In this project, I focused on enhancing contextual recognition in ASR systems using semantic lattice rescoring. During the exploratory phase, we initially aimed to use a conformer model as both a language and acoustic model. While this innovative approach did not materialize as planned, it led to valuable insights and a refined implementation. The project repository now features conformer model implementations specifically tailored for the LJ Speech and TED-LIUM datasets, demonstrating a solid approach to advancing ASR technology. *This was an experimental portion of my Lattice Rescoing paper.
-
Notifications
You must be signed in to change notification settings - Fork 0
ankithasudarshan/Lattice-Project-Conformers
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Repository files navigation
About
Exploratory part of the implementation of the paper titled - "Improving Contextual Recognition in Automatic Speech Recognition Systems using Semantic Lattice Rescoring"
Topics
Resources
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published