AIbase
EN
Home
Wav2vec2 Lar Xlsr Es Col
W

Wav2vec2 Lar Xlsr Es Col

Developed by Santiagot1105
This model is a fine-tuned version based on jonatasgrosman/wav2vec2-large-xlsr-53-spanish for Spanish speech recognition tasks.
Speech Recognition
Transformers
Open Source License:Apache-2.0#Spanish speech recognition#Low word error rate#XLSR fine-tuning
Downloads 26
Release Time : 3/2/2022

Model Overview

This is a speech recognition model optimized for Spanish, fine-tuned on the wav2vec2-large-xlsr-53 architecture, achieving a word error rate of 18.84% on the evaluation set.

Model Features

Low word error rate
Achieved an 18.84% word error rate on the evaluation set, demonstrating excellent performance.
Based on XLSR architecture
Utilizes the cross-lingual speech representation learning (XLSR) architecture, which has powerful speech feature extraction capabilities.
Optimized for Spanish
Specifically fine-tuned and optimized for Spanish.

Model Capabilities

Spanish speech recognition
Audio to text conversion

Use Cases

Speech transcription
Automatic meeting transcription
Automatically convert Spanish meeting recordings into text transcripts
Word error rate approximately 18.84%
Voice assistant
Used as the speech recognition module for Spanish voice assistants
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
EnglishįŽ€äŊ“中文įšéĢ”中文ãĢãģんご
Š 2025AIbase