AIbase
EN
Home
Afrispeech Large A100
A

Afrispeech Large A100

Developed by Seyfelislem
An African language speech recognition model fine-tuned from Whisper-large-v2, trained on the afrispeech-200 dataset with a word error rate (WER) of 14.81
Speech Recognition
Transformers
#African speech recognition#Low word error rate#Multilingual support
Downloads 20
Release Time : 4/3/2023

Model Overview

This model is an automatic speech recognition (ASR) system for African languages, obtained by fine-tuning the openai/whisper-large-v2 model, focusing on improving recognition accuracy for African languages.

Model Features

African language optimization
Specially fine-tuned for African languages, improving recognition accuracy for African languages
Low word error rate
Achieves a word error rate (WER) of 14.81% on the afrispeech-200 dataset
Based on Whisper architecture
Built on the powerful Whisper-large-v2 model, inheriting its excellent speech recognition capabilities

Model Capabilities

Speech recognition
African language processing
Automatic transcription

Use Cases

Speech-to-text
African language audio transcription
Automatically converts African language speech content into text
Word error rate 14.81%
Voice assistant
African language voice assistant
Supports voice interaction systems for African languages
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
EnglishįŽ€äŊ“中文įšéĢ”中文ãĢãģんご
Š 2025AIbase