mnist测试
第一步:进入caffe目录
第二步:获取mnist数据集
./data/mnist/get_mnist.sh
第三步:创建lmdb
./examples/mnist/create_mnist.sh
第四步:训练
./examples/mnist/train_lenet.sh
运行完结果如下:
生成四个文件
lenet_iter_10000.caffemodel
lenet_iter_10000.solverstate
lenet_iter_5000.caffemodel
lenet_iter_5000.solverstate
屏幕显示:
I0918 14:43:05.602397 8647 solver.cpp:317] Iteration 10000, loss = 0.00291481
I0918 14:43:05.602445 8647 solver.cpp:337] Iteration 10000, Testing net (#0)
I0918 14:43:11.950065 8647 solver.cpp:404] Test net output #0: accuracy = 0.9904
I0918 14:43:11.950124 8647 solver.cpp:404] Test net output #1: loss = 0.0309935 (* 1 = 0.0309935 loss)
I0918 14:43:11.950135 8647 solver.cpp:322] Optimization Done.
I0918 14:43:11.950142 8647 caffe.cpp:254] Optimization Done.