1
---
2
language: or
3
datasets:
4
- common_voice 
5
metrics:
6
- wer
7
tags:
8
- audio
9
- automatic-speech-recognition
10
- speech
11
- xlsr-fine-tuning-week
12
license: apache-2.0
13
model-index:
14
- name: XLSR Wav2Vec2 Large 53 Odia by Gunjan Chhablani
15
  results:
16
  - task: 
17
      name: Speech Recognition
18
      type: automatic-speech-recognition
19
    dataset:
20
      name: Common Voice or
21
      type: common_voice
22
      args: or
23
    metrics:
24
       - name: Test WER
25
         type: wer
26
         value: 52.64
27
---
28
# Wav2Vec2-Large-XLSR-53-Odia
29
Fine-tuned [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on Odia using the [Common Voice](https://huggingface.co/datasets/common_voice).
30
When using this model, make sure that your speech input is sampled at 16kHz.
31
32
## Usage
33
The model can be used directly (without a language model) as follows:
34
```python
35
import torch
36
import torchaudio
37
from datasets import load_dataset
38
from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
39
test_dataset = load_dataset("common_voice", "or", split="test[:2%]")
40
41
processor = Wav2Vec2Processor.from_pretrained("gchhablani/wav2vec2-large-xlsr-or")
42
model = Wav2Vec2ForCTC.from_pretrained("gchhablani/wav2vec2-large-xlsr-or")
43
resampler = torchaudio.transforms.Resample(48_000, 16_000)
44
# Preprocessing the datasets.
45
# We need to read the aduio files as arrays
46
def speech_file_to_array_fn(batch):
47
    speech_array, sampling_rate = torchaudio.load(batch["path"])
48
    batch["speech"] = resampler(speech_array).squeeze().numpy()
49
    return batch
50
test_dataset = test_dataset.map(speech_file_to_array_fn)
51
inputs = processor(test_dataset["speech"][:2], sampling_rate=16_000, return_tensors="pt", padding=True)
52
with torch.no_grad():
53
    logits = model(inputs.input_values, attention_mask=inputs.attention_mask).logits
54
predicted_ids = torch.argmax(logits, dim=-1)
55
print("Prediction:", processor.batch_decode(predicted_ids))
56
print("Reference:", test_dataset["sentence"][:2])
57
```
58
## Evaluation
59
The model can be evaluated as follows on the Odia test data of Common Voice.
60
```python
61
import torch
62
import torchaudio
63
from datasets import load_dataset, load_metric
64
from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
65
import re
66
test_dataset = load_dataset("common_voice", "or", split="test")
67
wer = load_metric("wer")
68
processor = Wav2Vec2Processor.from_pretrained("gchhablani/wav2vec2-large-xlsr-or")
69
model = Wav2Vec2ForCTC.from_pretrained("gchhablani/wav2vec2-large-xlsr-or")
70
model.to("cuda")
71
chars_to_ignore_regex = '[\,\?\.\!\-\;\:\"\“\%\‘\”\�\–\…\'\_\’\।\|]'
72
resampler = torchaudio.transforms.Resample(48_000, 16_000)
73
# Preprocessing the datasets.
74
# We need to read the aduio files as arrays
75
def speech_file_to_array_fn(batch):
76
    batch["sentence"] = re.sub(chars_to_ignore_regex, '', batch["sentence"]).lower()
77
    speech_array, sampling_rate = torchaudio.load(batch["path"])
78
    batch["speech"] = resampler(speech_array).squeeze().numpy()
79
    return batch
80
test_dataset = test_dataset.map(speech_file_to_array_fn)
81
# Preprocessing the datasets.
82
# We need to read the aduio files as arrays
83
def evaluate(batch):
84
    inputs = processor(batch["speech"], sampling_rate=16_000, return_tensors="pt", padding=True)
85
    with torch.no_grad():
86
        logits = model(inputs.input_values.to("cuda"), attention_mask=inputs.attention_mask.to("cuda")).logits
87
    pred_ids = torch.argmax(logits, dim=-1)
88
    batch["pred_strings"] = processor.batch_decode(pred_ids)
89
    return batch
90
result = test_dataset.map(evaluate, batched=True, batch_size=8)
91
print("WER: {:2f}".format(100 * wer.compute(predictions=result["pred_strings"], references=result["sentence"])))
92
```
93
**Test Result**: 52.64 % 
94
95
## Training
96
The Common Voice `train` and `validation` datasets were used for training.The colab notebook used can be found [here](https://colab.research.google.com/drive/1s8DrwgB5y4Z7xXIrPXo1rQA5_1OZ8WD5?usp=sharing).