Tao Hu dongzhuoyao

## gist:ca13fc90d6027dec692ba0a6a5e06d49
name: "xception"
layer {
  name: "data"
  type: "ImageData"
  top: "data"
  top: "label"
  image_data_param {
    new_dim: 256
    bicubic: true
    shuffle: true

## gist:d19116c29690ddf47baf9930ef29480b
name: "shufflenet"
#  transform_param {
#    scale: 0.017
#    mirror: false
#    crop_size: 224
#    mean_value: [103.94,116.78,123.68]
#  }
input: "data"
input_shape {
  dim: 1

## gist:ce196e665d5ae43529ed5ade940d096e
name: "ResNet-50"
input: "data"
input_dim: 1
input_dim: 3
input_dim: 224
input_dim: 224

layer {
	bottom: "data"
	top: "conv1"

## gist:b231dfa05b23a0eb0559da4d0e19f81c
# please cite:
# @article{SqueezeNet,
#     Author = {Forrest N. Iandola and Matthew W. Moskewicz and Khalid Ashraf and Song Han and William J. Dally and Kurt Keutzer},
#     Title = {SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and $<$1MB model size},
#     Journal = {arXiv:1602.07360},
#     Year = {2016}
#}

input: "data"
input_shape {

## gist:8457a56fba3c7c4bb3d3c6bc06abd79e
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

## train_voc_train_aug.prototxt
# VGG 16-layer network convolutional finetuning
# Network modified to have smaller receptive field (128 pixels)
# and smaller stride (8 pixels) when run in convolutional mode.
#
# In this model we also change max pooling size in the first 4 layer
# from 2 to 3 while retaining stride = 2
# which makes it easier to exactly align responses at different layer.
#
# For alignment to work, we set (we choose 32x so as to be able to evaluate
# the model for all different subsampling sizes):

## virtualenvwrapper.md

      
              1 file
            
          
              1 fork
            
          
              0 comments
            
          
              9 stars
            
          
                dongzhuoyao
                / virtualenvwrapper.md
            
            
              Last active
              September 11, 2023 19:22
                — forked from whhone/virtualenvwrapper.md
            
              
                virtualenvwrapper Cheat Sheet
              
          
    local setting

PATH=$PATH:~/.local/bin
source /home/tao/.local/bin/virtualenvwrapper.sh
pip install virtualenvwrapper --user #make everything locally!!

vim, zsh is very hard to install locally without sudo permission.

  
## xception
name: "xception"
layer {
  name: "data"
  type: "Data"
  top: "data"
  top: "label"
  transform_param {
    mirror: false
    crop_size: 224
    mean_value: 104.0

## gist:952fa930c43576a94c805a045037b943
name: "shufflenet"
#  transform_param {
#    scale: 0.017
#    mirror: false
#    crop_size: 224
#    mean_value: [103.94,116.78,123.68]
#  }
input: "data"
input_shape {
  dim: 1

## xception
name: "xception"
layer {
  name: "data"
  type: "Data"
  top: "data"
  top: "label"
  transform_param {
    mirror: false
    crop_size: 224
    mean_value: 104.0
	name: "xception"
	layer {
	name: "data"
	type: "ImageData"
	top: "data"
	top: "label"
	image_data_param {
	new_dim: 256
	bicubic: true
	shuffle: true
	name: "shufflenet"
	# transform_param {
	# scale: 0.017
	# mirror: false
	# crop_size: 224
	# mean_value: [103.94,116.78,123.68]
	# }
	input: "data"
	input_shape {
	dim: 1
	name: "ResNet-50"
	input: "data"
	input_dim: 1
	input_dim: 3
	input_dim: 224
	input_dim: 224

	layer {
	bottom: "data"
	top: "conv1"
	# please cite:
	# @article{SqueezeNet,
	# Author = {Forrest N. Iandola and Matthew W. Moskewicz and Khalid Ashraf and Song Han and William J. Dally and Kurt Keutzer},
	# Title = {SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and $<$1MB model size},
	# Journal = {arXiv:1602.07360},
	# Year = {2016}
	#}

	input: "data"
	input_shape {
	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
	# VGG 16-layer network convolutional finetuning
	# Network modified to have smaller receptive field (128 pixels)
	# and smaller stride (8 pixels) when run in convolutional mode.
	#
	# In this model we also change max pooling size in the first 4 layer
	# from 2 to 3 while retaining stride = 2
	# which makes it easier to exactly align responses at different layer.
	#
	# For alignment to work, we set (we choose 32x so as to be able to evaluate
	# the model for all different subsampling sizes):