File size: 4,205 Bytes
137287d c58cb21 137287d 8cd9d75 137287d 8cd9d75 137287d c58cb21 8cd9d75 137287d 0841e76 137287d 8cd9d75 137287d c58cb21 137287d c58cb21 137287d c58cb21 137287d |
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 |
---
language: cv
datasets:
- common_voice
metrics:
- wer
tags:
- audio
- automatic-speech-recognition
- speech
- xlsr-fine-tuning-week
license: apache-2.0
model-index:
- name: Chuvash XLSR Wav2Vec2 Large 53 by Anton Lozhkov
results:
- task:
name: Speech Recognition
type: automatic-speech-recognition
dataset:
name: Common Voice cv
type: common_voice
args: cv
metrics:
- name: Test WER
type: wer
value: 40.01
---
# Wav2Vec2-Large-XLSR-53-Chuvash
Fine-tuned [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on Chuvash using the [Common Voice](https://huggingface.co/datasets/common_voice) dataset.
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", "cv", split="test[:2%]")
processor = Wav2Vec2Processor.from_pretrained("anton-l/wav2vec2-large-xlsr-53-chuvash")
model = Wav2Vec2ForCTC.from_pretrained("anton-l/wav2vec2-large-xlsr-53-chuvash")
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):
speech_array, sampling_rate = torchaudio.load(batch["path"])
batch["speech"] = resampler(speech_array).squeeze().numpy()
return 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():
logits = 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 Chuvash test data of Common Voice.
```python
import torch
import torchaudio
import urllib.request
import tarfile
import pandas as pd
from tqdm.auto import tqdm
from datasets import load_metric
from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
# Download the raw data instead of using HF datasets to save disk space
data_url = "https://voice-prod-bundler-ee1969a6ce8178826482b88e843c335139bd3fb4.s3.amazonaws.com/cv-corpus-6.1-2020-12-11/cv.tar.gz"
filestream = urllib.request.urlopen(data_url)
data_file = tarfile.open(fileobj=filestream, mode="r|gz")
data_file.extractall()
wer = load_metric("wer")
processor = Wav2Vec2Processor.from_pretrained("anton-l/wav2vec2-large-xlsr-53-chuvash")
model = Wav2Vec2ForCTC.from_pretrained("anton-l/wav2vec2-large-xlsr-53-chuvash")
model.to("cuda")
cv_test = pd.read_csv("cv-corpus-6.1-2020-12-11/cv/test.tsv", sep='\t')
clips_path = "cv-corpus-6.1-2020-12-11/cv/clips/"
def clean_sentence(sent):
sent = sent.lower()
# replace non-alpha characters with space
sent = "".join(ch if ch.isalpha() else " " for ch in sent)
# remove repeated spaces
sent = " ".join(sent.split())
return sent
targets = []
preds = []
for i, row in tqdm(cv_test.iterrows(), total=cv_test.shape[0]):
row["sentence"] = clean_sentence(row["sentence"])
speech_array, sampling_rate = torchaudio.load(clips_path + row["path"])
resampler = torchaudio.transforms.Resample(sampling_rate, 16_000)
row["speech"] = resampler(speech_array).squeeze().numpy()
inputs = processor(row["speech"], sampling_rate=16_000, return_tensors="pt", padding=True)
with torch.no_grad():
logits = model(inputs.input_values.to("cuda"), attention_mask=inputs.attention_mask.to("cuda")).logits
pred_ids = torch.argmax(logits, dim=-1)
targets.append(row["sentence"])
preds.append(processor.batch_decode(pred_ids)[0])
print("WER: {:2f}".format(100 * wer.compute(predictions=preds, references=targets)))
```
**Test Result**: 40.01 %
## Training
The Common Voice `train` and `validation` datasets were used for training.
The script used for training can be found [here](github.com)
|