Skip to content

This code was used to perform Language recognition using phonotactic information obtained from speech signals

Notifications You must be signed in to change notification settings

Daromog/Transformer-for-LRE

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 
 
 
 
 

Repository files navigation

Transformer-for-LRE

This code was used to perform Language recognition using phonotactic information obtained from speech signals

The transformer encoder implements sliding attention windows proposed by "LongFormer" and "BigBird" to manage long input sequences. We take advantage of the attention mechanism to find discriminative combination of characters to differentiate between similar languages. This model was fused with a pure acoustic system based on MFCC-SDC-iVectors.

About

This code was used to perform Language recognition using phonotactic information obtained from speech signals

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages