zero.prototxt

name: "ZeroNet"
layer {
  type: "Data"
  name: "data"
  # outputs
  top: "data"
  include {
    phase: TRAIN
  }
  data_param {
    source: "zero_train"
    batch_size: 8
    backend: LEVELDB
    threads: 1
    parser_threads: 1
  }
}

layer {
  type: "Data"
  name: "data"
  # outputs
  top: "data"
  include {
    phase: TEST
  }
  data_param {
    source: "zero_test"
    batch_size: 8
    backend: LEVELDB
    threads: 1
    parser_threads: 1
  }
}

layer {
  type: "Data"
  name: "label"
  # outputs
  top: "label_unsliced"
  include {
    phase: TRAIN
  }
  data_param {
    source: "zero_train_label"
    batch_size: 8
    backend: LEVELDB
    threads: 1
    parser_threads: 1
  }
}

layer {
  type: "Data"
  name: "label"
  # outputs
  top: "label_unsliced"
  include {
    phase: TEST
  }
  data_param {
    source: "zero_test_label"
    batch_size: 8
    backend: LEVELDB
    threads: 1
    parser_threads: 1
  }
}

layer {
  name: "slicer_label"
  type: "Slice"
  bottom: "label_unsliced"
  top: "label_move"
  top: "label_won"
  slice_param {
    axis: 1
    slice_point: 1
  }
}

# input layer
layer {
  name: "conv1"
  type: "Convolution"
  bottom: "data"
  top: "conv1"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn1"
  type: "BatchNorm"
  bottom: "conv1"
  top: "conv1"
}

layer {
  name: "relu1"
  type: "ReLU"
  bottom: "conv1"
  top: "conv1"
}

# residual block 1
layer {
  name: "conv2"
  type: "Convolution"
  bottom: "conv1"
  top: "conv2"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn2"
  type: "BatchNorm"
  bottom: "conv2"
  top: "conv2"
}

layer {
  name: "relu2"
  type: "ReLU"
  bottom: "conv2"
  top: "conv2"
}

layer {
  name: "conv3"
  type: "Convolution"
  bottom: "conv2"
  top: "conv3"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn3"
  type: "BatchNorm"
  bottom: "conv3"
  top: "conv3"
}

layer {
  name: "elt3"
  type: "Eltwise"
  bottom: "conv1"
  bottom: "conv3"
  top: "elt3"
}

layer {
  name: "relu3"
  type: "ReLU"
  bottom: "elt3"
  top: "elt3"
}


# residual block 2
layer {
  name: "conv4"
  type: "Convolution"
  bottom: "elt3"
  top: "conv4"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn4"
  type: "BatchNorm"
  bottom: "conv4"
  top: "conv4"
}

layer {
  name: "relu4"
  type: "ReLU"
  bottom: "conv4"
  top: "conv4"
}

layer {
  name: "conv5"
  type: "Convolution"
  bottom: "conv4"
  top: "conv5"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn5"
  type: "BatchNorm"
  bottom: "conv5"
  top: "conv5"
}

layer {
  name: "elt5"
  type: "Eltwise"
  bottom: "elt3"
  bottom: "conv5"
  top: "elt5"
}

layer {
  name: "relu5"
  type: "ReLU"
  bottom: "elt5"
  top: "elt5"
}

# residual block 3
layer {
  name: "conv6"
  type: "Convolution"
  bottom: "elt5"
  top: "conv6"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn6"
  type: "BatchNorm"
  bottom: "conv6"
  top: "conv6"
}

layer {
  name: "relu6"
  type: "ReLU"
  bottom: "conv6"
  top: "conv6"
}

layer {
  name: "conv7"
  type: "Convolution"
  bottom: "conv6"
  top: "conv7"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn7"
  type: "BatchNorm"
  bottom: "conv7"
  top: "conv7"
}

layer {
  name: "elt7"
  type: "Eltwise"
  bottom: "elt5"
  bottom: "conv7"
  top: "elt7"
}

layer {
  name: "relu7"
  type: "ReLU"
  bottom: "elt7"
  top: "elt7"
}

# residual block 4
layer {
  name: "conv8"
  type: "Convolution"
  bottom: "elt7"
  top: "conv8"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn8"
  type: "BatchNorm"
  bottom: "conv8"
  top: "conv8"
}

layer {
  name: "relu8"
  type: "ReLU"
  bottom: "conv8"
  top: "conv8"
}

layer {
  name: "conv9"
  type: "Convolution"
  bottom: "conv8"
  top: "conv9"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn9"
  type: "BatchNorm"
  bottom: "conv9"
  top: "conv9"
}

layer {
  name: "elt9"
  type: "Eltwise"
  bottom: "elt7"
  bottom: "conv9"
  top: "elt9"
}

layer {
  name: "relu9"
  type: "ReLU"
  bottom: "elt9"
  top: "elt9"
}

# residual block 5
layer {
  name: "conv10"
  type: "Convolution"
  bottom: "elt9"
  top: "conv10"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn10"
  type: "BatchNorm"
  bottom: "conv10"
  top: "conv10"
}

layer {
  name: "relu10"
  type: "ReLU"
  bottom: "conv10"
  top: "conv10"
}

layer {
  name: "conv11"
  type: "Convolution"
  bottom: "conv10"
  top: "conv11"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn11"
  type: "BatchNorm"
  bottom: "conv11"
  top: "conv11"
}

layer {
  name: "elt11"
  type: "Eltwise"
  bottom: "elt9"
  bottom: "conv11"
  top: "elt11"
}

layer {
  name: "relu11"
  type: "ReLU"
  bottom: "elt11"
  top: "elt11"
}

# residual block 6
layer {
  name: "conv12"
  type: "Convolution"
  bottom: "elt11"
  top: "conv12"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn12"
  type: "BatchNorm"
  bottom: "conv12"
  top: "conv12"
}

layer {
  name: "relu12"
  type: "ReLU"
  bottom: "conv12"
  top: "conv12"
}

layer {
  name: "conv13"
  type: "Convolution"
  bottom: "conv12"
  top: "conv13"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn13"
  type: "BatchNorm"
  bottom: "conv13"
  top: "conv13"
}

layer {
  name: "elt13"
  type: "Eltwise"
  bottom: "elt11"
  bottom: "conv13"
  top: "elt13"
}

layer {
  name: "relu13"
  type: "ReLU"
  bottom: "elt13"
  top: "elt13"
}

# residual block 7
layer {
  name: "conv14"
  type: "Convolution"
  bottom: "elt13"
  top: "conv14"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn14"
  type: "BatchNorm"
  bottom: "conv14"
  top: "conv14"
}

layer {
  name: "relu14"
  type: "ReLU"
  bottom: "conv14"
  top: "conv14"
}

layer {
  name: "conv15"
  type: "Convolution"
  bottom: "conv14"
  top: "conv15"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn15"
  type: "BatchNorm"
  bottom: "conv15"
  top: "conv15"
}

layer {
  name: "elt15"
  type: "Eltwise"
  bottom: "elt13"
  bottom: "conv15"
  top: "elt15"
}

layer {
  name: "relu15"
  type: "ReLU"
  bottom: "elt15"
  top: "elt15"
}

# residual block 8
layer {
  name: "conv16"
  type: "Convolution"
  bottom: "elt15"
  top: "conv16"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn16"
  type: "BatchNorm"
  bottom: "conv16"
  top: "conv16"
}

layer {
  name: "relu16"
  type: "ReLU"
  bottom: "conv16"
  top: "conv16"
}

layer {
  name: "conv17"
  type: "Convolution"
  bottom: "conv16"
  top: "conv17"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn17"
  type: "BatchNorm"
  bottom: "conv17"
  top: "conv17"
}

layer {
  name: "elt17"
  type: "Eltwise"
  bottom: "elt15"
  bottom: "conv17"
  top: "elt17"
}

layer {
  name: "relu17"
  type: "ReLU"
  bottom: "elt17"
  top: "elt17"
}

# residual block 9
layer {
  name: "conv18"
  type: "Convolution"
  bottom: "elt17"
  top: "conv18"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn18"
  type: "BatchNorm"
  bottom: "conv18"
  top: "conv18"
}

layer {
  name: "relu18"
  type: "ReLU"
  bottom: "conv18"
  top: "conv18"
}

layer {
  name: "conv19"
  type: "Convolution"
  bottom: "conv18"
  top: "conv19"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn19"
  type: "BatchNorm"
  bottom: "conv19"
  top: "conv19"
}

layer {
  name: "elt19"
  type: "Eltwise"
  bottom: "elt17"
  bottom: "conv19"
  top: "elt19"
}

layer {
  name: "relu19"
  type: "ReLU"
  bottom: "elt19"
  top: "elt19"
}

# residual block 10
layer {
  name: "conv20"
  type: "Convolution"
  bottom: "elt19"
  top: "conv20"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn20"
  type: "BatchNorm"
  bottom: "conv20"
  top: "conv20"
}

layer {
  name: "relu20"
  type: "ReLU"
  bottom: "conv20"
  top: "conv20"
}

layer {
  name: "conv21"
  type: "Convolution"
  bottom: "conv20"
  top: "conv21"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn21"
  type: "BatchNorm"
  bottom: "conv21"
  top: "conv21"
}

layer {
  name: "elt21"
  type: "Eltwise"
  bottom: "elt19"
  bottom: "conv21"
  top: "elt21"
}

layer {
  name: "relu21"
  type: "ReLU"
  bottom: "elt21"
  top: "elt21"
}

# residual block 11
layer {
  name: "conv22"
  type: "Convolution"
  bottom: "elt21"
  top: "conv22"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn22"
  type: "BatchNorm"
  bottom: "conv22"
  top: "conv22"
}

layer {
  name: "relu22"
  type: "ReLU"
  bottom: "conv22"
  top: "conv22"
}

layer {
  name: "conv23"
  type: "Convolution"
  bottom: "conv22"
  top: "conv23"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn23"
  type: "BatchNorm"
  bottom: "conv23"
  top: "conv23"
}

layer {
  name: "elt23"
  type: "Eltwise"
  bottom: "elt21"
  bottom: "conv23"
  top: "elt23"
}

layer {
  name: "relu23"
  type: "ReLU"
  bottom: "elt23"
  top: "elt23"
}

# residual block 12
layer {
  name: "conv24"
  type: "Convolution"
  bottom: "elt23"
  top: "conv24"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn24"
  type: "BatchNorm"
  bottom: "conv24"
  top: "conv24"
}

layer {
  name: "relu24"
  type: "ReLU"
  bottom: "conv24"
  top: "conv24"
}

layer {
  name: "conv25"
  type: "Convolution"
  bottom: "conv24"
  top: "conv25"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn25"
  type: "BatchNorm"
  bottom: "conv25"
  top: "conv25"
}

layer {
  name: "elt25"
  type: "Eltwise"
  bottom: "elt23"
  bottom: "conv25"
  top: "elt25"
}

layer {
  name: "relu25"
  type: "ReLU"
  bottom: "elt25"
  top: "elt25"
}

# residual block 13
layer {
  name: "conv26"
  type: "Convolution"
  bottom: "elt25"
  top: "conv26"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn26"
  type: "BatchNorm"
  bottom: "conv26"
  top: "conv26"
}

layer {
  name: "relu26"
  type: "ReLU"
  bottom: "conv26"
  top: "conv26"
}

layer {
  name: "conv27"
  type: "Convolution"
  bottom: "conv26"
  top: "conv27"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn27"
  type: "BatchNorm"
  bottom: "conv27"
  top: "conv27"
}

layer {
  name: "elt27"
  type: "Eltwise"
  bottom: "elt25"
  bottom: "conv27"
  top: "elt27"
}

layer {
  name: "relu27"
  type: "ReLU"
  bottom: "elt27"
  top: "elt27"
}

# residual block 14
layer {
  name: "conv28"
  type: "Convolution"
  bottom: "elt27"
  top: "conv28"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn28"
  type: "BatchNorm"
  bottom: "conv28"
  top: "conv28"
}

layer {
  name: "relu28"
  type: "ReLU"
  bottom: "conv28"
  top: "conv28"
}

layer {
  name: "conv29"
  type: "Convolution"
  bottom: "conv28"
  top: "conv29"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn29"
  type: "BatchNorm"
  bottom: "conv29"
  top: "conv29"
}

layer {
  name: "elt29"
  type: "Eltwise"
  bottom: "elt27"
  bottom: "conv29"
  top: "elt29"
}

layer {
  name: "relu29"
  type: "ReLU"
  bottom: "elt29"
  top: "elt29"
}

# residual block 15
layer {
  name: "conv30"
  type: "Convolution"
  bottom: "elt29"
  top: "conv30"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn30"
  type: "BatchNorm"
  bottom: "conv30"
  top: "conv30"
}

layer {
  name: "relu30"
  type: "ReLU"
  bottom: "conv30"
  top: "conv30"
}

layer {
  name: "conv31"
  type: "Convolution"
  bottom: "conv30"
  top: "conv31"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn31"
  type: "BatchNorm"
  bottom: "conv31"
  top: "conv31"
}

layer {
  name: "elt31"
  type: "Eltwise"
  bottom: "elt29"
  bottom: "conv31"
  top: "elt31"
}

layer {
  name: "relu31"
  type: "ReLU"
  bottom: "elt31"
  top: "elt31"
}

# residual block 16
layer {
  name: "conv32"
  type: "Convolution"
  bottom: "elt31"
  top: "conv32"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn32"
  type: "BatchNorm"
  bottom: "conv32"
  top: "conv32"
}

layer {
  name: "relu32"
  type: "ReLU"
  bottom: "conv32"
  top: "conv32"
}

layer {
  name: "conv33"
  type: "Convolution"
  bottom: "conv32"
  top: "conv33"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn33"
  type: "BatchNorm"
  bottom: "conv33"
  top: "conv33"
}

layer {
  name: "elt33"
  type: "Eltwise"
  bottom: "elt31"
  bottom: "conv33"
  top: "elt33"
}

layer {
  name: "relu33"
  type: "ReLU"
  bottom: "elt33"
  top: "elt33"
}

# residual block 17
layer {
  name: "conv34"
  type: "Convolution"
  bottom: "elt33"
  top: "conv34"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn34"
  type: "BatchNorm"
  bottom: "conv34"
  top: "conv34"
}

layer {
  name: "relu34"
  type: "ReLU"
  bottom: "conv34"
  top: "conv34"
}

layer {
  name: "conv35"
  type: "Convolution"
  bottom: "conv34"
  top: "conv35"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn35"
  type: "BatchNorm"
  bottom: "conv35"
  top: "conv35"
}

layer {
  name: "elt35"
  type: "Eltwise"
  bottom: "elt33"
  bottom: "conv35"
  top: "elt35"
}

layer {
  name: "relu35"
  type: "ReLU"
  bottom: "elt35"
  top: "elt35"
}

# residual block 18
layer {
  name: "conv36"
  type: "Convolution"
  bottom: "elt35"
  top: "conv36"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn36"
  type: "BatchNorm"
  bottom: "conv36"
  top: "conv36"
}

layer {
  name: "relu36"
  type: "ReLU"
  bottom: "conv36"
  top: "conv36"
}

layer {
  name: "conv37"
  type: "Convolution"
  bottom: "conv36"
  top: "conv37"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn37"
  type: "BatchNorm"
  bottom: "conv37"
  top: "conv37"
}

layer {
  name: "elt37"
  type: "Eltwise"
  bottom: "elt35"
  bottom: "conv37"
  top: "elt37"
}

layer {
  name: "relu37"
  type: "ReLU"
  bottom: "elt37"
  top: "elt37"
}

# residual block 19
layer {
  name: "conv38"
  type: "Convolution"
  bottom: "elt37"
  top: "conv38"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn38"
  type: "BatchNorm"
  bottom: "conv38"
  top: "conv38"
}

layer {
  name: "relu38"
  type: "ReLU"
  bottom: "conv38"
  top: "conv38"
}

layer {
  name: "conv39"
  type: "Convolution"
  bottom: "conv38"
  top: "conv39"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn39"
  type: "BatchNorm"
  bottom: "conv39"
  top: "conv39"
}

layer {
  name: "elt39"
  type: "Eltwise"
  bottom: "elt37"
  bottom: "conv39"
  top: "elt39"
}

layer {
  name: "relu39"
  type: "ReLU"
  bottom: "elt39"
  top: "elt39"
}

# residual block 20
layer {
  name: "conv40"
  type: "Convolution"
  bottom: "elt39"
  top: "conv40"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn40"
  type: "BatchNorm"
  bottom: "conv40"
  top: "conv40"
}

layer {
  name: "relu40"
  type: "ReLU"
  bottom: "conv40"
  top: "conv40"
}

layer {
  name: "conv41"
  type: "Convolution"
  bottom: "conv40"
  top: "conv41"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn41"
  type: "BatchNorm"
  bottom: "conv41"
  top: "conv41"
}

layer {
  name: "elt41"
  type: "Eltwise"
  bottom: "elt39"
  bottom: "conv41"
  top: "elt41"
}

