Whisper large-v3 turbo model for CTranslate2

This repository contains the conversion of katrintomanek/whisper-large-v3-turbo_Akan_standardspeech_specaugment to the CTranslate2 model format.

This model can be used in CTranslate2 or projects based on CTranslate2 such as faster-whisper.

Example

from faster_whisper import WhisperModel

model = WhisperModel("fiifinketia/whisper-large-v3-turbo-akan")

segments, info = model.transcribe("audio.mp3")
for segment in segments:
    print("[%.2fs -> %.2fs] %s" % (segment.start, segment.end, segment.text))

Conversion details

The original model was converted with the following command after adding missing files from openai/whisper-large-v3-turbo to support ct2 transformers convertion:

ct2-transformers-converter --model katrintomanek/whisper-large-v3-turbo_Akan_standardspeech_specaugment --output_dir whisper-large-v3-turbo-akan \
    --copy_files tokenizer.json preprocessor_config.json --quantization float16

Note that the model weights are saved in FP16. This type can be changed when the model is loaded using the compute_type option in CTranslate2.

More information

For more information about the original model, see its model card.