Wav2vec2 Large Xlsr Moroccan Darija

🚀 Wav2Vec2-Large-XLSR-53-Moroccan-Darija

A new fine - tuned wav2vec2 - large - xlsr - 53 model for Moroccan Darija speech recognition

This model is fine - tuned on 57 hours of labeled Darija Audios. The data is extracted from MDVC (https://ijeecs.iaescore.com/index.php/IJEECS/article/view/35709), which contains over 1000 hours of Moroccan Darija "ary". Fine - tuning is running 24/7 to improve accuracy, and data is continuously added to the model daily. The team prefers not to add the entire MDVC Corpus at once to standardize the writing of Moroccan Darija.

Model Index

• Name: XLSR Wav2Vec2 Moroccan Arabic dialect by Boumehdi
• Results:
  • Task:
    • Name: Speech Recognition
    • Type: automatic - speech - recognition
  • Metrics:
    • Name: Test WER
    • Type: wer
    • Value: 0.084904

🚀 Quick Start

The model can be used directly as follows:

💻 Usage Examples

Basic Usage

import librosa
import torch
from transformers import Wav2Vec2CTCTokenizer, Wav2Vec2ForCTC, Wav2Vec2Processor, TrainingArguments, Wav2Vec2FeatureExtractor, Trainer

tokenizer = Wav2Vec2CTCTokenizer("./vocab.json", unk_token="[UNK]", pad_token="[PAD]", word_delimiter_token="|")
processor = Wav2Vec2Processor.from_pretrained('boumehdi/wav2vec2-large-xlsr-moroccan-darija', tokenizer=tokenizer)
model=Wav2Vec2ForCTC.from_pretrained('boumehdi/wav2vec2-large-xlsr-moroccan-darija')


# load the audio data (use your own wav file here!)
input_audio, sr = librosa.load('file.wav', sr=16000)

# tokenize
input_values = processor(input_audio, return_tensors="pt", padding=True).input_values

# retrieve logits
logits = model(input_values).logits

tokens = torch.argmax(logits, axis=-1)

# decode using n-gram
transcription = tokenizer.batch_decode(tokens)

# print the output
print(transcription)

Output: قالت ليا هاد السيد هادا ما كاينش بحالو

Contact Information:

• Email: souregh@gmail.com
• Author: BOUMEHDI Ahmed