ddlee-cn

name: "ResNeXt"
input: "data"
input_shape {
  dim: 1
  dim: 3
  dim: 224
  dim: 224
}
layer {
  name: "conv1"

ddlee-cn

name: "DenseNet"
input: "data"
input_shape {
  dim: 1
  dim: 3
  dim: 224
  dim: 224
}
layer {
  name: "conv1"

ddlee-cn

Xception

as in paper

https://arxiv.org/abs/1610.02357

ddlee-cn

name: "Inception-ResNet-v2"
input: "data"
input_shape {
  dim: 1
  dim: 3
  dim: 299 
  dim: 299
}
layer {
  name: "conv1_3x3_s2"

ddlee-cn

Inception-v4

as in paper

https://arxiv.org/abs/1602.07261

ddlee-cn

Inception-v3 architecture

as in paper

https://arxiv.org/abs/1512.00567

ddlee-cn

name: "MOBILENET"
#  transform_param {
#    scale: 0.017
#    mirror: false
#    crop_size: 224
#    mean_value: [103.94,116.78,123.68]
#  }
input: "data"
input_dim: 1
input_dim: 3

ddlee-cn

Xception: Depthwise Separable Convolutions

The Xception Paper: https://arxiv.org/abs/1610.02357

Caffe model from https://github.com/yihui-he/Xception-caffe

ddlee-cn

ResNet 50

The ResNet Paper: http://arxiv.org/abs/1512.03385

Note this file is only for visualization, not for training or deployment.

ddlee-cn

name: "nin_imagenet"
layers {
  top: "data"
  top: "label"
  name: "data"
  type: DATA
  data_param {
    source: "/IMAGENET-LMDB/imagenet-train-lmdb"
    backend: LMDB
    batch_size: 64