Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,71 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language: it
|
| 3 |
+
license: mit
|
| 4 |
+
tags:
|
| 5 |
+
- whisper
|
| 6 |
+
- automatic-speech-recognition
|
| 7 |
+
- italian
|
| 8 |
+
- ctranslate2
|
| 9 |
+
- faster-whisper
|
| 10 |
+
- whisperx
|
| 11 |
+
- localai
|
| 12 |
+
- int8
|
| 13 |
+
datasets:
|
| 14 |
+
- mozilla-foundation/common_voice_25_0
|
| 15 |
+
- facebook/multilingual_librispeech
|
| 16 |
+
- facebook/voxpopuli
|
| 17 |
+
base_model: openai/whisper-medium
|
| 18 |
+
pipeline_tag: automatic-speech-recognition
|
| 19 |
+
---
|
| 20 |
+
|
| 21 |
+
# whisper-medium-it-multi-ct2-int8
|
| 22 |
+
|
| 23 |
+
[CTranslate2](https://github.com/OpenNMT/CTranslate2) INT8 quantized version of [LocalAI-io/whisper-medium-it-multi](https://huggingface.co/LocalAI-io/whisper-medium-it-multi) for fast CPU inference.
|
| 24 |
+
|
| 25 |
+
**Author:** Ettore Di Giacinto
|
| 26 |
+
|
| 27 |
+
Brought to you by the [LocalAI](https://github.com/mudler/LocalAI) team. This model can be used directly with [LocalAI](https://localai.io).
|
| 28 |
+
|
| 29 |
+
## Usage with LocalAI
|
| 30 |
+
|
| 31 |
+
This model is ready to use with [LocalAI](https://localai.io) via the `whisperx` backend.
|
| 32 |
+
|
| 33 |
+
Save the following as `whisperx-medium-it-multi.yaml` in your LocalAI models directory:
|
| 34 |
+
|
| 35 |
+
```yaml
|
| 36 |
+
name: whisperx-medium-it-multi
|
| 37 |
+
backend: whisperx
|
| 38 |
+
known_usecases:
|
| 39 |
+
- transcript
|
| 40 |
+
parameters:
|
| 41 |
+
model: LocalAI-io/whisper-medium-it-multi-ct2-int8
|
| 42 |
+
language: it
|
| 43 |
+
```
|
| 44 |
+
|
| 45 |
+
Then transcribe audio via the OpenAI-compatible endpoint:
|
| 46 |
+
|
| 47 |
+
```bash
|
| 48 |
+
curl http://localhost:8080/v1/audio/transcriptions \
|
| 49 |
+
-H "Content-Type: multipart/form-data" \
|
| 50 |
+
-F file="@audio.mp3" \
|
| 51 |
+
-F model="whisperx-medium-it-multi"
|
| 52 |
+
```
|
| 53 |
+
|
| 54 |
+
## Usage
|
| 55 |
+
|
| 56 |
+
### faster-whisper
|
| 57 |
+
|
| 58 |
+
```python
|
| 59 |
+
from faster_whisper import WhisperModel
|
| 60 |
+
|
| 61 |
+
model = WhisperModel("LocalAI-io/whisper-medium-it-multi-ct2-int8", device="cpu", compute_type="int8")
|
| 62 |
+
segments, info = model.transcribe("audio.mp3", language="it")
|
| 63 |
+
for segment in segments:
|
| 64 |
+
print(f"[{segment.start:.1f}s - {segment.end:.1f}s] {segment.text}")
|
| 65 |
+
```
|
| 66 |
+
|
| 67 |
+
## Links
|
| 68 |
+
|
| 69 |
+
- **HF Safetensors:** [LocalAI-io/whisper-medium-it-multi](https://huggingface.co/LocalAI-io/whisper-medium-it-multi)
|
| 70 |
+
- **Project:** [github.com/localai-org/italian-whisper](https://github.com/localai-org/italian-whisper)
|
| 71 |
+
- **LocalAI:** [github.com/mudler/LocalAI](https://github.com/mudler/LocalAI)
|