File size: 4,579 Bytes
14359da
 
 
 
 
 
 
08ee2f5
 
14359da
 
 
 
38cc73b
14359da
f09e2da
 
14359da
 
 
 
48d7d91
 
 
 
14359da
f09e2da
 
 
 
 
 
 
 
 
 
 
 
 
3f9fa35
 
 
 
 
 
 
f09e2da
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
14359da
 
161497a
14359da
6ce253e
6cdc194
14359da
cbcda8d
6cdc194
cbcda8d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
14359da
 
 
 
 
 
 
 
 
 
 
 
 
 
3c1d5fd
 
 
 
 
161497a
0c94710
3c1d5fd
 
 
 
24ee097
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
---
language:
- en
license: apache-2.0
tags:
- automatic-speech-recognition
- en
- hf-asr-leaderboard
- mozilla-foundation/common_voice_8_0
- robust-speech-event
datasets:
- mozilla-foundation/common_voice_8_0
model-index:
- name: XLS-R Wav2Vec2 English by Jonatas Grosman
  results:
  - task:
      name: Automatic Speech Recognition
      type: automatic-speech-recognition
    dataset:
      name: Common Voice 8
      type: mozilla-foundation/common_voice_8_0
      config: en
      split: test
      args: 
        language: en
    metrics:
    - name: Test WER
      type: wer
      value: 21.05
    - name: Test CER
      type: cer
      value: 8.44
    - name: Test WER (+LM)
      type: wer
      value: 17.31
    - name: Test CER (+LM)
      type: cer
      value: 7.77
  - task:
      name: Automatic Speech Recognition
      type: automatic-speech-recognition
    dataset:
      name: Robust Speech Event - Dev Data
      type: speech-recognition-community-v2/dev_data
      args: en
    metrics:
    - name: Dev WER
      type: wer
      value: 20.53
    - name: Dev CER
      type: cer
      value: 9.31
    - name: Dev WER (+LM)
      type: wer
      value: 17.7
    - name: Dev CER (+LM)
      type: cer
      value: 8.93
  - task:
      name: Automatic Speech Recognition
      type: automatic-speech-recognition
    dataset:
      name: Robust Speech Event - Test Data
      type: speech-recognition-community-v2/eval_data
      args: en
    metrics:
    - name: Test WER
      type: wer
      value: 17.88
---

# Fine-tuned XLS-R 1B model for speech recognition in English

Fine-tuned [facebook/wav2vec2-xls-r-1b](https://huggingface.co/facebook/wav2vec2-xls-r-1b) on English using the train and validation splits of [Common Voice 8.0](https://huggingface.co/datasets/mozilla-foundation/common_voice_8_0), [Multilingual LibriSpeech](https://www.openslr.org/94/), [TED-LIUMv3](https://www.openslr.org/51/), and [Voxpopuli](https://github.com/facebookresearch/voxpopuli).
When using this model, make sure that your speech input is sampled at 16kHz.

This model has been fine-tuned by the [HuggingSound](https://github.com/jonatasgrosman/huggingsound) tool, and thanks to the GPU credits generously given by the [OVHcloud](https://www.ovhcloud.com/en/public-cloud/ai-training/) :)

## Usage

Using the [HuggingSound](https://github.com/jonatasgrosman/huggingsound) library:

```python
from huggingsound import SpeechRecognitionModel

model = SpeechRecognitionModel("jonatasgrosman/wav2vec2-xls-r-1b-english")
audio_paths = ["/path/to/file.mp3", "/path/to/another_file.wav"]

transcriptions = model.transcribe(audio_paths)
```

Writing your own inference script:

```python
import torch
import librosa
from datasets import load_dataset
from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor

LANG_ID = "en"
MODEL_ID = "jonatasgrosman/wav2vec2-xls-r-1b-english"
SAMPLES = 10

test_dataset = load_dataset("common_voice", LANG_ID, split=f"test[:{SAMPLES}]")

processor = Wav2Vec2Processor.from_pretrained(MODEL_ID)
model = Wav2Vec2ForCTC.from_pretrained(MODEL_ID)

# Preprocessing the datasets.
# We need to read the audio files as arrays
def speech_file_to_array_fn(batch):
    speech_array, sampling_rate = librosa.load(batch["path"], sr=16_000)
    batch["speech"] = speech_array
    batch["sentence"] = batch["sentence"].upper()
    return batch

test_dataset = test_dataset.map(speech_file_to_array_fn)
inputs = processor(test_dataset["speech"], sampling_rate=16_000, return_tensors="pt", padding=True)

with torch.no_grad():
    logits = model(inputs.input_values, attention_mask=inputs.attention_mask).logits

predicted_ids = torch.argmax(logits, dim=-1)
predicted_sentences = processor.batch_decode(predicted_ids)
```

## Evaluation Commands

1. To evaluate on `mozilla-foundation/common_voice_8_0` with split `test`

```bash
python eval.py --model_id jonatasgrosman/wav2vec2-xls-r-1b-english --dataset mozilla-foundation/common_voice_8_0 --config en --split test
```

2. To evaluate on `speech-recognition-community-v2/dev_data`

```bash
python eval.py --model_id jonatasgrosman/wav2vec2-xls-r-1b-english --dataset speech-recognition-community-v2/dev_data --config en --split validation --chunk_length_s 5.0 --stride_length_s 1.0
```

## Citation
If you want to cite this model you can use this:

```bibtex
@misc{grosman2021xlsr-1b-english,
  title={Fine-tuned {XLS-R} 1{B} model for speech recognition in {E}nglish},
  author={Grosman, Jonatas},
  howpublished={\url{https://huggingface.co/jonatasgrosman/wav2vec2-xls-r-1b-english}},
  year={2022}
}
```