aalst/nb-whisper-large-distil-turbo-beta-mlx

NB-Whisper Large Distil Turbo BETA — MLX

Note: This is the BETA distilled turbo variant. The original authors note it is a work in progress released for testing purposes. For production use, consider aalst/nb-whisper-large-mlx.

This model was converted to MLX format from NbAiLab/nb-whisper-large-distil-turbo-beta.

Refer to the original model card for detailed information about training data, performance benchmarks, and intended use.

About the Original Model

NB-Whisper Large Distil Turbo BETA is a lighter, faster version of the Norwegian ASR model developed by the National Library of Norway AI Lab (NB AI-Lab). It was distilled from NbAiLab/nb-whisper-large using a teacher-student framework, reducing parameters from 1,550M to 756M (~51% smaller) while retaining competitive transcription quality.

• Languages: Norwegian (Bokmal)
• Optimized for: Real-time transcription, low-latency applications, edge devices

Use with mlx-whisper

pip install mlx-whisper

import mlx_whisper

result = mlx_whisper.transcribe(
    "audio.mp3",
    path_or_hf_repo="aalst/nb-whisper-large-distil-turbo-beta-mlx",
    language="no",
)
print(result["text"])

Conversion Details

• Converted from: NbAiLab/nb-whisper-large-distil-turbo-beta (PyTorch/Safetensors)
• Converted to: MLX format (optimized for Apple Silicon)
• Precision: float16
• Conversion tool: mlx-examples/whisper/convert.py
• Changes: Format conversion only. No fine-tuning or weight modifications.

License

This model inherits the Apache 2.0 license from the original NB-Whisper model.

Citation

@misc{kummervold2024whisperingnorwegian,
    title={Whispering in Norwegian: Navigating Orthographic and Dialectic Challenges},
    author={Per Egil Kummervold and Javier de la Rosa and Freddy Wetjen and Rolv-Arild Braaten and Per Erik Solberg},
    year={2024},
    eprint={2402.01917},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}