1 paper · 1 filter
Jiefu Zhang, Haixiang Sun, Yang Xu +2
Pretrained macro-placement policies can reduce repeated optimization across circuits, but deployment often exposes them to unfamiliar designs when the original training data are un…