Caffe Multilabel没有按预期工作

Question

我和Caffe玩了很长时间但从未做过多标签分类，而且我似乎陷入困境：

我正在使用

• 首先，我创建了lmdb（train_lmdb，val_lmdb），标签（labels_train_lmdb，labels_val_lmdb）和平均值（{{1} }）Caffe-LMDBCreation-MultiLabel。
• 该模型有7类大约13000张图像。
• 2000个图像有两个类（例如，向量是[1,0,0,1,0,0,0]
• 其余图片只有一个类（例如，矢量为[0,0,1,0,0,0,0]

我期待的是

我期望至少从火车数据集中获取图像，例如：

  

img1.jpg 0 0 0 1 0 0 0

对其进行分类，其值类似于[0.0,1.0,0.0,0.0,0.0,0.0,0.0]

我改为

对于上面的图片（img1.jpg），我有这些类型的结果：

mean_lmdb.binaryproto

没有意义。我尝试了几个快照（每10000次迭代一次），结果相似，所有快照都接近0.50

我的[0.48112139105796814, 0.5486980676651001, 0.5396456122398376, 0.44233766198158264, 0.5605107545852661, 0.3539462387561798, 0.5215630531311035]

train_val.prototxt

prototxt

deploy.prototxt ：

name: "multi-class-alexnet"
# --------------------------------- TRAIN -------------------------------
# -----------------------------------------------------------------------
layer {
  name: "data"
  type: "Data"
  top: "data"
  include {
    phase: TRAIN
  }
  transform_param {
    mirror: true
    crop_size: 180
    mean_file: "./mean_lmdb.binaryproto"
  }
  data_param {
    source: "./train_lmdb"
    batch_size: 64
    backend: LMDB
  }
}

# ---------------------------- TRAIN LABELS -----------------------------
# -----------------------------------------------------------------------

layer {
  name: "data"
  type: "Data"
  top: "label"
  include {
    phase: TRAIN
  }
  transform_param {
    scale: 0.00390625
    mean_value: 0
  }
  data_param {
    source: "./labels_train_lmdb"
    batch_size: 64
    backend: LMDB
  }
}

# ---------------------------------- VAL --------------------------------
# -----------------------------------------------------------------------

layer {
  name: "data"
  type: "Data"
  top: "data"
  include {
    phase: TEST
  }
  transform_param {
    mirror: false
    crop_size: 180
    mean_file: "./mean_lmdb.binaryproto"
  }
  data_param {
    source: "./val_lmdb"
    batch_size: 32
    backend: LMDB
  }
}

# ----------------------------- VAL LABELS ------------------------------
# -----------------------------------------------------------------------

layer {
  name: "data"
  type: "Data"
  top: "label"
  include {
    phase: TEST
  }
  transform_param {
    scale: 0.00390625
    mean_value: 0
  }
  data_param {
    source: "./labels_val_lmdb"
    batch_size: 32
    backend: LMDB
  }
}