layer {
  name: "relu41"
  type: "ReLU"
  bottom: "elt41"
  top: "elt41"
}

# residual block 21
layer {
  name: "conv42"
  type: "Convolution"
  bottom: "elt41"
  top: "conv42"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn42"
  type: "BatchNorm"
  bottom: "conv42"
  top: "conv42"
}

layer {
  name: "relu42"
  type: "ReLU"
  bottom: "conv42"
  top: "conv42"
}

layer {
  name: "conv43"
  type: "Convolution"
  bottom: "conv42"
  top: "conv43"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn43"
  type: "BatchNorm"
  bottom: "conv43"
  top: "conv43"
}

layer {
  name: "elt43"
  type: "Eltwise"
  bottom: "elt41"
  bottom: "conv43"
  top: "elt43"
}

layer {
  name: "relu43"
  type: "ReLU"
  bottom: "elt43"
  top: "elt43"
}

# residual block 22
layer {
  name: "conv44"
  type: "Convolution"
  bottom: "elt43"
  top: "conv44"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn44"
  type: "BatchNorm"
  bottom: "conv44"
  top: "conv44"
}

layer {
  name: "relu44"
  type: "ReLU"
  bottom: "conv44"
  top: "conv44"
}

layer {
  name: "conv45"
  type: "Convolution"
  bottom: "conv44"
  top: "conv45"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn45"
  type: "BatchNorm"
  bottom: "conv45"
  top: "conv45"
}

layer {
  name: "elt45"
  type: "Eltwise"
  bottom: "elt43"
  bottom: "conv45"
  top: "elt45"
}

layer {
  name: "relu45"
  type: "ReLU"
  bottom: "elt45"
  top: "elt45"
}

# residual block 23
layer {
  name: "conv46"
  type: "Convolution"
  bottom: "elt45"
  top: "conv46"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn46"
  type: "BatchNorm"
  bottom: "conv46"
  top: "conv46"
}

layer {
  name: "relu46"
  type: "ReLU"
  bottom: "conv46"
  top: "conv46"
}

layer {
  name: "conv47"
  type: "Convolution"
  bottom: "conv46"
  top: "conv47"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn47"
  type: "BatchNorm"
  bottom: "conv47"
  top: "conv47"
}

layer {
  name: "elt47"
  type: "Eltwise"
  bottom: "elt45"
  bottom: "conv47"
  top: "elt47"
}

layer {
  name: "relu47"
  type: "ReLU"
  bottom: "elt47"
  top: "elt47"
}

# residual block 24
layer {
  name: "conv48"
  type: "Convolution"
  bottom: "elt47"
  top: "conv48"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn48"
  type: "BatchNorm"
  bottom: "conv48"
  top: "conv48"
}

layer {
  name: "relu48"
  type: "ReLU"
  bottom: "conv48"
  top: "conv48"
}

layer {
  name: "conv49"
  type: "Convolution"
  bottom: "conv48"
  top: "conv49"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn49"
  type: "BatchNorm"
  bottom: "conv49"
  top: "conv49"
}

layer {
  name: "elt49"
  type: "Eltwise"
  bottom: "elt47"
  bottom: "conv49"
  top: "elt49"
}

layer {
  name: "relu49"
  type: "ReLU"
  bottom: "elt49"
  top: "elt49"
}

# residual block 25
layer {
  name: "conv50"
  type: "Convolution"
  bottom: "elt49"
  top: "conv50"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn50"
  type: "BatchNorm"
  bottom: "conv50"
  top: "conv50"
}

layer {
  name: "relu50"
  type: "ReLU"
  bottom: "conv50"
  top: "conv50"
}

layer {
  name: "conv51"
  type: "Convolution"
  bottom: "conv50"
  top: "conv51"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn51"
  type: "BatchNorm"
  bottom: "conv51"
  top: "conv51"
}

layer {
  name: "elt51"
  type: "Eltwise"
  bottom: "elt49"
  bottom: "conv51"
  top: "elt51"
}

layer {
  name: "relu51"
  type: "ReLU"
  bottom: "elt51"
  top: "elt51"
}

# residual block 26
layer {
  name: "conv52"
  type: "Convolution"
  bottom: "elt51"
  top: "conv52"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn52"
  type: "BatchNorm"
  bottom: "conv52"
  top: "conv52"
}

