W

Wav2vec2 Large Xls R 1b Cv8 Mt

Developed by RuudVelo
An automatic speech recognition model fine-tuned on the Common Voice 8 Maltese dataset based on facebook/wav2vec2-xls-r-1b
Downloads 17
Release Time : 3/2/2022

Model Overview

This is an automatic speech recognition (ASR) model optimized for Maltese, based on Facebook's large-scale wav2vec2 XLS-R architecture and fine-tuned on the Common Voice 8 dataset.

Model Features

High-performance Maltese recognition
Achieves a word error rate (WER) of 17.57% and a character error rate (CER) of 3.86% on the Common Voice 8 Maltese test set
Based on large-scale pretrained model
Fine-tuned from the facebook/wav2vec2-xls-r-1b model, inheriting its powerful speech feature extraction capabilities
Robust speech processing
The model is optimized for various speech environments and can handle multiple speech scenarios

Model Capabilities

Maltese speech recognition
Speech-to-text
Robust speech event detection

Use Cases

Speech transcription
Maltese speech transcription
Convert Maltese speech content into text
Word error rate 17.57%, character error rate 3.86%
Voice assistants
Maltese voice interaction
Provide speech recognition capabilities for Maltese voice assistants
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
Š 2025AIbase