Caffe基礎（二）-使用命令行方式訓練預測mnist、cifar10及自己的數據集

在win10 vs2015 顯卡compute capability7.5 Python3.5.2環境下配置caffe及基本使用（一）介紹瞭如何編譯生成caffe工程及python、matlab接口。下面介紹通過命令行方式使用caffe訓練預測mnist數據集、訓練預測cifar10數據集，訓練預測自己的數據集。

(1) 訓練mnist數據集

在主目錄下的examples/minst文件夾下放入minist數據集

                               

數據集見鏈接

可以用如下兩個bat文件來做mnist數據集的轉換

如下是轉換爲leveldb數據格式

 

如下是轉換爲lmdb格式

                                            

在主目錄下新建一個bat文件my_add_mnist_run_train.bat(名字可以隨意定義），輸入如下內容：

打開lenet_solver.prototxt,可以看到

# The train/test net protocol buffer definition
net: "examples/mnist/lenet_train_test.prototxt"
# test_iter specifies how many forward passes the test should carry out.
# In the case of MNIST, we have test batch size 100 and 100 test iterations,
# covering the full 10,000 testing images.
test_iter: 100
# Carry out testing every 500 training iterations.
test_interval: 500
# The base learning rate, momentum and the weight decay of the network.
base_lr: 0.01
momentum: 0.9
weight_decay: 0.0005
# The learning rate policy
lr_policy: "inv"
gamma: 0.0001
power: 0.75
# Display every 100 iterations
display: 100
# The maximum number of iterations
max_iter: 10000
# snapshot intermediate results
snapshot: 5000
snapshot_prefix: "examples/mnist/lenet"
# solver mode: CPU or GPU
solver_mode: GPU

裏面有學習率，迭代步數，gpu還是使用cpu跑等配置

再打開lenet_train_test.prototxt文件，可以看到

可以將prototxt文件放在以下的網址，查看網絡的結構：http://ethereon.github.io/netscope/#/editor

name: "LeNet"
layer {
  name: "mnist"
  type: "Data"
  top: "data"
  top: "label"
  include {
    phase: TRAIN
  }
  transform_param {
    scale: 0.00390625
  }
  data_param {
    source: "examples/mnist/mnist_train_lmdb"
    batch_size: 64
    backend: LMDB
  }
}
layer {
  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 {
  name: "conv1"
  type: "Convolution"
  bottom: "data"
  top: "conv1"
  param {
    lr_mult: 1
  }
  param {
    lr_mult: 2
  }
  convolution_param {
    num_output: 20
    kernel_size: 5
    stride: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}
layer {
  name: "pool1"
  type: "Pooling"
  bottom: "conv1"
  top: "pool1"
  pooling_param {
    pool: MAX
    kernel_size: 2
    stride: 2
  }
}
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 {
  name: "ip1"
  type: "InnerProduct"
  bottom: "pool2"
  top: "ip1"
  param {
    lr_mult: 1
  }
  param {
    lr_mult: 2
  }
  inner_product_param {
    num_output: 500
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}
layer {
  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 {
  name: "accuracy"
  type: "Accuracy"
  bottom: "ip2"
  bottom: "label"
  top: "accuracy"
  include {
    phase: TEST
  }
}
layer {
  name: "loss"
  type: "SoftmaxWithLoss"
  bottom: "ip2"
  bottom: "label"
  top: "loss"
}

裏面有網絡結構、各層過濾器的大小，步長設置及數據集設置，有對應自己框架的數據結構

運行該訓練的bat文件，訓練完畢後結果如下：

（2）用訓練出來的模型測試mnist

首先計算mean.binaryproto

bat文件內容如下：

將lenet_train_test.prototxt複製一份出來，改名爲my_lenet_test.prototxt，取名隨意

對比lenet_train_test.prototxt，增加下面紅色標記處語句：

                                        

