You can deploy this model and start earning money today!

kramerthomas/
roest-v3-whisper-1.5b-coreml

Name: roest-v3-whisper-1.5b-coreml
Availability: InStock

whisperkit

coreml

whisper

automatic-speech-recognition

CoRal-project/roest-v3-whisper-1.5b

finetune

openrail

Views

No views yet

Røst v3 Whisper 1.5B — WhisperKit CoreML

CoreML conversion of CoRal-project/roest-v3-whisper-1.5b for on-device Danish speech recognition with WhisperKit.

The source model reports 11.6% CER on conversational CoRal-v3 and 4.5% CER on read-aloud CoRal-v3. This repository changes the runtime format, not the training data or intended use.

Reproduce the conversion

Generated on Apple Silicon with whisperkittools commit 84f77a83c8f530022ae55fbb1a64b3351ef63c7a:

bash

whisperkit-generate-model \
  --model-version CoRal-project/roest-v3-whisper-1.5b \
  --output-dir ./roest-v3-whisper-1.5b-coreml \
  --generate-quantized-variants \
  --allowed-nbits 8 \
  --force-recipe-nbits \
  --disable-default-tests

The published variant contains WhisperKit's AudioEncoder, MelSpectrogram, and TextDecoder compiled model bundles plus the original model configuration.

License and restrictions

The source model and these converted weights use the source model's custom OpenRAIL-M-derived license. Commercial use is allowed, but speech synthesis, voice cloning, and biometric identification are prohibited. See LICENSE for the complete terms.

Citation:

Nielsen, D. S., et al. CoRal: A Danish Corpus and Benchmark for Automatic Speech Recognition. Alexandra Institute, 2024.

kramerthomas/roest-v3-whisper-1.5b-coreml

Røst v3 Whisper 1.5B — WhisperKit CoreML

Reproduce the conversion

License and restrictions

kramerthomas/
roest-v3-whisper-1.5b-coreml