Instructions to use nvidia/Nemotron-3-Diarization with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- NeMo
How to use nvidia/Nemotron-3-Diarization with NeMo:
# tag did not correspond to a valid NeMo domain.
- Transformers
How to use nvidia/Nemotron-3-Diarization with Transformers:
# Load model directly from transformers import AutoProcessor, AutoModelForAudioFrameClassification processor = AutoProcessor.from_pretrained("nvidia/Nemotron-3-Diarization") model = AutoModelForAudioFrameClassification.from_pretrained("nvidia/Nemotron-3-Diarization", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Add Transformers weights, config and usage section
Adds the π€ Transformers weights and configuration for this checkpoint, converted from Nemotron-3-Diarization.nemo with the conversion script in the Transformers PR, plus a Transformers usage section in README.md (offline and streaming diarization).
The auxiliary activity head (only used by the training loss) is not converted. Weights are stored in float32, matching NeMo's inference precision.
Draft: Transformers support for this architecture is not merged yet, so the weight names and the configuration schema may still change during review. The snippets install Transformers from source for that reason. I will refresh this PR if review changes the layout.