Skip to content

Latest commit

 

History

History
318 lines (309 loc) · 4.78 KB

README.md

File metadata and controls

318 lines (309 loc) · 4.78 KB
Raw

MathorNet

This is just a small convolutional neural network designed for the final assignment

MathorNet is based on LeNet and Imitate Vgg. The accuracy of 84.33 was finally obtained on the cifar-10 test data set. You can refer to Inception V3 or ResNet for retrofit, which may be better

View Code

If you want to know more detailed parameter Settings, can copy the code below to http://ethereon.github.io/netscope/#/editor

name: "MathorNet"
input: "data"
input_dim: 60000
input_dim: 3
input_dim: 32
input_dim: 32
layers {
  bottom: "data"
  top: "conv1_1"
  name: "conv1_1"
  type: CONVOLUTION
  convolution_param {
    num_output: 64
    pad: 1
    kernel_size: 3
  }
}
layers {
  bottom: "conv1_1"
  top: "conv1_1"
  name: "conv1_2"
  type: CONVOLUTION
  convolution_param {
    num_output: 64
    pad: 1
    kernel_size: 3
  }
}
layers {
  bottom: "conv1_1"
  top: "conv1_2"
  name: "pool1"
  type: POOLING
  pooling_param {
    pool: MAX
    kernel_size: 2
    stride: 2
  }
}
layers {
  bottom: "conv1_2"
  top: "pool1"
  name: "relu1"
  type: RELU
}
################ 第一部分结束
layers {
  bottom: "pool1"
  top: "conv2_1"
  name: "conv2_1"
  type: CONVOLUTION
  convolution_param {
    num_output: 128
    pad: 1
    kernel_size: 3
  }
}
layers {
  bottom: "conv2_1"
  top: "conv2_1"
  name: "conv2_2"
  type: CONVOLUTION
  convolution_param {
    num_output: 128
    pad: 1
    kernel_size: 3
  }
}
layers {
  bottom: "conv2_1"
  top: "conv2_2"
  name: "pool2"
  type: POOLING
  pooling_param {
    pool: MAX
    kernel_size: 2
    pad: 1
    stride: 2
  }
}
layers {
  bottom: "conv2_2"
  top: "pool2"
  name: "relu2"
  type: RELU
}
################ 第二部分结束
layers {
  bottom: "pool2"
  top: "conv3_1"
  name: "conv3_1"
  type: CONVOLUTION
  convolution_param {
    num_output: 128
    pad: 1
    kernel_size: 3
  }
}
layers {
  bottom: "conv3_1"
  top: "conv3_1"
  name: "conv3_2"
  type: CONVOLUTION
  convolution_param {
    num_output: 128
    pad: 1
    kernel_size: 3
  }
}
layers {
  bottom: "conv3_1"
  top: "conv3_1"
  name: "conv3_3"
  type: CONVOLUTION
  convolution_param {
    num_output: 128
    pad: 1
    kernel_size: 1
  }
}
layers {
  bottom: "conv3_1"
  top: "conv3_3"
  name: "pool3"
  type: POOLING
  pooling_param {
    pool: MAX
    kernel_size: 2
    pad: 1
    stride: 2
  }
}
layers {
  bottom: "conv3_3"
  top: "pool3"
  name: "relu3"
  type: RELU
}
################ 第三部分结束
layers {
  bottom: "pool3"
  top: "conv4_1"
  name: "conv4_1"
  type: CONVOLUTION
  convolution_param {
    num_output: 256
    pad: 1
    kernel_size: 3
  }
}
layers {
  bottom: "conv4_1"
  top: "conv4_1"
  name: "conv4_2"
  type: CONVOLUTION
  convolution_param {
    num_output: 256
    pad: 1
    kernel_size: 3
  }
}
layers {
  bottom: "conv4_1"
  top: "conv4_1"
  name: "conv4_3"
  type: CONVOLUTION
  convolution_param {
    num_output: 256
    pad: 1
    kernel_size: 1
  }
}
layers {
  bottom: "conv4_1"
  top: "conv4_3"
  name: "pool4"
  type: POOLING
  pooling_param {
    pool: MAX
    kernel_size: 2
    pad: 1
    stride: 2
  }
}
layers {
  bottom: "conv4_3"
  top: "pool4"
  name: "relu4"
  type: RELU
}
################ 第四部分结束
layers {
  bottom: "pool4"
  top: "conv5_1"
  name: "conv5_1"
  type: CONVOLUTION
  convolution_param {
    num_output: 512
    pad: 1
    kernel_size: 3
  }
}
layers {
  bottom: "conv5_1"
  top: "conv5_1"
  name: "conv5_2"
  type: CONVOLUTION
  convolution_param {
    num_output: 512
    pad: 1
    kernel_size: 3
  }
}
layers {
  bottom: "conv5_1"
  top: "conv5_1"
  name: "conv5_3"
  type: CONVOLUTION
  convolution_param {
    num_output: 512
    pad: 1
    kernel_size: 1
  }
}
layers {
  bottom: "conv5_1"
  top: "conv5_3"
  name: "pool5"
  type: POOLING
  pooling_param {
    pool: MAX
    kernel_size: 2
    pad: 1
    stride: 2
  }
}
layers {
  bottom: "conv5_3"
  top: "pool5"
  name: "relu5"
  type: RELU
}
################ 第五部分结束
layers {
  bottom: "pool5"
  top: "fc1"
  name: "fc1"
  type: INNER_PRODUCT
  inner_product_param {
    num_output: 1024
  }
}
layers {
  bottom: "fc1"
  top: "fc1"
  name: "relu1"
  type: RELU
}
layers {
  bottom: "fc1"
  top: "fc1"
  name: "drop1"
  type: DROPOUT
  dropout_param {
    dropout_ratio: 0.5
  }
}
layers {
  bottom: "fc1"
  top: "fc2"
  name: "fc2"
  type: INNER_PRODUCT
  inner_product_param {
    num_output: 1024
  }
}
layers {
  bottom: "fc2"
  top: "fc2"
  name: "relu2"
  type: RELU
}
layers {
  bottom: "fc2"
  top: "fc2"
  name: "drop2"
  type: DROPOUT
  dropout_param {
    dropout_ratio: 0.5
  }
}
layers {
  bottom: "fc2"
  top: "fc3"
  name: "fc3"
  type: INNER_PRODUCT
  inner_product_param {
    num_output: 10
  }
}

################ 全连接结束