tion at the same layer (R1, R2, R3), and showed better performance than baselines and SE-Nets on image classification, 2

Open in new window