기본 EfficientNet B0~B7까지 다룸. V2 등등은 공부예정EfficientNet이란?이 글은 논문 " EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks " 을 참고하여 작성됨.Baseline: The original network without scaling.Width Scaling: Increasing the number of channels in each layer.Depth Scaling: Increasing the number of layers.Resolution Scaling: Increasing the input image resolution.Compound Scaling: Simultaneously ..