layer {
  name: "conv1"
  type: "Convolution"
  bottom: "data"
  top: "conv1"
  param {
    lr_mult: 1
    decay_mult: 1
  }
  param {
    lr_mult: 2
    decay_mult: 0
  }
  convolution_param {
    num_output: 96
    kernel_size: 11
    stride: 4
    weight_filler {
      type: "gaussian"
      std: 0.01
    }
    bias_filler {
      type: "constant"
      value: 0
    }
  }
}
layer {
  name: "relu1"
  type: "ReLU"
  bottom: "conv1"
  top: "conv1"
}
layer {
  name: "pool1"
  type: "Pooling"
  bottom: "conv1"
  top: "pool1"
  pooling_param {
    pool: MAX
    kernel_size: 3
    stride: 2
  }
}
layer {
  name: "norm1"
  type: "LRN"
  bottom: "pool1"
  top: "norm1"
  lrn_param {
    local_size: 5
    alpha: 0.0001
    beta: 0.75
  }
}
layer {
  name: "conv2"
  type: "Convolution"
  bottom: "norm1"
  top: "conv2"
  param {
    lr_mult: 1
    decay_mult: 1
  }
  param {
    lr_mult: 2
    decay_mult: 0
  }
  convolution_param {
    num_output: 256
    pad: 2
    kernel_size: 5
    group: 2
    weight_filler {
      type: "gaussian"
      std: 0.01
    }
    bias_filler {
      type: "constant"
      value: 1
    }
  }
}
layer {
  name: "relu2"
  type: "ReLU"
  bottom: "conv2"
  top: "conv2"
}
layer {
  name: "pool2"
  type: "Pooling"
  bottom: "conv2"
  top: "pool2"
  pooling_param {
    pool: MAX
    kernel_size: 3
    stride: 2
  }
}
layer {
  name: "norm2"
  type: "LRN"
  bottom: "pool2"
  top: "norm2"
  lrn_param {
    local_size: 5
    alpha: 0.0001
    beta: 0.75
  }
}
layer {
  name: "conv3"
  type: "Convolution"
  bottom: "norm2"
  top: "conv3"
  param {
    lr_mult: 1
    decay_mult: 1
  }
  param {
    lr_mult: 2
    decay_mult: 0
  }
  convolution_param {
    num_output: 384
    pad: 1
    kernel_size: 3
    weight_filler {
      type: "gaussian"
      std: 0.01
    }
    bias_filler {
      type: "constant"
      value: 0
    }
  }
}
layer {
  name: "relu3"
  type: "ReLU"
  bottom: "conv3"
  top: "conv3"
}
layer {
  name: "conv4"
  type: "Convolution"
  bottom: "conv3"
  top: "conv4"
  param {
    lr_mult: 1
    decay_mult: 1
  }
  param {
    lr_mult: 2
    decay_mult: 0
  }
  convolution_param {
    num_output: 384
    pad: 1
    kernel_size: 3
    group: 2
    weight_filler {
      type: "gaussian"
      std: 0.01
    }
    bias_filler {
      type: "constant"
      value: 1
    }
  }
}
layer {
  name: "relu4"
  type: "ReLU"
  bottom: "conv4"
  top: "conv4"
}
layer {
  name: "conv5"
  type: "Convolution"
  bottom: "conv4"
  top: "conv5"
  param {
    lr_mult: 1
    decay_mult: 1
  }
  param {
    lr_mult: 2
    decay_mult: 0
  }
  convolution_param {
    num_output: 256
    pad: 1
    kernel_size: 3
    group: 2
    weight_filler {
      type: "gaussian"
      std: 0.01
    }
    bias_filler {
      type: "constant"
      value: 1
    }
  }
}
layer {
  name: "relu5"
  type: "ReLU"
  bottom: "conv5"
  top: "conv5"
}
layer {
  name: "pool5"
  type: "Pooling"
  bottom: "conv5"
  top: "pool5"
  pooling_param {
    pool: MAX
    kernel_size: 3
    stride: 2
  }
}
layer {
  name: "fc6_"
  type: "InnerProduct"
  bottom: "pool5"
  top: "fc6_"
  param {
    lr_mult: 1
    decay_mult: 1
  }
  param {
    lr_mult: 2
    decay_mult: 0
  }
  inner_product_param {
    num_output: 4096
    weight_filler {
      type: "gaussian"
      std: 0.005
    }
    bias_filler {
      type: "constant"
      value: 1
    }
  }
}
layer {
  name: "relu6"
  type: "ReLU"
  bottom: "fc6_"
  top: "fc6_"
}
layer {
  name: "drop6"
  type: "Dropout"
  bottom: "fc6_"
  top: "fc6_"
  dropout_param {
    dropout_ratio: 0.5
  }
}
layer {
  name: "fc7"
  type: "InnerProduct"
  bottom: "fc6_"
  top: "fc7"
  param {
    lr_mult: 1
    decay_mult: 1
  }
  param {
    lr_mult: 2
    decay_mult: 0
  }
  inner_product_param {
    num_output: 4096
    weight_filler {
      type: "gaussian"
      std: 0.005
    }
    bias_filler {
      type: "constant"
      value: 1
    }
  }
}
layer {
  name: "relu7"
  type: "ReLU"
  bottom: "fc7"
  top: "fc7"
}
layer {
  name: "drop7"
  type: "Dropout"
  bottom: "fc7"
  top: "fc7"
  dropout_param {
    dropout_ratio: 0.5
  }
}
layer {
  name: "latent"
  type: "InnerProduct"
  bottom: "fc7"
  top: "latent"
  param {
    lr_mult: 1
    decay_mult: 1
  }
  param {
    lr_mult: 2
    decay_mult: 0
  }
  inner_product_param {
    num_output: 48
    weight_filler {
      type: "gaussian"
      std: 0.01
    }
    bias_filler {
      type: "constant"
      value: 0
    }
  }
}
layer {
  bottom: "latent"
  top: "latent_sigmoid"
  name: "latent_sigmoid"
  type: "Sigmoid"
}
layer {
  name: "fc9"
  type: "InnerProduct"
  bottom: "latent_sigmoid"
  top: "fc9"
  param {
    lr_mult: 10
    decay_mult: 1
  }
  param {
    lr_mult: 20
    decay_mult: 0
  }
  inner_product_param {
    num_output: 7
    weight_filler {
      type: "gaussian"
      std: 0.2
    }
    bias_filler {
      type: "constant"
      value: 0
    }
  }
}
layer {
  name: "accuracy"
  type: "MultiLabelAccuracy"
  bottom: "fc9"
  bottom: "label"
  top: "accuracy"
  include {
    phase: TEST
  }
}
# ----------------------------------------------------------------
# ----------------- Multi-label Loss Function  -------------------
# ----------------------------------------------------------------
layer {
  name: "loss"
  type: "SigmoidCrossEntropyLoss"
  bottom: "fc9"
  bottom: "label"
  top: "loss"
}

