W

WHISPER SMALL SWAHILI ASR CV 14

Developed by dmusingu
This model is a fine-tuned speech recognition model based on OpenAI's Whisper large on the Common Voice 14.0 Swahili (SW) dataset, achieving a word error rate (WER) of 25.13%.
Downloads 28
Release Time : 4/19/2024

Model Overview

An automatic speech recognition (ASR) model optimized for Swahili, fine-tuned based on the Whisper architecture, suitable for speech-to-text tasks.

Model Features

Low Word Error Rate
Achieves a word error rate (WER) of 25.13% on the Common Voice 14.0 Swahili test set
Based on Whisper Architecture
Fine-tuned on OpenAI's powerful Whisper-large model, inheriting its excellent speech recognition capabilities
Optimized Specifically for Swahili
Trained on the Common Voice 14.0 Swahili dataset for better recognition performance in this language

Model Capabilities

Speech-to-Text
Swahili Speech Recognition
Long Audio Processing

Use Cases

Speech Transcription
Swahili Speech Transcription
Convert Swahili speech content into text
Word error rate 25.13%, character error rate 9.83%
Voice Assistants
Swahili Voice Assistant
Provide voice interaction capabilities for Swahili users
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
Š 2025AIbase