A

Asr Whisper Medium Commonvoice Ar

Developed by speechbrain
A Whisper medium speech recognition model fine-tuned on the CommonVoice Arabic dataset, developed by the SpeechBrain team
Downloads 17
Release Time : 7/20/2023

Model Overview

This model is an automatic speech recognition system based on the Whisper medium architecture, specifically optimized for Arabic and fine-tuned on the CommonVoice Arabic dataset

Model Features

High-accuracy Arabic recognition
Achieves a WER of 14.82% on the CommonVoice Arabic test set
Based on Whisper architecture
Utilizes the OpenAI Whisper medium pre-trained model for fine-tuning
End-to-end training
Complete encoder-decoder architecture that directly outputs text results
Automatic audio processing
Built-in audio normalization (resampling + mono channel selection)

Model Capabilities

Arabic speech recognition
Audio transcription
16kHz mono audio processing

Use Cases

Speech transcription
Arabic speech-to-text
Convert Arabic speech content into text
Test set WER 14.82%, CER 4.95%
Voice assistants
Arabic voice command recognition
Front-end speech recognition module for Arabic voice assistants
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
Š 2025AIbase