(Caffe,Lenet5)初始化测试网络(四)

本文地址:http://blog.csdn.net/mounty_fsc/article/details/51090698

在 (Caffe,Lenet5)初始化训练网络(三)中介绍了训练网络的初始化,本文介绍测试网络的初始化。

1 测试网络结构

Layer layer Type Bottom Blob Top Blob Top Blob Shape
minst Data data
&&
label
100 1 28 28 (78400)
&&
100 (100)
label_mnist
_1_split
Split label label_mnist_1_split_0
&&
label_mnist_1_split_1
100 (100)
&&
100 (100)
conv1 Convolution data conv1 100 20 24 24 (1152000)
pool1 Pooling conv1 pool1 100 20 12 12 (288000)
conv2 Convolution pool1 conv2 100 50 8 8 (320000)
pool2 Pooling conv2 pool2 100 50 4 4 (80000)
ip1 InnerProduct pool2 ip1 100 500 (50000)
relu1 ReLU ip1 ip1(in-place) 100 500 (50000)
ip2 InnerProduct ip1 ip2 100 10 (1000)
ip2_ip2_0_split Split ip2 ip2_ip2_0_split_0
&&
ip2_ip2_0_split_1
100 10 (1000)
&&
100 10 (1000)
accuracy Accuracy ip2_ip2_0_split_0
&&
label_mnist_1_split_0
accuracy (1)
loss SoftmaxWithLoss ip2_ip2_0_split_1
&&
label_mnist_1_split_1
loss (1)

注:Top Blob Shape格式为:BatchSize,ChannelSize,Height,Width(Total Count)

2 与训练网络对比

  1. 训练网络9层,测试网络12层
  2. 训练网络没有的是:多了label_mnist_1_split、layer ip2_ip2_0_split、accuracy,类型为Split LayerAccuracy Layer

3 Split Layer

4 Accuracy Layer

你可能感兴趣的:(caffe)