3

caffe'doc 说:

layer {
  name: "upsample", type: "Deconvolution"
  bottom: "{{bottom_name}}" top: "{{top_name}}"
  convolution_param {
    kernel_size: {{2 * factor - factor % 2}} stride: {{factor}}
    num_output: {{C}} group: {{C}}
    pad: {{ceil((factor - 1) / 2.)}}
    weight_filler: { type: "bilinear" } bias_term: false
  }
  param { lr_mult: 0 decay_mult: 0 }
}

我不知道为什么要这样设置 kenrel_size、stride 和 pad?

4

1 回答 1

1

对于上采样,如果您希望调整大小因子为 2,则参数将为 kernel_size: 4, stride:2, pad:1

于 2018-08-08T06:46:04.993 回答