Sharif-wav2vec2

This is a fine-tuned version of Sharif Wav2vec2 for Farsi. The base model went through a fine-tuning process in which 108 hours of Commonvoice's Farsi samples with a sampling rate equal to 16kHz. Afterward, we trained a 5gram using kenlm toolkit and used it in the processor which increased our accuracy on online ASR.

Usage

When using the model, ensure that your speech input is sampled at 16Khz. Prior to the usage, you may need to install the below dependencies:

pip install pyctcdecode
pip install pypi-kenlm

For testing, you can use the hosted inference API at the hugging face (There are provided examples from common-voice). It may take a while to transcribe the given voice; Or you can use the bellow code for a local run:

import tensorflow
import torchaudio
import torch
import numpy as np

from transformers import AutoProcessor, AutoModelForCTC

processor = AutoProcessor.from_pretrained("SLPL/Sharif-wav2vec2")
model = AutoModelForCTC.from_pretrained("SLPL/Sharif-wav2vec2")

speech_array, sampling_rate = torchaudio.load("path/to/your.wav")
speech_array = speech_array.squeeze().numpy()

features = processor(
    speech_array,
    sampling_rate=processor.feature_extractor.sampling_rate,
    return_tensors="pt",
    padding=True)

with torch.no_grad():
    logits = model(
        features.input_values,
        attention_mask=features.attention_mask).logits
    prediction = processor.batch_decode(logits.numpy()).text

print(prediction[0])
# تست

Evaluation

For the evaluation, you can use the code below. Ensure your dataset to be in following form in order to avoid any further conflict:

path	reference
path/to/audio_file.wav	"TRANSCRIPTION"

also, make sure you have installed pip install jiwer prior to running.

import tensorflow
import torchaudio
import torch
import librosa
from datasets import load_dataset,load_metric
import numpy as np
from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
from transformers import Wav2Vec2ProcessorWithLM

model = Wav2Vec2ForCTC.from_pretrained("SLPL/Sharif-wav2vec2") 
processor = Wav2Vec2ProcessorWithLM.from_pretrained("SLPL/Sharif-wav2vec2") 

def speech_file_to_array_fn(batch):
    speech_array, sampling_rate = torchaudio.load(batch["path"])
    speech_array = speech_array.squeeze().numpy()
    speech_array = librosa.resample(
        np.asarray(speech_array),
        sampling_rate,
        processor.feature_extractor.sampling_rate)
    batch["speech"] = speech_array
    return batch

def predict(batch):
    features = processor(
        batch["speech"], 
        sampling_rate=processor.feature_extractor.sampling_rate, 
        return_tensors="pt", 
        padding=True
    )

    with torch.no_grad():
        logits = model(
            features.input_values,
            attention_mask=features.attention_mask).logits
    batch["prediction"] = processor.batch_decode(logits.numpy()).text
    return batch
    
dataset = load_dataset(
    "csv",
    data_files={"test":"dataset.eval.csv"},
    delimiter=",")["test"]
dataset = dataset.map(speech_file_to_array_fn)

result = dataset.map(predict, batched=True, batch_size=4)
wer = load_metric("wer")

print("WER: {:.2f}".format(wer.compute(
    predictions=result["prediction"],
    references=result["reference"])))

Result (WER) on common-voice 6.1:

cleaned	other
0.06	0.16

Citation

If you want to cite this model you can use this:

?

Contributions

Thanks to @sarasadeghii and @sadrasabouri for adding this model.