File size: 4,007 Bytes
0b2cea5
0ce8c65
0b2cea5
0ce8c65
0b2cea5
 
 
 
 
 
 
 
 
0ce8c65
0b2cea5
 
 
 
 
0ce8c65
0b2cea5
0ce8c65
0b2cea5
 
 
0ce8c65
0b2cea5
 
 
 
0ce8c65
0b2cea5
 
 
 
 
 
 
 
 
 
 
0ce8c65
0b2cea5
0ce8c65
 
0b2cea5
 
 
 
 
 
0ce8c65
 
 
0b2cea5
 
 
 
 
0ce8c65
0b2cea5
 
 
 
 
 
 
 
 
 
0ce8c65
0b2cea5
 
 
 
 
 
 
 
 
0ce8c65
0b2cea5
 
0ce8c65
 
0b2cea5
 
0ce8c65
0b2cea5
 
 
 
 
0ce8c65
 
 
 
0b2cea5
 
 
 
 
 
0ce8c65
0b2cea5
0ce8c65
 
0b2cea5
0ce8c65
 
 
0b2cea5
 
 
 
 
 
0ce8c65
0b2cea5
 
 
 
 
0ce8c65
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
---
language: bem 
datasets:
- BembaSpeech 
metrics:
- wer
tags:
- audio
- automatic-speech-recognition
- speech
- xlsr-fine-tuning-week
license: apache-2.0
model-index:
- name: XLSR Wav2Vec2 Bemba by Claytone Sikasote     
  results:
  - task: 
      name: Speech Recognition
      type: automatic-speech-recognition
    dataset:
      name: BembaSpeech bem 
      type: bembaspeech
      args: bem 
    metrics:
       - name: Test WER
         type: wer
         value: 42.14 
---

# Wav2Vec2-Large-XLSR-53-Bemba #TODO: replace language with your {language}, *e.g.* French

Fine-tuned [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on Bemba using the [BembaSpeech](https://csikasote.github.io/BembaSpeech). When using this model, make sure that your speech input is sampled at 16kHz.

## Usage

The model can be used directly (without a language model) as follows:

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

test_dataset = load_dataset("common_voice", "{lang_id}", split="test[:2%]") 

processor = Wav2Vec2Processor.from_pretrained("{model_id}") 
model = Wav2Vec2ForCTC.from_pretrained("{model_id}") 

resampler = torchaudio.transforms.Resample(48_000, 16_000)

# Preprocessing the datasets.
# We need to read the aduio files as arrays
def speech_file_to_array_fn(batch):
\tspeech_array, sampling_rate = torchaudio.load(batch["path"])
\tbatch["speech"] = resampler(speech_array).squeeze().numpy()
\treturn batch

test_dataset = test_dataset.map(speech_file_to_array_fn)
inputs = processor(test_dataset["speech"][:2], sampling_rate=16_000, return_tensors="pt", padding=True)

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

predicted_ids = torch.argmax(logits, dim=-1)

print("Prediction:", processor.batch_decode(predicted_ids))
print("Reference:", test_dataset["sentence"][:2])
```


## Evaluation

The model can be evaluated as follows on the Bemba test data of BembaSpeech.  


```python
import torch
import torchaudio
from datasets import load_dataset, load_metric
from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
import re

test_dataset = load_dataset("common_voice", "{lang_id}", split="test") 
wer = load_metric("wer")

processor = Wav2Vec2Processor.from_pretrained("csikasote/wav2vec2-large-xlsr-bemba") 
model = Wav2Vec2ForCTC.from_pretrained("csikasote/wav2vec2-large-xlsr-bemba") 
model.to("cuda")

chars_to_ignore_regex = '[\\,\\?\\.\\!\\-\\;\\:\\"\\“]'  
#resampler = torchaudio.transforms.Resample(48_000, 16_000)

# Preprocessing the datasets.
# We need to read the aduio files as arrays
def speech_file_to_array_fn(batch):
\tbatch["sentence"] = re.sub(chars_to_ignore_regex, '', batch["sentence"]).lower()
\tspeech_array, sampling_rate = torchaudio.load(batch["path"])
\tbatch["speech"] = speech_array.squeeze().numpy() 
\treturn batch

test_dataset = test_dataset.map(speech_file_to_array_fn)

# Preprocessing the datasets.
# We need to read the aduio files as arrays
def evaluate(batch):
\tinputs = processor(batch["speech"], sampling_rate=16_000, return_tensors="pt", padding=True)

\twith torch.no_grad():
\t\tlogits = model(inputs.input_values.to("cuda"), attention_mask=inputs.attention_mask.to("cuda")).logits

\tpred_ids = torch.argmax(logits, dim=-1)
\tbatch["pred_strings"] = processor.batch_decode(pred_ids)
\treturn batch

result = test_dataset.map(evaluate, batched=True, batch_size=8)

print("WER: {:2f}".format(100 * wer.compute(predictions=result["pred_strings"], references=result["sentence"])))
```

**Test Result**: 42.14 %  

## Training

The BembaSpeech `train`, `dev` and `test` datasets were used for training, development and evaluation respectively  # TODO: adapt to state all the datasets that were used for training.

The script used for training can be found [here](https://colab.research.google.com/drive/1IgdR-EQq5rgmBqw5O6tcfJpmXM8rDX55?usp=sharing).