wav2vec2-large-xlsr-kazakh

Wav2Vec2-Large-XLSR-53-Kazakh

Fine-tuned facebook/wav2vec2-large-xlsr-53 for Kazakh ASR using the Kazakh Speech Corpus v1.1

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:

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

from utils import get_test_dataset

test_dataset = get_test_dataset("ISSAI_KSC_335RS_v1.1")

processor = Wav2Vec2Processor.from_pretrained("wav2vec2-large-xlsr-kazakh")
model = Wav2Vec2ForCTC.from_pretrained("wav2vec2-large-xlsr-kazakh")


# Preprocessing the datasets.
# We need to read the audio files as arrays
def speech_file_to_array_fn(batch):
    speech_array, sampling_rate = torchaudio.load(batch["path"])
    batch["speech"] = torchaudio.transforms.Resample(sampling_rate, 16_000)(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 test set of Kazakh Speech Corpus v1.1. To evaluate, download the archive, untar and pass the path to data to get_test_dataset as below:

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

from utils import get_test_dataset

test_dataset = get_test_dataset("ISSAI_KSC_335RS_v1.1")
wer = load_metric("wer")

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


# Preprocessing the datasets.
# We need to read the audio files as arrays
def speech_file_to_array_fn(batch):
    batch["sentence"] = re.sub(chars_to_ignore_regex, '', batch["sentence"]).lower()
    speech_array, sampling_rate = torchaudio.load(batch["path"])
    batch["speech"] = torchaudio.transforms.Resample(sampling_rate, 16_000)(speech_array).squeeze().numpy()
    return batch

test_dataset = test_dataset.map(speech_file_to_array_fn)

def evaluate(batch):
    inputs = processor(batch["text"], 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)
    batch["pred_strings"] = processor.batch_decode(pred_ids)
    return 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: 19.65%

Training

The Kazakh Speech Corpus v1.1 train dataset was used for training.