babelbit-uid142-whisper-large-v3

Whisper large-v3 model for CTranslate2

This repository contains the conversion of openai/whisper-large-v3 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("large-v3")

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:

ct2-transformers-converter --model openai/whisper-large-v3 --output_dir faster-whisper-large-v3 \
    --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.

Babelbit subnet 59 deployment

This public revision is the CTranslate2 ASR artifact bound to miner UID 142. The managed runtime uses Faster-Whisper task="translate" for direct speech-to-English translation and Kokoro af_heart for speech synthesis. The Whisper weights are the unmodified upstream conversion; the low-latency streaming, stability, pacing, and output logic lives in the paired managed container.

Runtime metadata and the exact scoring configuration are recorded in babelbit_runtime.json.