W

Wav2vec2 Large Xls R 1b Nl Lm

Developed by RuudVelo
An automatic speech recognition model fine-tuned on the Dutch Common Voice 8 dataset based on facebook/wav2vec2-xls-r-1b, combined with a KenLM 5-gram language model
Downloads 21
Release Time : 3/2/2022

Model Overview

This is a model for Dutch automatic speech recognition (ASR), fine-tuned on the Common Voice 8 dataset with an added language model to improve recognition accuracy.

Model Features

High-performance Dutch recognition
Achieves a 9.73% word error rate (WER) on the Common Voice 8 test set
Integrated language model
Includes a KenLM 5-gram language model to improve recognition accuracy
Large-scale pre-training
Fine-tuned from the 1-billion parameter wav2vec2-XLS-R-1B model

Model Capabilities

Dutch speech recognition
Speech-to-text
Dialogue transcription

Use Cases

Speech transcription
Dutch speech-to-text
Convert Dutch speech content into text
9.73% WER on Common Voice 8 test set
Dialogue analysis
Dutch dialogue transcription
Transcribe Dutch dialogue content
27.67% WER on robust speech event test data
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
Š 2025AIbase