zealbell/Yoru.A.I

A DeepLearning4J project with a fiercely trained model that decodes the curves and tones of the Yoruba alphabet — one of Africa’s most sxually syllabic* and culturally loaded writing systems. # Yoru.A.I. A DeepLearning4J model trained to recognize the alphabets of one of Africa’s boldest and baddest languages: Yoruba. Yes, Yoruba—a language so rich, expressive, and… unapologetically razz—that it somehow manages to be both chaotic and irresistibly s*xy. To back this up, we present: the global phenomenon of Yoruba demonds, smooth-talking heartbreakers with PhDs in charm and zero regard for emotional stability. Google them at your own risk. Contact them even less. Especially the ones based in Nigeria—one text, and your peace of mind is over. Now that the warnings are out of the way… let’s get neural. Let’s get Yoruba. Let’s get dirty ## The S\*xy Alphabets These aren’t just letters. In the hands of a Yoruba demond, they become the Yoruphalbet—an alphabet so smooth it comes with a warning: protect your girl ## Quick Links - training data - audio files > Despite the wildly unverifiable truths shared above, the actual goal of this library is to demonstrate > how to train a Convolutional Neural Network model with pictures from scratch using the Deeplearning4j framework.