损失函数

• 不知何故，我的模型显示了两个损失，我不明白，输出＃0和＃5的损失。这些是最后一行的name: "multi-class-alexnet" input: "data" input_shape { dim: 10 dim: 3 dim: 180 dim: 180 } layer { name: "conv1" type: "Convolution" bottom: "data" top: "conv1" param { lr_mult: 1 decay_mult: 1 } param { lr_mult: 2 decay_mult: 0 } convolution_param { num_output: 96 kernel_size: 11 stride: 4 weight_filler { type: "gaussian" std: 0.01 } bias_filler { type: "constant" value: 0 } } } layer { name: "relu1" type: "ReLU" bottom: "conv1" top: "conv1" } layer { name: "pool1" type: "Pooling" bottom: "conv1" top: "pool1" pooling_param { pool: MAX kernel_size: 3 stride: 2 } } layer { name: "norm1" type: "LRN" bottom: "pool1" top: "norm1" lrn_param { local_size: 5 alpha: 0.0001 beta: 0.75 } } layer { name: "conv2" type: "Convolution" bottom: "norm1" top: "conv2" param { lr_mult: 1 decay_mult: 1 } param { lr_mult: 2 decay_mult: 0 } convolution_param { num_output: 256 pad: 2 kernel_size: 5 group: 2 weight_filler { type: "gaussian" std: 0.01 } bias_filler { type: "constant" value: 1 } } } layer { name: "relu2" type: "ReLU" bottom: "conv2" top: "conv2" } layer { name: "pool2" type: "Pooling" bottom: "conv2" top: "pool2" pooling_param { pool: MAX kernel_size: 3 stride: 2 } } layer { name: "norm2" type: "LRN" bottom: "pool2" top: "norm2" lrn_param { local_size: 5 alpha: 0.0001 beta: 0.75 } } layer { name: "conv3" type: "Convolution" bottom: "norm2" top: "conv3" param { lr_mult: 1 decay_mult: 1 } param { lr_mult: 2 decay_mult: 0 } convolution_param { num_output: 384 pad: 1 kernel_size: 3 weight_filler { type: "gaussian" std: 0.01 } bias_filler { type: "constant" value: 0 } } } layer { name: "relu3" type: "ReLU" bottom: "conv3" top: "conv3" } layer { name: "conv4" type: "Convolution" bottom: "conv3" top: "conv4" param { lr_mult: 1 decay_mult: 1 } param { lr_mult: 2 decay_mult: 0 } convolution_param { num_output: 384 pad: 1 kernel_size: 3 group: 2 weight_filler { type: "gaussian" std: 0.01 } bias_filler { type: "constant" value: 1 } } } layer { name: "relu4" type: "ReLU" bottom: "conv4" top: "conv4" } layer { name: "conv5" type: "Convolution" bottom: "conv4" top: "conv5" param { lr_mult: 1 decay_mult: 1 } param { lr_mult: 2 decay_mult: 0 } convolution_param { num_output: 256 pad: 1 kernel_size: 3 group: 2 weight_filler { type: "gaussian" std: 0.01 } bias_filler { type: "constant" value: 1 } } } layer { name: "relu5" type: "ReLU" bottom: "conv5" top: "conv5" } layer { name: "pool5" type: "Pooling" bottom: "conv5" top: "pool5" pooling_param { pool: MAX kernel_size: 3 stride: 2 } } layer { name: "fc6_" type: "InnerProduct" bottom: "pool5" top: "fc6_" param { lr_mult: 1 decay_mult: 1 } param { lr_mult: 2 decay_mult: 0 } inner_product_param { num_output: 4096 weight_filler { type: "gaussian" std: 0.005 } bias_filler { type: "constant" value: 1 } } } layer { name: "relu6" type: "ReLU" bottom: "fc6_" top: "fc6_" } layer { name: "drop6" type: "Dropout" bottom: "fc6_" top: "fc6_" dropout_param { dropout_ratio: 0.5 } } layer { name: "fc7" type: "InnerProduct" bottom: "fc6_" top: "fc7" param { lr_mult: 1 decay_mult: 1 } param { lr_mult: 2 decay_mult: 0 } inner_product_param { num_output: 4096 weight_filler { type: "gaussian" std: 0.005 } bias_filler { type: "constant" value: 1 } } } layer { name: "relu7" type: "ReLU" bottom: "fc7" top: "fc7" } layer { name: "drop7" type: "Dropout" bottom: "fc7" top: "fc7" dropout_param { dropout_ratio: 0.5 } } layer { name: "latent_" type: "InnerProduct" bottom: "fc7" top: "latent_" param { lr_mult: 1 decay_mult: 1 } param { lr_mult: 2 decay_mult: 0 } inner_product_param { num_output: 7 weight_filler { type: "gaussian" std: 0.01 } bias_filler { type: "constant" value: 0 } } } layer { bottom: "latent_" top: "latent_sigmoid" name: "latent_sigmoid" type: "Sigmoid" } （超过110000次迭代）：

