whisper-large-v3-turbo-it-yodas-only-ct2-int8

CTranslate2 INT8 quantized version of LocalAI-io/whisper-large-v3-turbo-it-yodas-only for fast CPU inference.

Author: Ettore Di Giacinto

Brought to you by the LocalAI team. This model can be used directly with LocalAI.

Training

• Base model: openai/whisper-large-v3-turbo, fine-tuned on YODAS-Granary Italian (asr_only + ast capped at 200000)
• Quantization: INT8 via CTranslate2

Usage

faster-whisper

from faster_whisper import WhisperModel

model = WhisperModel("LocalAI-io/whisper-large-v3-turbo-it-yodas-only-ct2-int8", device="cpu", compute_type="int8")
segments, info = model.transcribe("audio.mp3", language="it")
for segment in segments:
    print(f"[{segment.start:.1f}s - {segment.end:.1f}s] {segment.text}")

WhisperX

import whisperx

model = whisperx.load_model("LocalAI-io/whisper-large-v3-turbo-it-yodas-only-ct2-int8", device="cpu", compute_type="int8")
result = model.transcribe("audio.mp3", language="it")

Links

• HF Safetensors: LocalAI-io/whisper-large-v3-turbo-it-yodas-only
• Code: github.com/mudler/italian-asr
• LocalAI: github.com/mudler/LocalAI