File size: 9,733 Bytes
5bca4d0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f516bd1
5bca4d0
 
f516bd1
 
5bca4d0
 
f516bd1
5bca4d0
 
f516bd1
 
5bca4d0
 
f516bd1
 
5bca4d0
 
f516bd1
5bca4d0
f516bd1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5bca4d0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
---
tags: 
- pyannote
- pyannote-audio
- pyannote-audio-pipeline
- audio
- voice
- speech
- speaker
- speaker-diarization
- speaker-change-detection
- voice-activity-detection
- overlapped-speech-detection
datasets:
- ami
- dihard
- voxconverse
- aishell
- repere
- voxceleb
license: mit
---

# 🎹 Speaker diarization

Relies on pyannote.audio 2.0: see [installation instructions](https://github.com/pyannote/pyannote-audio/tree/develop#installation).


## TL;DR

```python
# load the pipeline from Hugginface Hub
from pyannote.audio import Pipeline
pipeline = Pipeline.from_pretrained("pyannote/[email protected]")

# apply the pipeline to an audio file
diarization = pipeline("audio.wav")

# dump the diarization output to disk using RTTM format
with open("audio.rttm", "w") as rttm:
    diarization.write_rttm(rttm)
```

## Advanced usage

If the number of speakers is known in advance, you can include the num_speakers parameter in the parameters dictionary:

```python
handler = EndpointHandler()
diarization = handler({"inputs": base64_audio, "parameters": {"num_speakers": 2}})
```

You can also provide lower and/or upper bounds on the number of speakers using the min_speakers and max_speakers parameters:

```python
handler = EndpointHandler()
diarization = handler({"inputs": base64_audio, "parameters": {"min_speakers": 2, "max_speakers": 5}})
```

If you're feeling adventurous, you can experiment with various pipeline hyperparameters. 
For instance, you can use a more aggressive voice activity detection by increasing the value of segmentation_onset threshold:

```python
hparams = handler.pipeline.parameters(instantiated=True)
hparams["segmentation_onset"] += 0.1
handler.pipeline.instantiate(hparams)
```
To apply the updated handler for the API inference that can handle the number of speakers, use the following code:
```python
from typing import Dict
from pyannote.audio import Pipeline
import torch 
import base64
import numpy as np

SAMPLE_RATE = 16000

class EndpointHandler():
    def __init__(self, path=""):
        # load the model
        self.pipeline = Pipeline.from_pretrained("KIFF/pyannote-speaker-diarization-endpoint")

    def __call__(self, data: Dict[str, bytes]) -> Dict[str, str]:
        """
        Args:
            data (:obj:):
                includes the deserialized audio file as bytes
        Return:
            A :obj:`dict`:. base64 encoded image
        """
        # process input
        inputs = data.pop("inputs", data)
        parameters = data.pop("parameters", None) #  min_speakers=2, max_speakers=5

        # decode the base64 audio data
        audio_data = base64.b64decode(inputs)
        audio_nparray = np.frombuffer(audio_data, dtype=np.int16)

        # prepare pynannote input
        audio_tensor= torch.from_numpy(audio_nparray).float().unsqueeze(0)
        pyannote_input = {"waveform": audio_tensor, "sample_rate": SAMPLE_RATE}
        
        # apply pretrained pipeline
        # pass inputs with all kwargs in data
        if parameters is not None:
            diarization = self.pipeline(pyannote_input, **parameters)
        else:
            diarization = self.pipeline(pyannote_input)

        # postprocess the prediction
        processed_diarization = [
            {"label": str(label), "start": str(segment.start), "stop": str(segment.end)}
            for segment, _, label in diarization.itertracks(yield_label=True)
        ]
        
        return {"diarization": processed_diarization}
```
## Benchmark 

### Real-time factor

Real-time factor is around 5% using one Nvidia Tesla V100 SXM2 GPU (for the neural inference part) and one Intel Cascade Lake 6248 CPU (for the clustering part).

In other words, it takes approximately 3 minutes to process a one hour conversation.

### Accuracy

This pipeline is benchmarked on a growing collection of datasets.  

