caffe学习系列(6):其他层介绍

主要包括softmax-loss层(与softmax有区别),全连接层(Inner Prouduct),accuracy层,reshape层,

Dropout层。

softmax:

layers {
  bottom: "cls3_fc"
  top: "prob"//概率似然值
  name: "prob"
  type: “Softmax"
}

accuracy

layer {
  name: "accuracy"
  type: "Accuracy"
  bottom: "ip2"
  bottom: "label"
  top: "accuracy"
  include {
    phase: TEST //test阶段才有
  }
}

 reshape

 layer { //改变输入的维度
    name: "reshape"
    type: "Reshape"
    bottom: "input"
    top: "output" 
    reshape_param {
      shape {//指定blob数据各维的值n*c*w*h
        dim: 0  # copy the dimension from below维度不变
        dim: 2  #将原来的维度变成2
        dim: 3
        dim: -1 # infer it from the other dimensions根据其他三维来计算当前维度
      }
    }
  }

Dropout层

layer {
  name: "drop7"
  type: "Dropout"
  bottom: "fc7-conv"
  top: "fc7-conv"
  dropout_param {
    dropout_ratio: 0.5
  }
}

参考:http://www.cnblogs.com/denny402/p/5072746.html

原文地址:https://www.cnblogs.com/573177885qq/p/5805013.html