wav2vec2-common_voice-nl-demo Open-source Model - Free Implementation of Dutch Automatic Speech Recognition

Wav2vec2 Common Voice Nl Demo

Developed by MatsUy

This is an automatic speech recognition (ASR) model fine-tuned on the Dutch COMMON_VOICE dataset based on the facebook/wav2vec2-large-xlsr-53 model.

Speech Recognition

Transformers

OtherOpen Source License:Apache-2.0 #Dutch speech recognition #Low word error rate #Multi-scenario speech transcription

Downloads 16

Release Time : 3/2/2022

Model Overview

This model is specifically designed for Dutch speech recognition tasks, trained on the COMMON_VOICE-NL dataset, capable of converting Dutch speech into text.

Model Features

Dutch speech recognition

Speech recognition capabilities optimized specifically for Dutch

Based on wav2vec2 architecture

Uses facebook's wav2vec2-large-xlsr-53 as the base model

Fine-tuned on COMMON_VOICE dataset

Trained using high-quality Dutch speech datasets

Model Capabilities

Dutch speech-to-text

Automatic speech recognition

Use Cases

Speech transcription

Dutch meeting minutes

Automatically transcribe Dutch meeting recordings into written records

Word error rate (WER) approximately 20.46%

Voice assistant

Provide speech recognition capabilities for Dutch voice assistants

Training Loss	Epoch	Step	Validation Loss	Wer
3.0536	1.12	500	0.5349	0.4338
0.2543	2.24	1000	0.3859	0.3029
0.1472	3.36	1500	0.3471	0.2818
0.1088	4.47	2000	0.3489	0.2731
0.0855	5.59	2500	0.3582	0.2558
0.0721	6.71	3000	0.3457	0.2471
0.0653	7.83	3500	0.3299	0.2357
0.0527	8.95	4000	0.3440	0.2334
0.0444	10.07	4500	0.3417	0.2289
0.0404	11.19	5000	0.3691	0.2204
0.0345	12.3	5500	0.3453	0.2102
0.0288	13.42	6000	0.3634	0.2089
0.027	14.54	6500	0.3532	0.2044

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご

Wav2vec2 Common Voice Nl Demo

Model Overview

Model Features

Model Capabilities

Use Cases

🚀 wav2vec2-common_voice-nl-demo

🚀 Quick Start

📚 Documentation

Training and Evaluation Data

Training Procedure

Training Hyperparameters

Training Results

Framework Versions

📄 License