Comment by rocauc

Purpose-built, small, and fast models appears to be the inevitable evolution for computer vision.

Where can the "Easy Set, "Medium Set, and "Hard Set" evaluations referenced in the "Wider Face Val" be found?