Processing is fully automatic:

* no manual voice activity detection (as is sometimes the case in the literature)
* no manual number of speakers (though it is possible to provide it to the pipeline)
* no fine-tuning of the internal models nor tuning of the pipeline hyper-parameters to each dataset

... with the least forgiving diarization error rate (DER) setup (named *"Full"* in [this paper](https://doi.org/10.1016/j.csl.2021.101254)):

* no forgiveness collar
* evaluation of overlapped speech


| Benchmark                                                                                                                          | [DER%](. "Diarization error rate") | [FA%](. "False alarm rate") | [Miss%](. "Missed detection rate") | [Conf%](. "Speaker confusion rate") | Expected output                                                                            | File-level evaluation                                                                      |
| ---------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------- | --------------------------- | ---------------------------------- | ----------------------------------- | ------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------ |
| [AISHELL-4](http://www.openslr.org/111/)                                                                                           | 14.61                              | 3.31                        | 4.35                               | 6.95                                | [RTTM](reproducible_research/AISHELL.SpeakerDiarization.Full.test.rttm)                    | [eval](reproducible_research/AISHELL.SpeakerDiarization.Full.test.eval)                    |
| [AMI *Mix-Headset*](https://groups.inf.ed.ac.uk/ami/corpus/) [*only_words*](https://github.com/BUTSpeechFIT/AMI-diarization-setup) | 18.21                              | 3.28                        | 11.07                              | 3.87                                | [RTTM](reproducible_research/2022.07/AMI.SpeakerDiarization.only_words.test.rttm)          | [eval](reproducible_research/2022.07/AMI.SpeakerDiarization.only_words.test.eval)          |
| [AMI *Array1-01*](https://groups.inf.ed.ac.uk/ami/corpus/) [*only_words*](https://github.com/BUTSpeechFIT/AMI-diarization-setup)   | 29.00                              | 2.71                        | 21.61                              | 4.68                                | [RTTM](reproducible_research/2022.07/AMI-SDM.SpeakerDiarization.only_words.test.rttm)      | [eval](reproducible_research/2022.07/AMI-SDM.SpeakerDiarization.only_words.test.eval)      |
| [CALLHOME](https://catalog.ldc.upenn.edu/LDC2001S97) [*Part2*](https://github.com/BUTSpeechFIT/CALLHOME_sublists/issues/1)         | 30.24                              | 3.71                        | 16.86                              | 9.66                                | [RTTM](reproducible_research/2022.07/CALLHOME.SpeakerDiarization.CALLHOME.test.rttm)       | [eval](reproducible_research/2022.07/CALLHOME.SpeakerDiarization.CALLHOME.test.eval)       |
| [DIHARD 3 *Full*](https://arxiv.org/abs/2012.01477)                                                                                | 20.99                              | 4.25                        | 10.74                              | 6.00                                | [RTTM](reproducible_research/2022.07/DIHARD.SpeakerDiarization.Full.test.rttm)             | [eval](reproducible_research/2022.07/DIHARD.SpeakerDiarization.Full.test.eval)             |
| [REPERE *Phase 2*](https://islrn.org/resources/360-758-359-485-0/)                                                                 | 12.62                              | 1.55                        | 3.30                               | 7.76                                | [RTTM](reproducible_research/2022.07/REPERE.SpeakerDiarization.Full.test.rttm)             | [eval](reproducible_research/2022.07/REPERE.SpeakerDiarization.Full.test.eval)             |
| [VoxConverse *v0.0.2*](https://github.com/joonson/voxconverse)                                                                     | 12.76                              | 3.45                        | 3.85                               | 5.46                                | [RTTM](reproducible_research/2022.07/VoxConverse.SpeakerDiarization.VoxConverse.test.rttm) | [eval](reproducible_research/2022.07/VoxConverse.SpeakerDiarization.VoxConverse.test.eval) |


## Support

For commercial enquiries and scientific consulting, please contact [me](mailto:[email protected]).  
For [technical questions](https://github.com/pyannote/pyannote-audio/discussions) and [bug reports](https://github.com/pyannote/pyannote-audio/issues), please check [pyannote.audio](https://github.com/pyannote/pyannote-audio) Github repository.


## Citations

```bibtex
@inproceedings{Bredin2021,
  Title = {{End-to-end speaker segmentation for overlap-aware resegmentation}},
  Author = {{Bredin}, Herv{\'e} and {Laurent}, Antoine},
  Booktitle = {Proc. Interspeech 2021},
  Address = {Brno, Czech Republic},
  Month = {August},
  Year = {2021},
}
```

```bibtex
@inproceedings{Bredin2020,
  Title = {{pyannote.audio: neural building blocks for speaker diarization}},
  Author = {{Bredin}, Herv{\'e} and {Yin}, Ruiqing and {Coria}, Juan Manuel and {Gelly}, Gregory and {Korshunov}, Pavel and {Lavechin}, Marvin and {Fustes}, Diego and {Titeux}, Hadrien and {Bouaziz}, Wassim and {Gill}, Marie-Philippe},
  Booktitle = {ICASSP 2020, IEEE International Conference on Acoustics, Speech, and Signal Processing},
  Address = {Barcelona, Spain},
  Month = {May},
  Year = {2020},
}
```