ships at $699 with custom personality, skills and an app economy — dee_hw · 2026-11-11
- Must-read papers of the week: HALO humanoid motion, GR00T 2, VLA policies, agent tooling — TheTuringPost · 2026-11-11
- Microsoft's DLSS rival Auto SR launches on Xbox Ally X with handheld 1080p and 120fps support — tomwarren · 2026-11-11
- NVIDIA's PhysicalAI-Human-1M-100K robot dataset trends on Hugging Face — nvidia · 2026-11-11