W

Wav2vec2 Large Xlsr 53 Polish

Developed by jonatasgrosman
XLSR-53 large model speech recognition system optimized for Polish, fine-tuned based on facebook/wav2vec2-large-xlsr-53, supports Polish automatic speech recognition
Downloads 412.13k
Release Time : 3/2/2022

Model Overview

This is a Polish speech recognition model based on the XLSR-53 architecture, fine-tuned using the Common Voice 6.1 Polish dataset, suitable for Polish speech-to-text tasks.

Model Features

Polish Optimization
Specially fine-tuned for Polish, achieving a word error rate of 14.21% on the Common Voice Polish test set
Language Model Integration Support
Can be combined with a language model to further improve recognition accuracy, reducing the word error rate to 10.98%
Robust Speech Processing
Performs well on robust speech event datasets, capable of handling speech input in various environments

Model Capabilities

Polish speech recognition
Audio-to-text conversion
Supports 16kHz sample rate audio processing

Use Cases

Speech Transcription
Polish Speech Transcription
Convert Polish speech content into text
Word error rate of 14.21% and character error rate of 3.49% on the Common Voice test set
Voice Assistant
Polish Voice Command Recognition
Recognize and understand Polish voice commands
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase