wav2vec2-xls-r-300m-ar Open-source Speech Recognition Model - Accurately Recognize Arabic Speech Content

Wav2vec2 Xls R 300m Ar

Developed by nouamanetazi

This model is an automatic speech recognition model fine-tuned on the Common Voice Arabic dataset based on facebook/wav2vec2-xls-r-300m.

Speech Recognition

Transformers

ArabicOpen Source License:Apache-2.0 #Arabic speech recognition #Common Voice dataset #XLS-R architecture

Downloads 19

Release Time : 3/2/2022

Model Overview

This is an automatic speech recognition model optimized for Arabic, fine-tuned based on the XLS-R-300M architecture, suitable for Arabic speech-to-text tasks.

Model Features

Arabic optimization

Specially fine-tuned for Arabic speech, improving recognition accuracy for Arabic speech.

Based on XLS-R architecture

Uses Facebook's XLS-R-300M pre-trained model as the foundation, with powerful speech feature extraction capabilities.

Trained on Common Voice dataset

Trained using the Common Voice Arabic dataset, covering diverse speech scenarios.

Model Capabilities

Arabic speech recognition

Speech-to-text

Long audio processing (supports chunk processing)

Use Cases

Speech transcription

Arabic speech transcription

Convert Arabic speech content into text

Voice assistants

Arabic voice command recognition

Used for Arabic voice assistants or control system voice command recognition

🚀 wav2vec2-xls-r-300m-ar

This model fine - tunes facebook/wav2vec2-xls-r-300m on the COMMON_VOICE - AR dataset, aiming to solve the problem of automatic speech recognition in Arabic and providing high - quality speech - to - text conversion services.

🚀 Quick Start

To evaluate on speech-recognition-community-v2/dev_data, use the following command:

python eval.py --model_id nouamanetazi/wav2vec2-xls-r-300m-ar --dataset speech-recognition-community-v2/dev_data --config ar --split validation --chunk_length_s 5.0 --stride_length_s 1.0

✨ Features

This model is a fine - tuned version of facebook/wav2vec2-xls-r-300m on the COMMON_VOICE - AR dataset. It achieves the following results on the evaluation set:

eval_loss: 3.0191
eval_wer: 1.0
eval_runtime: 252.2389
eval_samples_per_second: 30.217
eval_steps_per_second: 0.476
epoch: 1.0
step: 340

📚 Documentation

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 0.0005
train_batch_size: 64
eval_batch_size: 64
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e - 08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 2000
num_epochs: 5
mixed_precision_training: Native AMP

Framework versions

Transformers 4.17.0.dev0
Pytorch 1.10.2+cu102
Datasets 1.18.2.dev0
Tokenizers 0.11.0

🔧 Technical Details

Property	Details
Language	Arabic
License	Apache - 2.0
Tags	ar, automatic - speech - recognition, common_voice, generated_from_trainer, hf - asr - leaderboard, robust - speech - event
Datasets	common_voice
Model Name	XLS - R - 300M - Arabic
Evaluation Metrics	On the Robust Speech Event - Dev Data (ar): Test WER = 1.0, Test CER = 1.0; On the evaluation set: eval_loss = 3.0191, eval_wer = 1.0, eval_runtime = 252.2389, eval_samples_per_second = 30.217, eval_steps_per_second = 0.476, epoch = 1.0, step = 340

📄 License

This model is released under the Apache - 2.0 license.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご