Глубокое обучение (Deep Learning)
Дюша Грицевский; 16+. Class will be taught in English.
Дюша Грицевский; 16+. Class will be taught in English.
Prerequisites: linear algebra, multivariable calculus
If we show a computer a cat and a dog, will it dream of demonic dog-cat hybrids? How do we automatically translate from any language to any other language? How does a robot learn how to walk by only sensing hunger and pain?
Come join this course for a mathematical treatment of the above questions, complete with interactive exercises that will let you generate your own dolphin-cats, turn horses into zebras, and build your own basic GPT-2 model. (Powerful computers not needed, we'll train everything on Google's servers for free.)
Dataset Sample
Дополнительные материалы
Дополнительные материалы
UToronto CSC413/2516 Neural Networks and Deep Learning (Winter 2020): https://csc413-2020.github.io/
Видеозапись
Видеозапись
Материалы
Материалы
Google Colaboratory notebooks: