Caffe-CIFAR10实验

本日志用于交流学习，如果问题，欢迎交流。

教程

1-实验步骤

cd $CAFFE_ROOT

下载数据

./data/cifar10/get_cifar10.sh

转换数据格式

./examples/cifar10/create_cifar10.sh

开始训练

./examples/cifar10/train_quick.sh

结束，将在

/home/yourname/caffe/examples/cifar10

下产生如下4个文件
cifar10_quick_iter_4000.caffemodel.h5
cifar10_quick_iter_4000.solverstate.h5
cifar10_quick_iter_5000.caffemodel.h5
cifar10_quick_iter_5000.caffemodel.h5

训练过程输出信息

输出信息

I1210 20:03:49.067945  3704 net.cpp:156] Memory required for data: 1229200I1210 20:03:49.067953  3704 layer_factory.hpp:77] Creating layer conv1I1210 20:03:49.067976  3704 net.cpp:91] Creating Layer conv1I1210 20:03:49.067983  3704 net.cpp:425] conv1 <- dataI1210 20:03:49.067996  3704 net.cpp:399] conv1 -> conv1

网络初始化

I0317 21:52:49.309370 2008298256 net.cpp:166] Network initialization done.I0317 21:52:49.309376 2008298256 net.cpp:167] Memory required for Data 23790808I0317 21:52:49.309422 2008298256 solver.cpp:36] Solver scaffolding done.I0317 21:52:49.309447 2008298256 solver.cpp:47] Solving CIFAR10_quick_train

训练参数

I0317 21:53:12.179772 2008298256 solver.cpp:208] Iteration 100, lr = 0.001I0317 21:53:12.185698 2008298256 solver.cpp:65] Iteration 100, loss = 1.73643...I0317 21:54:41.150030 2008298256 solver.cpp:87] Iteration 500, Testing netI0317 21:54:47.129461 2008298256 solver.cpp:114] Test score #0: 0.5504I0317 21:54:47.129500 2008298256 solver.cpp:114] Test score #1: 1.27805

输出结束

I0317 22:12:19.666914 2008298256 solver.cpp:87] Iteration 5000, Testing netI0317 22:12:25.580330 2008298256 solver.cpp:114] Test score #0: 0.7533I0317 22:12:25.580379 2008298256 solver.cpp:114] Test score #1: 0.739837I0317 22:12:25.587262 2008298256 solver.cpp:130] Snapshotting to cifar10_quick_iter_5000I0317 22:12:25.590215 2008298256 solver.cpp:137] Snapshotting solver state to cifar10_quick_iter_5000.solverstateI0317 22:12:25.592813 2008298256 solver.cpp:81] Optimization Done.

2-学习曲线的绘制

将训练的输出日志文件重定向到cifar.log文件

.examples/cifar.train_quick.sh >& cifar.log &

如果要想中途查看输出情况用下面这条指令，如果退出用Ｃtrl+C

利用awk将损失值提取出来

cat cifar.log |grep “Train net output” | awk ‘{print $11}’

tail -f cifar.log

损失函数曲线如下所示：

用Python画出模型结构图如下图所示：