Open qclqcl opened 10 months ago
Same question here, any updates on this please ? (Thanks to google translate :-) the question is => "The network I tested was ConvTranspose+ BatchNormalization. But the BN layer did not disappear after simplification. Can the integration of ConvTranspose+BN be supported?")
我测试的网络是ConvTranspose+ BatchNormalization. 但是经过简化之后BN层没有消失。 请问可以支持ConvTranspose+BN的融合吗?