输出＃0 ：

  

I0427 10：20：04.475754 1817 solver.cpp：238]火车净输出＃0：损失= 0.0867133（* 1 = 0.0867133损失）

     

I0427 10：20：38.257825 1817 solver.cpp：238]火车净输出＃0：损失= 0.0477974（* 1 = 0.0477974损失）

     

I0427 10：21：11.794013 1817 solver.cpp：238]火车净输出＃0：损失= 0.0390092（* 1 = 0.0390092损失）

     

I0427 10：21：45.620671 1817 solver.cpp：238]火车净输出＃0：损失= 0.039954（* 1 = 0.039954损失）

     

I0427 10：22：19.271747 1817 solver.cpp：238]火车净输出＃0：损失= 0.0477802（* 1 = 0.0477802损失）

     

I0427 10：22：53.160802 1817 solver.cpp：238]列车净输出＃0：损失= 0.0406158（* 1 = 0.0406158损失）

     

I0427 10：23：26.843694 1817 solver.cpp：238]火车净输出＃0：损失= 0.0355715（* 1 = 0.0355715损失）

     

I0427 10：24：31.727321 1817 solver.cpp：238]火车净输出＃0：损失= 0.0396538（* 1 = 0.0396538损失）

     

I0427 10：25：05.019598 1817 solver.cpp：238]火车净输出＃0：损失= 0.037121（* 1 = 0.037121损失）

     

I0427 10：25：38.730303 1817 solver.cpp：238]火车净输出＃0：损失= 0.0362058（* 1 = 0.0362058损失）

输出＃5 ：

  

I0427 09：26：52.251719 1817 solver.cpp：398]测试净输出＃5：损失= 6.98116（* 1 = 6.98116损失）

     

I0427 09：33：01.639736 1817 solver.cpp：398]测试净输出＃5：损失= 6.99285（* 1 = 6.99285损失）

     

I0427 09：39：09.991879 1817 solver.cpp：398]测试净输出＃5：损失= 7.02165（* 1 = 7.02165损失）

     

I0427 09：45：18.013739 1817 solver.cpp：398]测试净输出＃5：损失= 7.01533（* 1 = 7.01533损失）

     

I0427 09：51：27.065721 1817 solver.cpp：398]测试净输出＃5：损失= 7.02347（* 1 = 7.02347损失）

     

I0427 09：58：13.271441 1817 solver.cpp：398]测试净输出＃5：损失= 6.98176（* 1 = 6.98176损失）

     

I0427 10：05：31.896226 1817 solver.cpp：398]测试净输出＃5：损失= 6.99103（* 1 = 6.99103损失）

     

I0427 10：12：12.693677 1817 solver.cpp：398]测试净输出＃5：损失= 7.02868（* 1 = 7.02868损失）

     

I0427 10：18：23.250385 1817 solver.cpp：398]测试净输出＃5：损失= 7.03427（* 1 = 7.03427损失）

     

I0427 10：24：31.239820 1817 solver.cpp：398]测试净输出＃5：损失= 6.97721（* 1 = 6.97721损失）