metadata

language: ary
metrics:
  - wer
tags:
  - audio
  - automatic-speech-recognition
  - speech
  - xlsr-fine-tuning-week
license: apache-2.0
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: 0.172139

Wav2Vec2-Large-XLSR-53-Moroccan-Darija

wav2vec2-large-xlsr-53

Fine-tuned on 31 hours (31 people) of labeled Darija Audios.
Each hour of audio is pronounced by a different person.
Transcriptions are performed by a single individual.
Fine-tuning is ongoing 24/7 to enhance accuracy.
We are consistently adding more data to the model every day.
Audio database is organized (by sex, age, region, ..)

Training Loss	Validation	Loss Wer
0.018800	0.262561	0.172139

Usage

The model can be used directly as follows:

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

email: [email protected]

BOUMEHDI Ahmed