Bengali.AI

is a non-profit initiative from Bangladesh that is focused on 1) crowdsourcing datasets 2) crowdsourcing solutions on Kaggle 3) open-sourcing!

[Bengali.AI @bengali_ai]

We have created OOD-Speech, the largest Bengali ASR dataset (1200+ hrs, 22,546+ unique voices) and first Out-of-Distribution Benchmarking dataset with Massively Crowdsourced (MaCro) scripted speech and 17 unique domains of spontaneous speech in test. Dataset+paper coming soon!

[Kaggle @kaggle]

📣 Competition launch alert! Bengali AI speech recognition competition hosted by Bengali AI, a non profit community.
🎯: to recognize Bengali speech from out-of-distribution audio recordings
💰: $53,000 prize pool
⏰: Oct 10, 2023, entry deadline
https://t.co/g15hynBj7R

[Bengali.AI @bengali_ai]

Bangla Unicode Normalizer

For example:
(a)'আরো'==(b)'আরো' -> False ; although they look exactly the same they are **NOT** same
(a) breaks as:['আ', 'র', 'ে', 'া']
(b) breaks as:['আ', 'র', 'ো']

https://t.co/54pliad2cR

[Bengali.AI @bengali_ai]

Shout out to @shohag_ansary @farigys @rezwan249 @Reasat8 @SadiMohammad_

[Bengali.AI @bengali_ai]

Grapheme Parser for indic languages.

Available languages: Bangla, Malyalam, Tamil, Gujrati, Panjabi, Odiya, Hindi.

https://t.co/l2QDcG1c6q

[Bengali.AI @bengali_ai]

IML GPU Support is live now!

Like https://t.co/2MSAAJfWcb, Intelligent Machines Limited(IML) is also working on promoting data science in bangladeshi students. So we hope this... https://t.co/uQvOdM8eer

[Bengali.AI @bengali_ai]

It's only first week and 200 teams milestone has been reached. Kagglers from different backgrounds are there committing and submitting. Happy modelling!
#bengali #bangladesh #kaggle

[Bengali.AI @bengali_ai]

Who says It's only first week. Lot's of submission going on https://t.co/Cxu5FSz05l

[Kaggle @kaggle]

Can you separately classify 3 constituent elements in a handwritten language? 🤔📝
Join the https://t.co/y95NjMvlMM research competition today: https://t.co/vZ8bmOjR9Q #kagglecompetition