在主目錄下新建一個bat文件，文件裏內容爲：

運行該bat文件，結果如下：

                      

創建一個bat文件，對單張mnist圖片進行預測，文件里語句如下：

test image文件夾中裏的內容如下：

                                              

result.txt中的內容如下（這裏是caffe預測的結果爲一個10維向量，這裏預測的圖片是2，會得到該向量中最大值所處的索引值，憑該索引去result中找對應的類別標籤）：

     

                                                                

該bat文件的執行結果如下：

（3）訓練cifar10

首先轉化cifar10爲caffe能夠支持的數據集格式，這裏轉爲lmdb,原來的數據集格式如下：

                                                     

轉化的bat文件中的語句如下：

創建bat文件計算cifar10的mean.binaryproto文件

創建bat文件訓練網絡

cifar10_quick_solver.prototxt文件如下：

                               

cifar10_quick_train_test.prototxt文件如下：

name: "CIFAR10_quick"
layer {
  name: "cifar"
  type: "Data"
  top: "data"
  top: "label"
  include {
    phase: TRAIN
  }
  transform_param {
    mean_file: "./examples/cifar10/cifar_outputdata/mean.binaryproto"
  }
  data_param {
    source: "./examples/cifar10/cifar_outputdata/cifar10_train_lmdb"
    batch_size: 100
    backend: LMDB
  }
}
layer {
  name: "cifar"
  type: "Data"
  top: "data"
  top: "label"
  include {
    phase: TEST
  }
  transform_param {
    mean_file: "./examples/cifar10/cifar_outputdata/mean.binaryproto"
  }
  data_param {
    source: "./examples/cifar10/cifar_outputdata/cifar10_test_lmdb"
    batch_size: 100
    backend: LMDB
  }
}
layer {
  name: "conv1"
  type: "Convolution"
  bottom: "data"
  top: "conv1"
  param {
    lr_mult: 1
  }
  param {
    lr_mult: 2
  }
  convolution_param {
    num_output: 32
    pad: 2
    kernel_size: 5
    stride: 1
    weight_filler {
      type: "gaussian"
      std: 0.0001
    }
    bias_filler {
      type: "constant"
    }
  }
}
layer {
  name: "pool1"
  type: "Pooling"
  bottom: "conv1"
  top: "pool1"
  pooling_param {
    pool: MAX
    kernel_size: 3
    stride: 2
  }
}
layer {
  name: "relu1"
  type: "ReLU"
  bottom: "pool1"
  top: "pool1"
}
layer {
  name: "conv2"
  type: "Convolution"
  bottom: "pool1"
  top: "conv2"
  param {
    lr_mult: 1
  }
  param {
    lr_mult: 2
  }
  convolution_param {
    num_output: 32
    pad: 2
    kernel_size: 5
    stride: 1
    weight_filler {
      type: "gaussian"
      std: 0.01
    }
    bias_filler {
      type: "constant"
    }
  }
}
layer {
  name: "relu2"
  type: "ReLU"
  bottom: "conv2"
  top: "conv2"
}
layer {
  name: "pool2"
  type: "Pooling"
  bottom: "conv2"
  top: "pool2"
  pooling_param {
    pool: AVE
    kernel_size: 3
    stride: 2
  }
}
layer {
  name: "conv3"
  type: "Convolution"
  bottom: "pool2"
  top: "conv3"
  param {
    lr_mult: 1
  }
  param {
    lr_mult: 2
  }
  convolution_param {
    num_output: 64
    pad: 2
    kernel_size: 5
    stride: 1
    weight_filler {
      type: "gaussian"
      std: 0.01
    }
    bias_filler {
      type: "constant"
    }
  }
}
layer {
  name: "relu3"
  type: "ReLU"
  bottom: "conv3"
  top: "conv3"
}
layer {
  name: "pool3"
  type: "Pooling"
  bottom: "conv3"
  top: "pool3"
  pooling_param {
    pool: AVE
    kernel_size: 3
    stride: 2
  }
}
layer {
  name: "ip1"
  type: "InnerProduct"
  bottom: "pool3"
  top: "ip1"
  param {
    lr_mult: 1
  }
  param {
    lr_mult: 2
  }
  inner_product_param {
    num_output: 64
    weight_filler {
      type: "gaussian"
      std: 0.1
    }
    bias_filler {
      type: "constant"
    }
  }
}
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: "gaussian"
      std: 0.1
    }
    bias_filler {
      type: "constant"
    }
  }
}
layer {
  name: "accuracy"
  type: "Accuracy"
  bottom: "ip2"
  bottom: "label"
  top: "accuracy"
  include {
    phase: TEST
  }
}
layer {
  name: "loss"
  type: "SoftmaxWithLoss"
  bottom: "ip2"
  bottom: "label"
  top: "loss"
}

