本文介绍了GoogleNet Inception架构，详细解释了22层深度的网络结构及其输入尺寸为224×224 RGB图像的要求。文中还讨论了1×1卷积核在减少参数数量方面的作用。
GoogleNet Inception 1
22 layers deep when counting only layers with parameters (or 27 layers if count pooling )
Inception model:
GoogLeNet network:
Input: 224×224in
the RGB color space with zero mean.
“#3×3reduce”
and “#5×5reduce”stands for the number of
1×1filters in the reduction layer used before the
3×3and
5×5convolutions .
扫一扫
1137
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
