W

Wav2vec2 Base 100h

Developed by vuiseng9
Wav2Vec2 base version speech recognition model trained on 100 hours of LibriSpeech data
Downloads 26
Release Time : 3/2/2022

Model Overview

This is an automatic speech recognition (ASR) model based on the Wav2Vec2 architecture, trained on 100 hours of English speech data from the LibriSpeech dataset, suitable for English speech-to-text tasks.

Model Features

Efficient Speech Recognition
Achieves a word error rate (WER) of 6.1 (clean) and 13.5 (other) on the LibriSpeech test set
Lightweight Base Model
Compared to larger-scale models, this 100-hour trained base version is more suitable for resource-constrained environments
Strong Compatibility
Verified compatible with transformers v4.15.0 and datasets 1.18.0 versions

Model Capabilities

English Speech Recognition
Audio to Text Conversion
Batch Speech Processing

Use Cases

Speech Transcription
Meeting Minutes Transcription
Automatically convert English meeting recordings into text transcripts
Achieves a 6.1% word error rate in clear speech environments
Educational Content Transcription
Convert English educational audio content into text
Achieves a 13.5% word error rate in complex speech environments
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase