wav2vec2-indonesian-javanese-sundanese/README.md

language, datasets, metrics, tags, license, model-index

language	datasets	metrics	tags	license	model-index
id jv sun	mozilla-foundation/common_voice_7_0 openslr magic_data titml	wer	audio automatic-speech-recognition hf-asr-leaderboard id jv robust-speech-event speech su	apache-2.0	name results Wav2Vec2 Indonesian Javanese and Sundanese by Indonesian NLP task dataset metrics name type Automatic Speech Recognition automatic-speech-recognition name type args Common Voice 6.1 common_voice id name type value Test WER wer 4.056 name type value Test CER cer 1.472 task dataset metrics name type Automatic Speech Recognition automatic-speech-recognition name type args Common Voice 7 mozilla-foundation/common_voice_7_0 id name type value Test WER wer 4.492 name type value Test CER cer 1.577 task dataset metrics name type Automatic Speech Recognition automatic-speech-recognition name type args Robust Speech Event - Dev Data speech-recognition-community-v2/dev_data id name type value Test WER wer 48.94 task dataset metrics name type Automatic Speech Recognition automatic-speech-recognition name type args Robust Speech Event - Test Data speech-recognition-community-v2/eval_data id name type value Test WER wer 68.95

Multilingual Speech Recognition for Indonesian Languages

This is the model built for the project Multilingual Speech Recognition for Indonesian Languages. It is a fine-tuned facebook/wav2vec2-large-xlsr-53 model on the Indonesian Common Voice dataset, High-quality TTS data for Javanese - SLR41, and High-quality TTS data for Sundanese - SLR44 datasets.

We also provide a live demo to test the model.

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

test_dataset = load_dataset("common_voice", "id", split="test[:2%]")

processor = Wav2Vec2Processor.from_pretrained("indonesian-nlp/wav2vec2-indonesian-javanese-sundanese")
model = Wav2Vec2ForCTC.from_pretrained("indonesian-nlp/wav2vec2-indonesian-javanese-sundanese")

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[:2]["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)

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

Evaluation

The model can be evaluated as follows on the Indonesian test data of Common Voice.

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

test_dataset = load_dataset("common_voice", "id", split="test")
wer = load_metric("wer")

processor = Wav2Vec2Processor.from_pretrained("indonesian-nlp/wav2vec2-indonesian-javanese-sundanese")
model = Wav2Vec2ForCTC.from_pretrained("indonesian-nlp/wav2vec2-indonesian-javanese-sundanese") 
model.to("cuda")

chars_to_ignore_regex = '[\,\?\.\!\-\;\:\"\“\%\‘\'\”\<5C>]'

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

# 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"] = resampler(speech_array).squeeze().numpy()
    return batch

test_dataset = test_dataset.map(speech_file_to_array_fn)

# Preprocessing the datasets.
# We need to read the audio files as arrays
def evaluate(batch):
    inputs = processor(batch["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)
    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: 11.57 %

Training

The Common Voice train, validation, and ... datasets were used for training as well as ... and ... # TODO

The script used for training can be found here (will be available soon)