Training-free Neural Architecture Search through Variance of Knowledge of Deep Network Weights.

Ondrej Tbl, Luks Neumann

Browse the full CVPR paper archive.