File size: 2,437 Bytes
b6bc6ea
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6c655b3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b6bc6ea
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f8fdeed
b6bc6ea
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
---
language: it
license: mit
tags:
- whisper
- automatic-speech-recognition
- italian
- ctranslate2
- faster-whisper
- whisperx
- localai
- int8
datasets:
- mozilla-foundation/common_voice_25_0
base_model: openai/whisper-tiny
pipeline_tag: automatic-speech-recognition
---

# whisper-tiny-it-ct2-int8

[CTranslate2](https://github.com/OpenNMT/CTranslate2) INT8 quantized version of [LocalAI-io/whisper-tiny-it](https://huggingface.co/LocalAI-io/whisper-tiny-it) for fast CPU inference.

**Author:** Ettore Di Giacinto

Brought to you by the [LocalAI](https://github.com/mudler/LocalAI) team. This model can be used directly with [LocalAI](https://localai.io).

## Usage with LocalAI

This model is ready to use with [LocalAI](https://localai.io) via the `whisperx` backend.

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

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

Then transcribe audio via the OpenAI-compatible endpoint:

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

## Model Details

- **Base model:** openai/whisper-tiny, fine-tuned on Common Voice 25.0 Italian
- **Quantization:** INT8 via CTranslate2
- **Size:** 39MB (vs 578MB for the original fp32 model)
- **WER:** 27.1% on Common Voice 25.0 Italian test set

## Usage

### faster-whisper

```python
from faster_whisper import WhisperModel

model = WhisperModel("LocalAI-io/whisper-tiny-it-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

```python
import whisperx

model = whisperx.load_model("LocalAI-io/whisper-tiny-it-ct2-int8", device="cpu", compute_type="int8")
result = model.transcribe("audio.mp3", language="it")
```

### LocalAI

This model is compatible with [LocalAI](https://github.com/mudler/LocalAI) for local, self-hosted AI inference.

## Links

- **HF Safetensors version:** [LocalAI-io/whisper-tiny-it](https://huggingface.co/LocalAI-io/whisper-tiny-it)
- **Code:** [github.com/mudler/italian-asr](https://github.com/mudler/italian-asr)
- **LocalAI:** [github.com/mudler/LocalAI](https://github.com/mudler/LocalAI)