whisper-tiny-it-multi

Fine-tuned openai/whisper-tiny (39M params) for Italian ASR on multiple datasets.

Author: Ettore Di Giacinto

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

Usage with LocalAI

This model is ready to use with LocalAI via the whisperx backend.

Save the following as whisperx-tiny-it-multi.yaml in your LocalAI models directory:

name: whisperx-tiny-it-multi
backend: whisperx
known_usecases:
  - transcript
parameters:
  model: LocalAI-io/whisper-tiny-it-multi-ct2-int8
  language: it

Then transcribe audio via the OpenAI-compatible endpoint:

curl http://localhost:8080/v1/audio/transcriptions \
  -H "Content-Type: multipart/form-data" \
  -F file="@audio.mp3" \
  -F model="whisperx-tiny-it-multi"

Results

Evaluated on combined test set (Common Voice + MLS + VoxPopuli, 17,598 samples):

Step	WER
1000	39.8%
3000	33.5%
5000	31.4%
10000	29.4%

Training Details

• Base model: openai/whisper-tiny (39M parameters)
• Datasets: Common Voice 25.0 Italian (173k) + MLS Italian (60k) + VoxPopuli Italian (23k) = 255k train samples
• Steps: 10,000 (batch size 32)
• Learning rate: 1e-5 with 500 warmup steps
• Precision: bf16 on NVIDIA GB10

Usage

Transformers

from transformers import pipeline

pipe = pipeline("automatic-speech-recognition", model="LocalAI-io/whisper-tiny-it-multi")
result = pipe("audio.mp3", generate_kwargs={"language": "it", "task": "transcribe"})
print(result["text"])

CTranslate2 / faster-whisper

For optimized CPU inference: LocalAI-io/whisper-tiny-it-multi-ct2-int8

Links

• CV-only version: LocalAI-io/whisper-tiny-it (WER 27.1% on CV test)
• CTranslate2 INT8: LocalAI-io/whisper-tiny-it-multi-ct2-int8
• Code: github.com/localai-org/italian-whisper
• LocalAI: github.com/mudler/LocalAI