caffe 网络结构参数介绍及可视化

caffe/examples/mnist/lenet_solver.prototxt

# The train/test net protocol buffer definition
net: "examples/mnist/lenet_train_test.prototxt"
//用户训练/预测的网络描述文件(ProtoBuffer文本格式)
# test_iter specifies how many forward passes the test should carry out.
//预测阶段迭代次数。在mnist例程下,预测样本组(test batch)大小为100
# In the case of MNIST, we have test batch size 100 and 100 test iterations,
# covering the full 10,000 testing images.
//设置预测阶段迭代次数为100可以覆盖全部10000个测试样本
test_iter: 100
# Carry out testing every 500 training iterations.
//训练时每迭代500次,进行一次预测
test_interval: 500
# The base learning rate, momentum and the weight decay of the network.
网络的基础学习速率、冲量和权衰量
base_lr: 0.01
momentum: 0.9
weight_decay: 0.0005
# The learning rate policy
lr_policy: "inv"
//学习速率的衰减策略
gamma: 0.0001
power: 0.75
# Display every 100 iterations
//每经过100次迭代,在屏幕上打印一次log
display: 100
# The maximum number of iterations
//最大迭代次数
max_iter: 10000
# snapshot intermediate results
//每500次打印一次快照
snapshot: 5000
snapshot_prefix: "examples/mnist/lenet"
# solver mode: CPU or GPU
//求解模式CPU或者GPU
solver_mode: CPU

caffe/examples/mnist/lenet_train_test.prototxt

name: "LeNet" //网络的名称
layer {       //定义一个层(Layer)
  name: "mnist" //层名称
  type: "Data"  //层类型:数据层
  top: "data"  //层输出 :data和label
  top: "label"
  include {
    phase: TRAIN //本层只在训练阶段有效
  }
  transform_param {
    scale: 0.00390625 //数据变换使用的数据缩放因子
  }
  data_param { //数据层参数
    source: "examples/mnist/mnist_train_lmdb" //lmdb数据源路径
    batch_size: 64 //批次大小,一次读取64张图
    backend: LMDB //数据格式
  }
}
layer { //一个新的数据层,名字也叫mnist,输出blog也是data和label,但是这里定义的参数只在分类阶段有效
  name: "mnist"
  type: "Data"
  top: "data"
  top: "label"
  include {
    phase: TEST
  }
  transform_param {
    scale: 0.00390625
  }
  data_param {
    source: "examples/mnist/mnist_test_lmdb"
    batch_size: 100
    backend: LMDB
  }
}
layer { //卷积层conv1,输入blog为data,输出blog为conv1
  name: "conv1"
  type: "Convolution"
  bottom: "data"
  top: "conv1"
  param {
    lr_mult: 1 //权值学习速率倍乘因子,1表示与全局参数保持一致
  }
  param {
    lr_mult: 2 //bias学习速率倍乘因子,是全局参数的2倍
  }
  convolution_param { //卷积计算参数
    num_output: 20 //输出feature map数目为20
    kernel_size: 5 //卷积核尺寸
    stride: 1 //卷积步长
    weight_filler { //权值初始化策略
      type: "xavier"
    }
    bias_filler { //bias初始化策略
      type: "constant"
    }
  }
}
layer { //定义下采样层pool1,输入blob为conv1,输出blob为pool1
  name: "pool1"
  type: "Pooling"
  bottom: "conv1"
  top: "pool1"
  pooling_param { //下采样参数
    pool: MAX //下采样方法
    kernel_size: 2 //下采样窗口尺寸
    stride: 2 //下采样窗口步长
  }
}
layer {
  name: "conv2"
  type: "Convolution"
  bottom: "pool1"
  top: "conv2"
  param {
    lr_mult: 1
  }
  param {
    lr_mult: 2
  }
  convolution_param {
    num_output: 50
    kernel_size: 5
    stride: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}
layer {
  name: "pool2"
  type: "Pooling"
  bottom: "conv2"
  top: "pool2"
  pooling_param {
    pool: MAX
    kernel_size: 2
    stride: 2
  }
}
layer { //定义全连接层,输入blob为pool2,输出blog为ip1
  name: "ip1"
  type: "InnerProduct"
  bottom: "pool2"
  top: "ip1"
  param {
    lr_mult: 1
  }
  param {
    lr_mult: 2
  }
  inner_product_param { //全连接层参数
    num_output: 500
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}
layer { //定义relu非线性层
  name: "relu1"
  type: "ReLU"
  bottom: "ip1"
  top: "ip1"
}
layer {
  name: "ip2"
  type: "InnerProduct"
  bottom: "ip1"
  top: "ip2"
  param {
    lr_mult: 1
  }
  param {
    lr_mult: 2
  }
  inner_product_param {
    num_output: 10
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}
layer { //分类准确率层,只在test阶段有效,该层用于计算分类准确率
  name: "accuracy"
  type: "Accuracy"
  bottom: "ip2"
  bottom: "label"
  top: "accuracy"
  include {
    phase: TEST
  }
}
layer { //损失层,损失函数为SoftmaxLoss,输入blob为ip2和label,输出blob为loss
  name: "loss"
  type: "SoftmaxWithLoss"
  bottom: "ip2"
  bottom: "label"
  top: "loss"
}

lenet_train_test.prototxt 可视化

可视化工具网页:
http://ethereon.github.io/netscope/#/editor
在左边输入prototxt文本,按下shift+Enter,右侧会输出网络可视化图像
caffe 网络结构参数介绍及可视化_第1张图片

你可能感兴趣的:(深度学习,caffe)