（4）命令行方式使用caffe預測一張cifar10圖片

如上mnist預測一樣，預測圖片如下：

                                                                              

同樣需要類別標籤名文件synset_words.txt，供caffe的結果去索引，該文件內容如下：

                                                                      

預測bat文件如下：

運行結果如下：

 

（5）訓練自己的數據集

在根目錄下data文件夾下新建mydata文件夾，需要訓練的數據集會存放在該文件夾內。

新建一個train和val文件夾,每個文件夾內如下：

                                                       

cat文件夾內圖片如下：

dog文件夾內圖片如下：

同時在mydata文件夾內新建train.txt和val.txt，以記錄這些圖片的路徑

                                                                   

                                                                     

 

在主目錄下創建bat文件，該文件將jpg圖片轉化爲LMDB數據格式

新建bat文件，計算mean.binaryproto

這裏選擇利用cifar10中的cifar10_full_solver.prototxt、cifar10_full_train_test.prototxt，在其基礎上進行修改，主要是修改了batch size，因爲我這邊的數據集很小，此外輸出類別種類數由10改爲了2

新建bat文件，訓練網絡

solve.prototxt中的內容如下：

# reduce learning rate after 120 epochs (60000 iters) by factor 0f 10
# then another factor of 10 after 10 more epochs (5000 iters)

# The train/test net protocol buffer definition
net: "./data/mydata/train_val.prototxt"
# test_iter specifies how many forward passes the test should carry out.
# In the case of CIFAR10, we have test batch size 100 and 100 test iterations,
# covering the full 10,000 testing images.
test_iter: 4
# Carry out testing every 1000 training iterations.
test_interval: 4
# The base learning rate, momentum and the weight decay of the network.
base_lr: 0.001
momentum: 0.9
weight_decay: 0.004
# The learning rate policy
lr_policy: "fixed"
# Display every 200 iterations
display: 4
# The maximum number of iterations
max_iter: 6000
# snapshot intermediate results
snapshot: 2000
snapshot_format: HDF5
snapshot_prefix: "./data/mydata/full"
# solver mode: CPU or GPU
solver_mode: GPU

