【caffe】cifar10例子之quick_train.sh在windows下的解决方案

@tags caffe

照例还是转写为python脚本：

import os

caffe_root=os.environ['caffe_root']
caffe_build=os.environ['caffe_build']

cmd1=caffe_build+"\caffe.exe train --solver="+caffe_root+"\examples\cifar10\cifar10_quick_solver.prototxt"

print cmd1
os.system(cmd1)

# reduce learning rate by factor of 10 after 8 epochs
cmd2=caffe_build+"\caffe.exe train --solver="+caffe_root+"\examples\cifar10\cifar10_quick_solver_lr1.prototxt --snapshot="+caffe_root+"\examples\cifar10\cifar10_quick_iter_4000.solverstate.h5"

print cmd2
os.system(cmd2)