F

FYP ARABIZI

Developed by ali-issa
This model is a speech recognition model fine-tuned on an unknown dataset based on facebook/wav2vec2-large-xlsr-53, supporting recognition of Arabic dialects (Arabizi).
Downloads 33
Release Time : 5/12/2022

Model Overview

This is a speech recognition model optimized for Arabic dialects (Arabizi), fine-tuned on the wav2vec2-large-xlsr-53 architecture, achieving a word error rate of 0.4042 on the evaluation set.

Model Features

Arabic Dialect Optimization
Speech recognition model specifically optimized for Arabizi (Arabic Latin alphabet transcription)
Fine-tuned on wav2vec2-large
Fine-tuned on the powerful wav2vec2-large-xlsr-53, inheriting its excellent speech feature extraction capabilities
Relatively Low Word Error Rate
Achieved a word error rate of 0.4042 on the evaluation set, indicating good recognition accuracy

Model Capabilities

Arabic dialect speech recognition
Speech to text
GPU acceleration support

Use Cases

Speech transcription
Arabic dialect speech transcription
Convert Arabic dialect speech content into text
Word error rate 0.4042
Voice assistant
Arabic dialect voice command recognition
Used to recognize Arabic dialect voice commands
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
Š 2025AIbase