layer {
  name: "relu52"
  type: "ReLU"
  bottom: "conv52"
  top: "conv52"
}

layer {
  name: "conv53"
  type: "Convolution"
  bottom: "conv52"
  top: "conv53"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn53"
  type: "BatchNorm"
  bottom: "conv53"
  top: "conv53"
}

layer {
  name: "elt53"
  type: "Eltwise"
  bottom: "elt51"
  bottom: "conv53"
  top: "elt53"
}

layer {
  name: "relu53"
  type: "ReLU"
  bottom: "elt53"
  top: "elt53"
}

# residual block 27
layer {
  name: "conv54"
  type: "Convolution"
  bottom: "elt53"
  top: "conv54"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn54"
  type: "BatchNorm"
  bottom: "conv54"
  top: "conv54"
}

layer {
  name: "relu54"
  type: "ReLU"
  bottom: "conv54"
  top: "conv54"
}

layer {
  name: "conv55"
  type: "Convolution"
  bottom: "conv54"
  top: "conv55"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn55"
  type: "BatchNorm"
  bottom: "conv55"
  top: "conv55"
}

layer {
  name: "elt55"
  type: "Eltwise"
  bottom: "elt53"
  bottom: "conv55"
  top: "elt55"
}

layer {
  name: "relu55"
  type: "ReLU"
  bottom: "elt55"
  top: "elt55"
}

# residual block 28
layer {
  name: "conv56"
  type: "Convolution"
  bottom: "elt55"
  top: "conv56"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn56"
  type: "BatchNorm"
  bottom: "conv56"
  top: "conv56"
}

layer {
  name: "relu56"
  type: "ReLU"
  bottom: "conv56"
  top: "conv56"
}

layer {
  name: "conv57"
  type: "Convolution"
  bottom: "conv56"
  top: "conv57"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn57"
  type: "BatchNorm"
  bottom: "conv57"
  top: "conv57"
}

layer {
  name: "elt57"
  type: "Eltwise"
  bottom: "elt55"
  bottom: "conv57"
  top: "elt57"
}

layer {
  name: "relu57"
  type: "ReLU"
  bottom: "elt57"
  top: "elt57"
}

# residual block 29
layer {
  name: "conv58"
  type: "Convolution"
  bottom: "elt57"
  top: "conv58"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn58"
  type: "BatchNorm"
  bottom: "conv58"
  top: "conv58"
}

layer {
  name: "relu58"
  type: "ReLU"
  bottom: "conv58"
  top: "conv58"
}

layer {
  name: "conv59"
  type: "Convolution"
  bottom: "conv58"
  top: "conv59"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn59"
  type: "BatchNorm"
  bottom: "conv59"
  top: "conv59"
}

layer {
  name: "elt59"
  type: "Eltwise"
  bottom: "elt57"
  bottom: "conv59"
  top: "elt59"
}

layer {
  name: "relu59"
  type: "ReLU"
  bottom: "elt59"
  top: "elt59"
}

# residual block 30
layer {
  name: "conv60"
  type: "Convolution"
  bottom: "elt59"
  top: "conv60"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn60"
  type: "BatchNorm"
  bottom: "conv60"
  top: "conv60"
}

layer {
  name: "relu60"
  type: "ReLU"
  bottom: "conv60"
  top: "conv60"
}

layer {
  name: "conv61"
  type: "Convolution"
  bottom: "conv60"
  top: "conv61"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn61"
  type: "BatchNorm"
  bottom: "conv61"
  top: "conv61"
}

layer {
  name: "elt61"
  type: "Eltwise"
  bottom: "elt59"
  bottom: "conv61"
  top: "elt61"
}

layer {
  name: "relu61"
  type: "ReLU"
  bottom: "elt61"
  top: "elt61"
}

# residual block 31
layer {
  name: "conv62"
  type: "Convolution"
  bottom: "elt61"
  top: "conv62"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn62"
  type: "BatchNorm"
  bottom: "conv62"
  top: "conv62"
}

layer {
  name: "relu62"
  type: "ReLU"
  bottom: "conv62"
  top: "conv62"
}

layer {
  name: "conv63"
  type: "Convolution"
  bottom: "conv62"
  top: "conv63"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn63"
  type: "BatchNorm"
  bottom: "conv63"
  top: "conv63"
}

layer {
  name: "elt63"
  type: "Eltwise"
  bottom: "elt61"
  bottom: "conv63"
  top: "elt63"
}

