Skip to content

Translating English to Malayalam and vice versa using pretrained Helsinki Transformer models.

Notifications You must be signed in to change notification settings

Midhun134/Bilingual-Language-Translator-using-Raspberry-Pi5

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Bilingual-Language-Translator-using-Raspberry-Pi5

This project aims to develop a bilingual headphone translator capable of translating between English and Malayalam in real-time, leveraging the power of the Raspberry Pi 5 as the operating system. The device is designed to provide seamless communication for speakers of these two languages, addressing the challenges of language barriers in various social, educational, and professional settings. The Raspberry Pi 5, with its enhanced processing power, serves as the core of the system, managing both speech recognition and translation tasks efficiently.

The translation process involves multiple stages: capturing the spoken input via a microphone, processing the audio to text using advanced speech recognition algorithms, translating the text from English to Malayalam or vice versa using natural language processing (NLP) models, and finally converting the translated text back to speech for output simultaneously through both the headphone and speaker. The system utilizes in-built python libraries for speech-to-text and text-to-speech conversion, ensuring high accuracy and natural-sounding audio output.

There are two main code files both in python and one is for audio outputting through windows laptop while the other linux file is for outputting audio through the raspberry pi5. The other folder contains files for finetuning the models and the parallel corpora datasets used for finetuning the Helsinki models...

About

Translating English to Malayalam and vice versa using pretrained Helsinki Transformer models.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published