编辑一个write_net.py文件进行生成,代码如下:
import sys
caffe_root = '/home/yourname/caffe/'
sys.path.insert(0,caffe_root+'python')
from caffe import layers as L
from caffe import params as P
import caffe
#有两个参数,分别是数据源和batch_size
def lenet(lmdb,batch_size):
#数据层的设置
n.data,n.label=L.Data(batch_size=batch_size,backend=P.Data.LMDB,source=lmdb,transform_param=dict(scale=1./255),ntop=2)
n.conv1 = L.Convolution(n.data,kenel_size=5,num_output=20,weight_filterdict(type='xavier'))
n.pool1 = L.Pooling(n.conv1,kernel_size=2,stride=2,pool=P.Pooling.MAX)
n.conv1 = L.Convolution(n.conv1,kenel_size=5,num_output=50,weight_filterdict(type='xavier'))
n.pool1 = L.Pooling(n.conv2,kernel_size=2,stride=2,pool=P.Pooling.MAX)
n.ip1 = L.InnerProduct(n.conv2,num_output=500,weight_filterdict(type='xavier'))
n.relu1 = L.ReLU(n.ip1,in_place=True)
n.ip2 = L.InnerProduct(n.relu1,num_output=10,weight_filterdict(type='xavier'))
n.loss = L.SoftmaxWithLoss(n.ip2,n.label)
#让n以prototxt的形式返回
return n.to_proto()
#分别生成两个网络模型(训练和测试)的prototxt文件
with open('lenet_auti_train.prototxt','w') as f:
f.write(str(lenet('example/mnist/mnist_train_ldmb',64)))
with open('lenet_auto_test.prototxt','w') as f:
f.write(str(lenet('example/mnist/mnist_test_lmdb',100)))