layer {
  name: "relu63"
  type: "ReLU"
  bottom: "elt63"
  top: "elt63"
}

# residual block 32
layer {
  name: "conv64"
  type: "Convolution"
  bottom: "elt63"
  top: "conv64"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn64"
  type: "BatchNorm"
  bottom: "conv64"
  top: "conv64"
}

layer {
  name: "relu64"
  type: "ReLU"
  bottom: "conv64"
  top: "conv64"
}

layer {
  name: "conv65"
  type: "Convolution"
  bottom: "conv64"
  top: "conv65"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn65"
  type: "BatchNorm"
  bottom: "conv65"
  top: "conv65"
}

layer {
  name: "elt65"
  type: "Eltwise"
  bottom: "elt63"
  bottom: "conv65"
  top: "elt65"
}

layer {
  name: "relu65"
  type: "ReLU"
  bottom: "elt65"
  top: "elt65"
}

# residual block 33
layer {
  name: "conv66"
  type: "Convolution"
  bottom: "elt65"
  top: "conv66"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn66"
  type: "BatchNorm"
  bottom: "conv66"
  top: "conv66"
}

layer {
  name: "relu66"
  type: "ReLU"
  bottom: "conv66"
  top: "conv66"
}

layer {
  name: "conv67"
  type: "Convolution"
  bottom: "conv66"
  top: "conv67"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn67"
  type: "BatchNorm"
  bottom: "conv67"
  top: "conv67"
}

layer {
  name: "elt67"
  type: "Eltwise"
  bottom: "elt65"
  bottom: "conv67"
  top: "elt67"
}

layer {
  name: "relu67"
  type: "ReLU"
  bottom: "elt67"
  top: "elt67"
}

# residual block 34
layer {
  name: "conv68"
  type: "Convolution"
  bottom: "elt67"
  top: "conv68"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn68"
  type: "BatchNorm"
  bottom: "conv68"
  top: "conv68"
}

layer {
  name: "relu68"
  type: "ReLU"
  bottom: "conv68"
  top: "conv68"
}

layer {
  name: "conv69"
  type: "Convolution"
  bottom: "conv68"
  top: "conv69"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn69"
  type: "BatchNorm"
  bottom: "conv69"
  top: "conv69"
}

layer {
  name: "elt69"
  type: "Eltwise"
  bottom: "elt67"
  bottom: "conv69"
  top: "elt69"
}

layer {
  name: "relu69"
  type: "ReLU"
  bottom: "elt69"
  top: "elt69"
}

# residual block 35
layer {
  name: "conv70"
  type: "Convolution"
  bottom: "elt69"
  top: "conv70"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn70"
  type: "BatchNorm"
  bottom: "conv70"
  top: "conv70"
}

layer {
  name: "relu70"
  type: "ReLU"
  bottom: "conv70"
  top: "conv70"
}

layer {
  name: "conv71"
  type: "Convolution"
  bottom: "conv70"
  top: "conv71"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn71"
  type: "BatchNorm"
  bottom: "conv71"
  top: "conv71"
}

layer {
  name: "elt71"
  type: "Eltwise"
  bottom: "elt69"
  bottom: "conv71"
  top: "elt71"
}

layer {
  name: "relu71"
  type: "ReLU"
  bottom: "elt71"
  top: "elt71"
}

# residual block 36
layer {
  name: "conv72"
  type: "Convolution"
  bottom: "elt71"
  top: "conv72"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn72"
  type: "BatchNorm"
  bottom: "conv72"
  top: "conv72"
}

layer {
  name: "relu72"
  type: "ReLU"
  bottom: "conv72"
  top: "conv72"
}

layer {
  name: "conv73"
  type: "Convolution"
  bottom: "conv72"
  top: "conv73"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn73"
  type: "BatchNorm"
  bottom: "conv73"
  top: "conv73"
}

layer {
  name: "elt73"
  type: "Eltwise"
  bottom: "elt71"
  bottom: "conv73"
  top: "elt73"
}

layer {
  name: "relu73"
  type: "ReLU"
  bottom: "elt73"
  top: "elt73"
}

# residual block 37
layer {
  name: "conv74"
  type: "Convolution"
  bottom: "elt73"
  top: "conv74"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn74"
  type: "BatchNorm"
  bottom: "conv74"
  top: "conv74"
}

