SENTINEL | Dynamics of Gradient Descent and its loss landscape | Deep Learning

SENTINEL visualizes the optimization process of a convnet in training mode, moving from a high loss value through the creation of an edge horizon to the final convexity and minimum. More details and related analysis about this and other visualizations will be published in the future. Sentinel is a loss Landscape generated with real data: convnet, imagenette dataset, sgd-adam, bs=16, bn, lr sched, train mod, 100k pts, 0.5 w range, 20p-interp, log scaled (orig loss nums) & vis-adapted, trained with the awes