# Transformers 5.18.0 adds streaming speaker diarization model

> Hugging Face’s new release includes Nemotron 3 Diarization, a streaming model that can label up to eight speakers in live audio.

Oossa · 2026-09-30 · https://oossa.com/en/transformers-5-18-0-adds-streaming-speaker-diarization-model

Produced and translated with AI assistance. Check the original sources below.

Hugging Face released Transformers v5.18.0. The update adds Nemotron 3 Diarization, an open‑weight model that can tell “who spoke when” in real‑time or offline audio. It works with up to eight speakers and lets developers choose latency from 80 ms to 30.4 seconds.

The release also brings multimodal models NemotronH Omni and HyperCLOVAX Vision V2, plus a range of bug fixes and performance tweaks.

## The facts

- Version 5.18.0 released on 2026‑09‑30
- Nemotron 3 Diarization supports up to 8 speakers and latency from 80 ms to 30.4 s

## Why it matters

Developers can now add live speaker‑labelled transcription to apps, making meetings and podcasts easier to follow.

## Sources & references

1. [huggingface/transformers Release 5.18.0](https://github.com/huggingface/transformers/releases/tag/v5.18.0) – Transformers, 2026-09-30

Last updated: 2026-09-30