# Nvidia 发布开源语音分离模型 Nemotron 3 Diarization，实时区分最多八位说话人

- 来源：The Decoder：AI News（RSS）
- 作者：Jonathan Kemper
- 发布时间：2026-09-27 19:01
- AIHOT 分数：56
- AIHOT 链接：https://aihot.news/items/cmujqnapk0q66ro9hpzs5l6vy
- 原文链接：https://the-decoder.com/nvidia-drops-a-free-100m-parameter-model-that-identifies-up-to-eight-speakers-in-real-time

## AI 摘要

Nvidia 发布约 100M 参数的 Nemotron 3 Diarization 模型，权重免费开放，可实时区分最多八位说话人并检测重叠语音，支持录音和实时音频。

## 正文

Nvidia released Nemotron 3 Diarization, an AI model that identifies which speaker is talking at any given moment in a conversation. The model has about 100 million parameters, and its weights are freely available. It can tell apart up to eight speakers and detect when multiple people talk at the same time. More participants, heavy background noise, or reverb push error rates higher. Paired with a speech recognition system like Parakeet, the model can produce transcripts with speaker labels, though only anonymous ones like "speaker_2." It works with both recordings and live audio.

In the VoiceArena Diarization Benchmark v1, the freely available Nemotron 3 leads with a DER of 14.7% and outperforms its predecessor, Streaming Sortformer, by 41%.

The audio buffer can be set to four levels ranging from 30.4 down to 0.32 seconds. Shorter buffers generally reduce accuracy. On the Diarization-Bench from VoiceArena, the model currently sits in first place with a 14.72 percent error rate, ahead of the next best system at 19.3 percent. The benchmark is strict. Overlapping speech counts, and even tiny misalignments at speaker transitions are scored as errors. Compared to its predecessor, Streaming Sortformer, the new model cuts the error rate by an average of 41 percent across eight test scenarios when using a 1.04-second buffer.