train_val.prototxt中的內容如下：

name: "CIFAR10_full"
layer {
  name: "cifar"
  type: "Data"
  top: "data"
  top: "label"
  include {
    phase: TRAIN
  }
  transform_param {
    mean_file: "./data/mydata/mean.binaryproto"
  }
  data_param {
    source: "./data/mydata/mtrain"
    batch_size: 4
    backend: LMDB
  }
}
layer {
  name: "cifar"
  type: "Data"
  top: "data"
  top: "label"
  include {
    phase: TEST
  }
  transform_param {
    mean_file: "./data/mydata/mean.binaryproto"
  }
  data_param {
    source: "./data/mydata/mval"
    batch_size: 2
    backend: LMDB
  }
}
layer {
  name: "conv1"
  type: "Convolution"
  bottom: "data"
  top: "conv1"
  param {
    lr_mult: 1
  }
  param {
    lr_mult: 2
  }
  convolution_param {
    num_output: 32
    pad: 2
    kernel_size: 5
    stride: 1
    weight_filler {
      type: "gaussian"
      std: 0.0001
    }
    bias_filler {
      type: "constant"
    }
  }
}
layer {
  name: "pool1"
  type: "Pooling"
  bottom: "conv1"
  top: "pool1"
  pooling_param {
    pool: MAX
    kernel_size: 3
    stride: 2
  }
}
layer {
  name: "relu1"
  type: "ReLU"
  bottom: "pool1"
  top: "pool1"
}
layer {
  name: "norm1"
  type: "LRN"
  bottom: "pool1"
  top: "norm1"
  lrn_param {
    local_size: 3
    alpha: 5e-05
    beta: 0.75
    norm_region: WITHIN_CHANNEL
  }
}
layer {
  name: "conv2"
  type: "Convolution"
  bottom: "norm1"
  top: "conv2"
  param {
    lr_mult: 1
  }
  param {
    lr_mult: 2
  }
  convolution_param {
    num_output: 32
    pad: 2
    kernel_size: 5
    stride: 1
    weight_filler {
      type: "gaussian"
      std: 0.01
    }
    bias_filler {
      type: "constant"
    }
  }
}
layer {
  name: "relu2"
  type: "ReLU"
  bottom: "conv2"
  top: "conv2"
}
layer {
  name: "pool2"
  type: "Pooling"
  bottom: "conv2"
  top: "pool2"
  pooling_param {
    pool: AVE
    kernel_size: 3
    stride: 2
  }
}
layer {
  name: "norm2"
  type: "LRN"
  bottom: "pool2"
  top: "norm2"
  lrn_param {
    local_size: 3
    alpha: 5e-05
    beta: 0.75
    norm_region: WITHIN_CHANNEL
  }
}
layer {
  name: "conv3"
  type: "Convolution"
  bottom: "norm2"
  top: "conv3"
  convolution_param {
    num_output: 64
    pad: 2
    kernel_size: 5
    stride: 1
    weight_filler {
      type: "gaussian"
      std: 0.01
    }
    bias_filler {
      type: "constant"
    }
  }
}
layer {
  name: "relu3"
  type: "ReLU"
  bottom: "conv3"
  top: "conv3"
}
layer {
  name: "pool3"
  type: "Pooling"
  bottom: "conv3"
  top: "pool3"
  pooling_param {
    pool: AVE
    kernel_size: 3
    stride: 2
  }
}
layer {
  name: "ip1"
  type: "InnerProduct"
  bottom: "pool3"
  top: "ip1"
  param {
    lr_mult: 1
    decay_mult: 250
  }
  param {
    lr_mult: 2
    decay_mult: 0
  }
  inner_product_param {
    num_output: 2
    weight_filler {
      type: "gaussian"
      std: 0.01
    }
    bias_filler {
      type: "constant"
    }
  }
}
layer {
  name: "accuracy"
  type: "Accuracy"
  bottom: "ip1"
  bottom: "label"
  top: "accuracy"
  include {
    phase: TEST
  }
}
layer {
  name: "loss"
  type: "SoftmaxWithLoss"
  bottom: "ip1"
  bottom: "label"
  top: "loss"
}

運行bat文件，生成caffe.model文件

新建deploy.prototxt文件，該文件是由上面的train_val.prototxt修改而成

deploy.prototxt文件和train_val.prototxt文件不同的地方在於：

（1）輸入的數據不再是LMDB，也不分爲測試集和訓練集，輸入的類型爲Input，定義的維度，和訓練集的數據維度保持一致，121*121，否則會報錯；

（2）去掉weight_filler和bias_filler，這些參數已經存在於caffemodel中了，由caffemodel進行初始化。

（3）去掉最後的Accuracy層和loss層，換位Softmax層，表示分爲某一類的概率。

deploy.prototxt文件如下：

name: "CIFAR10_full"
layer {
  name: "data"
  type: "Input"
  top: "data"
  input_param{shape: {dim:1 dim:3 dim:121 dim:121}}
}

