mohamedaminelayachi/North

North is a multimodal classifier, trained on North-African dialects to determine to which country the given text belongs. North is a multimodal classifier, trained on North-African dialects to determine to which country the given sentence belongs. Developed by **Mohamed Amine Layachi & Otmane El Bourki** Models used: - Multinomial Naive Bayes for Latin Sequences Classification - BERT for Arabic Sequences Classification Datasets used: - North Latin Dataset -- We collected it by web-scraping data from various places, one of which is DODA. - North Arabic Dataset -- We collected it by web-scraping data from various places. The supported dialects: - Moroccan (Arabic, Latin) - Tunisian (Arabic, Latin) - Algerian (Arabic) - Egyptian (Arabic) **Please note that the main notebook you might want to interact with is _'North - Dialectal Variance Classifier.ipynb'_. All the other files were used just for training purposes (Preprocessing is integrated).** ## An Example - Enter your text - Select the alphabet (Latin or Arabic). - Check the classification