File size: 2,155 Bytes
bb66f96
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4ade5d7
 
 
 
 
 
bb66f96
 
 
 
 
 
 
4ade5d7
bb66f96
4ade5d7
 
 
 
bb66f96
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: cc-by-nc-sa-4.0
datasets:
- voice-is-cool/voxtube
base_model:
- openai/whisper-tiny
library_name: transformers
tags:
- speaker-verification
- voice
- audio
- speaker-recognition
- speaker-embedding
- speaker-identification
- speaker
- whisper
- voxtube
---
# Whisper Speaker Identification (WSI)

**Whisper Speaker Identification (WSI)** is a state-of-the-art speaker identification model designed for multilingual scenarios.The WSI model adapts OpenAI's Whisper encoder and fine-tunes it with a projection head using triplet loss-based metric learning. This approach enhances its ability to generate discriminative, language-agnostic speaker embeddings.WSI demonstrates state-of-the-art performance on multilingual datasets, achieving lower Equal Error Rates (EER) and higher F1 Scores compared to models such as **pyannote/wespeaker-voxceleb-resnet34-LM** and **speechbrain/spkrec-ecapa-voxceleb**.

## Installation

Install the `whisper-speaker-id` library via pip:

```
pip install whisper-speaker-id 
```

## Usage

The `wsi` library provides a simple interface to use the WSI model for embedding generation and speaker similarity tasks. 

## Download the model from Huggingface

[WSI Model on Hugging Face](https://huggingface.co/emon-j/WSI)

### Generate Speaker Embeddings  

```python
from whisper-speaker-id import load_model, process_single_audio
model, feature_extractor = load_model(
    model_path_or_repo_id="emon-j/WSI",
    filename="wsi.pth"
)
# Process an audio file
embedding = process_single_audio(model, feature_extractor, "path/to/audio.wav")
print("Speaker Embedding:", embedding)
```

### Calculate Similarity Between Two Audio Files

```python
from whisper-speaker-id import load_model, process_audio_pair

model, feature_extractor = load_model(
    model_path_or_repo_id="emon-j/WSI",
    filename="wsi.pth"
)

# Compute similarity between two audio files
similarity = process_audio_pair(
    model, feature_extractor, "path/to/audio1.wav", "path/to/audio2.wav"
)
print("Similarity Score:", similarity)
```

### Cite This Work

Comming Soon!

### License

This project is licensed under the CC BY-NC-SA 4.0 License.