Nvidia drops a free 100M-parameter model that identifies up to eight speakers in real time

Technology · en

Nvidia drops a free 100M-parameter model that identifies up to eight speakers in real time

The Decoder · Sep 27, 2026, 11:01 AM UTC

Nvidia released Nemotron 3 Diarization, an AI model that identifies which speaker is talking at any given moment in a conversation. The article Nvidia drops a free 100M-parameter model that identifies up to eight speakers in real time appeared first on The Decoder .

Read original report Visit The Decoder

Technology · WIRED

Best Party Speakers (2026): JBL, Sony, Marshall, and More

These speakers combine serious volume, big bass, portability, and the occasional karaoke session to bring the party anywhere.Read original ↗

Technology · Engadget

Why we won't know how visible the iPhone Duo's crease is for a long time

The iPhone Duo has an impressively minimal crease, but that's only how it looks at the start. We'll have to wait a while to see what develops over time.Read original ↗

Technology · Engadget

How high is too high for speakers in your home sound system?

Too much height makes dialogue and fine details sound less direct.Read original ↗

Technology · Engadget

Garmin Fenix 9 vs Fenix 8: How much of an upgrade is the new model?

Which sport smartwatch should you get?Read original ↗

Technology · Engadget

These Xbox players got GTA 6 for free the hard way

GTA VI could be yours for free, or at least for a big discount, using a few tricks that Xbox offers.Read original ↗

Technology · WIRED

Thieves Stole ‘Nvidia’ Trailers. They Got 20 Tons of Sand

They wanted the silicon. They got the sand.Read original ↗