Counterintuitive Deep LearningDouble DescentDouble Descent is a phenomenon in which the test error of a model first decreases, then increases, and then decreases again as the model complexity increases.
Funny ProjectsThread EverythingA project that aims to make communications between different machines easier.
