Modelpublic

LocalAI-io/whisper-large-v3-turbo-it-multi-yodas-asr_only-nocv

sourceHugging Facemitupdated 5mo agoView on Hugging Face

0likes8downloads

Model Card

whisper-large-v3-turbo-it-multi-yodas-asr_only-nocv

Fine-tuned openai/whisper-large-v3-turbo for Italian automatic speech recognition.

Training mix: MLS + VoxPopuli + YODAS-Granary (asr_only) Italian (~165k samples). 10,000 steps, lr 1e-5, bf16 on an NVIDIA GB10.

Brought to you by the LocalAI team.

Available formats

—CTranslate2 INT8 (faster-whisper / WhisperX / LocalAI): LocalAI-io/whisper-large-v3-turbo-it-multi-yodas-asr_only-nocv-ct2-int8
—whisper.cpp GGML (q40/q50/q80): [LocalAI-io/whisper-large-v3-turbo-it-multi-yodas-asronly-nocv-ggml](https://huggingface.co/LocalAI-io/whisper-large-v3-turbo-it-multi-yodas-asr_only-nocv-ggml)
—safetensors (this repo) — for 🤗 Transformers

Usage

python

from transformers import pipeline
pipe = pipeline("automatic-speech-recognition", model="LocalAI-io/whisper-large-v3-turbo-it-multi-yodas-asr_only-nocv")
result = pipe("audio.mp3", generate_kwargs={"language": "it", "task": "transcribe"})
print(result["text"])