layer {
  name: "conv1"
  type: "Convolution"
  bottom: "data"
  top: "conv1"
  param {
    lr_mult: 1
  }
  param {
    lr_mult: 2
  }
  convolution_param {
    num_output: 32
    pad: 2
    kernel_size: 5
    stride: 1    
  }
}
layer {
  name: "pool1"
  type: "Pooling"
  bottom: "conv1"
  top: "pool1"
  pooling_param {
    pool: MAX
    kernel_size: 3
    stride: 2
  }
}
layer {
  name: "relu1"
  type: "ReLU"
  bottom: "pool1"
  top: "pool1"
}
layer {
  name: "norm1"
  type: "LRN"
  bottom: "pool1"
  top: "norm1"
  lrn_param {
    local_size: 3
    alpha: 5e-05
    beta: 0.75
    norm_region: WITHIN_CHANNEL
  }
}
layer {
  name: "conv2"
  type: "Convolution"
  bottom: "norm1"
  top: "conv2"
  param {
    lr_mult: 1
  }
  param {
    lr_mult: 2
  }
  convolution_param {
    num_output: 32
    pad: 2
    kernel_size: 5
    stride: 1
  }
}
layer {
  name: "relu2"
  type: "ReLU"
  bottom: "conv2"
  top: "conv2"
}
layer {
  name: "pool2"
  type: "Pooling"
  bottom: "conv2"
  top: "pool2"
  pooling_param {
    pool: AVE
    kernel_size: 3
    stride: 2
  }
}
layer {
  name: "norm2"
  type: "LRN"
  bottom: "pool2"
  top: "norm2"
  lrn_param {
    local_size: 3
    alpha: 5e-05
    beta: 0.75
    norm_region: WITHIN_CHANNEL
  }
}
layer {
  name: "conv3"
  type: "Convolution"
  bottom: "norm2"
  top: "conv3"
  convolution_param {
    num_output: 64
    pad: 2
    kernel_size: 5
    stride: 1 
  }
}
layer {
  name: "relu3"
  type: "ReLU"
  bottom: "conv3"
  top: "conv3"
}
layer {
  name: "pool3"
  type: "Pooling"
  bottom: "conv3"
  top: "pool3"
  pooling_param {
    pool: AVE
    kernel_size: 3
    stride: 2
  }
}
layer {
  name: "ip1"
  type: "InnerProduct"
  bottom: "pool3"
  top: "ip1"
  param {
    lr_mult: 1
    decay_mult: 250
  }
  param {
    lr_mult: 2
    decay_mult: 0
  }
  inner_product_param {
    num_output: 2
    weight_filler {
      type: "gaussian"
      std: 0.01
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "prob"
  type: "Softmax"
  bottom: "ip1"
  top: "prob"
}

運行結果如下：

測試圖片爲

可知預測結果正確

這裏附上以上相關文件鏈接

主目錄下的examples\mnist文件夾內文件鏈接：https://pan.baidu.com/s/1H2tlBFEun-EPXSDn5uH8xg 
提取碼：15xl 
主目錄下的examples\cifar10文件夾內文件python、matlab調用）鏈接：https://pan.baidu.com/s/17Sy5mKtaFMEyB4_CEyAyrA 
提取碼：5c1j 
主目錄下的\data\mydata文件夾內的文件 鏈接：https://pan.baidu.com/s/1deI-ka1CoIvJdgx9cyEhSQ 
提取碼：xgct

command line 文件 鏈接：https://pan.baidu.com/s/1kJONutsQlkN5KJ-n-lPmyg 
提取碼：5snd

後續會有博客介紹如何使用matlab、python、c++接口訓練網絡和預測圖片