W

Wav2vec2 Large Xlsr Portuguese

Developed by joaoalvarenga
A Portuguese automatic speech recognition model fine-tuned based on Facebook's wav2vec2-large-xlsr-53 architecture, trained on the Common Voice dataset with a word error rate of 13.77%.
Downloads 83
Release Time : 3/2/2022

Model Overview

This is an automatic speech recognition (ASR) model specifically optimized for Portuguese, based on the large-scale self-supervised pre-trained wav2vec2 architecture, suitable for Portuguese speech-to-text tasks.

Model Features

High-precision Portuguese recognition
Achieves a word error rate of 13.77% on the Common Voice Portuguese test set, demonstrating excellent performance.
Based on XLSR architecture
Utilizes a large-scale pre-trained model for cross-lingual speech representation, featuring powerful speech feature extraction capabilities.
No language model required
Can be used directly without additional language models for speech-to-text functionality.
Open-source license
Adopts the Apache-2.0 license, allowing for both commercial and research use.

Model Capabilities

Portuguese speech recognition
Audio-to-text conversion
Speech transcription

Use Cases

Speech transcription
Portuguese meeting minutes
Automatically converts Portuguese meeting recordings into text transcripts
Accuracy approximately 86.23%
Voice assistant
Provides speech recognition capabilities for Portuguese voice assistants
Education
Language learning applications
Helps learners practice Portuguese pronunciation and listening
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase