GPU下caffe训练cifar10

一、训练
1.获取训练数据
cd caffe
./data/cifar10/get_cifar10.sh
./examples/cifar10/create_cifar10.sh
2.开始训练
cd caffe
./examples/cifar10/train
*说明：先以0.001的学习率迭代4000次，再以0.01的学习率接着再迭代1000次，共5000次。训练后生成模型文件：
cifar10_quick_iter_4000.caffemodel.h5，
cifar10_quick_iter_4000.solverstate.h5，
cifar10_quick_iter_5000.caffemodel.h5，
cifar10_quick_iter_5000.solcerstate.h5
二、测试
使用caffe/python/classify.py进行测试
python python/classify.py –model_def examples/cifar10/cifar10_quick.prototxt –pretrained_model examples/cifar10/cifar10_quick_iter_4000.caffemodel.h5 –center_only examples/images/cat.jpg foo
*报错：解决方案见http://blog.csdn.net/fangjin_kl/article/details/50936925