Hugging Face Blog·· 3 h ago
NVIDIA Nemotron 3.5 Transcribe and Granary v2: Intelligent Multilingual Transcription You Can Prompt and Adapt
NVIDIA Nemotron 3.5 Transcribe and Granary v2: Intelligent Multilingual Transcription You Can Prompt and Adapt
AI summary
One open model for multilingual transcription, speaker attribution, translation, and audio Q&A, with top leaderboard results, plus the open Granary v2 data pipeline and fine-tuning recipes to adapt it.
Selection record
Threshold 60Official, first-handFirst 78Second 78
AdmittedSum of both 156 ≥ twice the threshold 120
- Source tier
- Official, first-hand; this tier's threshold is 60
- Pre-filter
- passed：NVIDIA开源语音模型及数据管线，实质AI技术
- Why it was chosen
- Covers a single set of weights for multilingual transcription, speaker attribution and translation, plus an open data pipeline and fine-tuning recipes, which shows how far speech workflows could be reworked.
- Same event
A model scores each item twice, independently, against one written standard, out of 100. An item is admitted only when the two scores add up to twice the threshold. The threshold is set per source tier.
Source: Hugging Face Blog · huggingface.co