Selbyggen/nb-whisper-large-coreml

NB-Whisper large – WhisperKit / Core ML

NbAiLab/nb-whisper-large from the National Library of Norway, converted to Core ML for WhisperKit so it runs on the Apple Neural Engine.

This is a format conversion, not a new model. The weights and behaviour are NbAiLab's. All credit for the model belongs to NbAiLab.

Used by the Mac, iPhone and iPad app Sagt for local Norwegian transcription.

Conversion

• Tool: whisperkittools (whisperkit-generate-model --model-version NbAiLab/nb-whisper-large)
• Precision: float16, no quantization
• The conversion's built-in tests (audio encoder and text decoder against the PyTorch reference) passed.
• config.json and generation_config.json are copied unchanged from the original model.

Usage with WhisperKit

let folder = try await WhisperKit.download(
    variant: "nb-whisper-large",
    from: "Selbyggen/nb-whisper-large-coreml"
)
let kit = try await WhisperKit(WhisperKitConfig(modelFolder: folder.path))
let result = try await kit.transcribe(audioPath: "opptak.m4a",
                                      decodeOptions: DecodingOptions(language: "no"))

The model uses the same tokenizer as openai/whisper-large-v3.

Try it without code

Sagt uses this model to transcribe recordings and meetings on Mac, iPhone and iPad. Everything runs on the device, and no audio leaves it.

License

Apache 2.0, the same as the original model.