WTF Is Going On(Weights, Tools & Frameworks)

한국어
microsoft/VibeVoice-ASR-Streaming-1.5B
automatic-speech-recognition mit

What this is

2.8B parameters. Turns recorded speech into text.

mit — commercial use allowed.

Release chainonly the steps this archive can prove

  1. weights posted2026-09-02

Measured herefrom our own daily capture — upstream publishes today only

← the board