nemotron-3-diarization-ov is an open-weight model from Shameez, released under other. Its published files total 398.7 MB.
This is an OpenVINO IR conversion of NVIDIA's nvidia/Nemotron-3-Diarization speaker diarization model ("who spoke when").
Model Card
This is an OpenVINO IR conversion of NVIDIA's nvidia/Nemotron-3-Diarization speaker diarization model ("who spoke when"). It was made for Scrybe, a local Windows transcription app, to add speaker labels ("Speaker 1", "Speaker 2",...) to transcripts on Intel CPUs and GPUs. Not affiliated with or endorsed by NVIDIA. This is an independent conversion. For the original model, its training data, evaluation and responsible-use information, see NVIDIA's model card. The IR holds one offline processing step of the model: the feature embedder, the 31-layer encoder, the projection, the upsampler and the speaker classifier. It does not include the audio front end or the speaker-cache logic. The calling…
Excerpt from the card by Shameez, licensed other.
Configuration
- Architecture
- Nemotron3DiarizationForAudioFrameClassification
- Model type
- nemotron3_diarization
Identity and Version
- Repository
- shameez/nemotron-3-diarization-ov
- Publisher
- Shameez
- Task
- Not stated by the source
- Modality
- Other
- Library
- openvino
- Parameters
- Not stated by the source
- Languages
- Not stated by the source
- Revision
- f07820236f05c02b0a83d298cfac67e6b520cb6b
- First published
- 2026-10-09
- Last updated
- 2026-10-09
Files and Weights
25 files, 398.7 MB in total. The weights are 1 file totalling 396.9 MB in bin.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| nemotron3_diar_step.bin | Weights | 396.9 MB | 56aca6be06d3 |
| config.json | Configuration | 1.5 KB | — |
| convert/e1_reference.py | Configuration | 7.8 KB | — |
| convert/e1_rescore.py | Configuration | 6.5 KB | — |
| convert/e2_make_inputs.py | Configuration | 5.9 KB | — |
| convert/t2_parity.py | Configuration | 7.2 KB | — |
| convert/t2_step_wrapper.py | Configuration | 7.4 KB | — |
| convert/t3_speed.py | Configuration | 20.2 KB | — |
| convert/t4_a0_clips.json | Configuration | 2.4 KB | — |
| convert/t4_a0_padding_check.py | Configuration | 8.3 KB | — |
| convert/t4_convert.py | Configuration | 12.9 KB | — |
| convert/t4_make_clips.py | Configuration | 2.7 KB | — |
| convert/t4_parity.py | Configuration | 9.0 KB | — |
| convert/t4_step_wrapper.py | Configuration | 10.2 KB | — |
| convert/t4_summarize.py | Configuration | 2.7 KB | — |
| processor_config.json | Configuration | 623 B | — |
| scrybe_model.json | Configuration | 2.1 KB | — |
| LICENSE | Documentation | 2.6 KB | — |
| NOTICE | Documentation | 1.7 KB | — |
| README.md | Documentation | 9.4 KB | — |
| convert/pyproject.toml | Other | 586 B | — |
| convert/uv.lock | Other | 114.8 KB | — |
| nemotron3_diar_step.xml | Other | 1.6 MB | — |
| silence_embeds.f32 | Other | 2.0 KB | — |
| .gitattributes | Repository | 1.5 KB | — |
License and Download
- License
- other
- Access
- Open weights, no gate
- Download size
- 396.9 MB
Released by Shameez through its official repository on Hugging Face.
Built From
- Derived from nvidia/Nemotron-3-Diarization
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 396.9 MB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About nemotron-3-diarization-ov
What license is nemotron-3-diarization-ov released under?
other, as its publisher declares it. Read the license text before commercial use.