layer {
  name: "relu74"
  type: "ReLU"
  bottom: "conv74"
  top: "conv74"
}

layer {
  name: "conv75"
  type: "Convolution"
  bottom: "conv74"
  top: "conv75"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn75"
  type: "BatchNorm"
  bottom: "conv75"
  top: "conv75"
}

layer {
  name: "elt75"
  type: "Eltwise"
  bottom: "elt73"
  bottom: "conv75"
  top: "elt75"
}

layer {
  name: "relu75"
  type: "ReLU"
  bottom: "elt75"
  top: "elt75"
}

# residual block 38
layer {
  name: "conv76"
  type: "Convolution"
  bottom: "elt75"
  top: "conv76"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn76"
  type: "BatchNorm"
  bottom: "conv76"
  top: "conv76"
}

layer {
  name: "relu76"
  type: "ReLU"
  bottom: "conv76"
  top: "conv76"
}

layer {
  name: "conv77"
  type: "Convolution"
  bottom: "conv76"
  top: "conv77"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn77"
  type: "BatchNorm"
  bottom: "conv77"
  top: "conv77"
}

layer {
  name: "elt77"
  type: "Eltwise"
  bottom: "elt75"
  bottom: "conv77"
  top: "elt77"
}

layer {
  name: "relu77"
  type: "ReLU"
  bottom: "elt77"
  top: "elt77"
}

# residual block 39
layer {
  name: "conv78"
  type: "Convolution"
  bottom: "elt77"
  top: "conv78"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn78"
  type: "BatchNorm"
  bottom: "conv78"
  top: "conv78"
}

layer {
  name: "relu78"
  type: "ReLU"
  bottom: "conv78"
  top: "conv78"
}

layer {
  name: "conv79"
  type: "Convolution"
  bottom: "conv78"
  top: "conv79"
  convolution_param {
    num_output: 256
    kernel_size: 3
    pad: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "bn79"
  type: "BatchNorm"
  bottom: "conv79"
  top: "conv79"
}

layer {
  name: "elt79"
  type: "Eltwise"
  bottom: "elt77"
  bottom: "conv79"
  top: "elt79"
}

layer {
  name: "relu79"
  type: "ReLU"
  bottom: "elt79"
  top: "elt79"
}

layer {
  name: "p_conv1"
  type: "Convolution"
  bottom: "elt79"
  top: "pol1"
  convolution_param {
    num_output: 2
    kernel_size: 1
    pad: 0
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "p_bn1"
  type: "BatchNorm"
  bottom: "pol1"
  top: "pol1"
}

layer {
  name: "p_relu1"
  type: "ReLU"
  bottom: "pol1"
  top: "pol1"
}

layer {
  name: "p_ip1"
  type: "InnerProduct"
  inner_product_param {
    num_output: 362
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
  bottom: "pol1"
  top: "pol2"
}

layer {
  name: "v_conv1"
  type: "Convolution"
  bottom: "elt79"
  top: "val1"
  convolution_param {
    num_output: 1
    kernel_size: 1
    pad: 0
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}

layer {
  name: "v_bn1"
  type: "BatchNorm"
  bottom: "val1"
  top: "val1"
}

layer {
  name: "v_relu1"
  type: "ReLU"
  bottom: "val1"
  top: "val1"
}

layer {
  name: "v_ip1"
  type: "InnerProduct"
  inner_product_param {
    num_output: 256
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
  bottom: "val1"
  top: "val2"
}

layer {
  name: "v_relu2"
  type: "ReLU"
  bottom: "val2"
  top: "val2"
}

layer {
  name: "v_ip2"
  type: "InnerProduct"
  inner_product_param {
    num_output: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
  bottom: "val2"
  top: "val3"
}

layer {
  name: "v_tanh"
  type: "TanH"
  bottom: "val3"
  top: "score"
}

layer {
  name: "loss_move"
  type: "SoftmaxWithLoss"
  bottom: "pol2"
  bottom: "label_move"
  top: "loss_move"
  loss_weight: 0.99
}

layer {
  name: "loss_score"
  type: "EuclideanLoss"
  bottom: "score"
  bottom: "label_won"
  top: "loss_score"
  loss_weight: 0.01
}

layer {
  name: "accuracy_1"
  type: "Accuracy"
  bottom: "pol2"
  bottom: "label_move"
  top: "accuracy_1"
  include { phase: TEST }
}