模型压缩(4) - SqueezeNet

SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size 论文地址: A ...