classification.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Classifying images with Caffe\n",
    "\n",
    "We will classify some images with the pretrained GoogLeNet model that is available from [here](https://github.com/BVLC/caffe/tree/master/models/bvlc_googlenet). GoogLeNet is an convolutional network architecture that won the Imagenet competition in 2014. For more details on the architecture, see [this paper](http://arxiv.org/abs/1409.4842).\n",
    "\n",
    "![img](./images/googlenet_diagram.png)\n",
    "\n",
    "We assume that we're running from EC2 AMI, otherwise you need to download the googlenet model and set ```caffe_root``` and ```model_root``` variables below.\n",
    "\n",
    "What we need for a pretrained model, is these files:\n",
    "* deploy.prototxt - the model definition file\n",
    "* bvlc_googlenet.caffemodel - a binary file with the learned model parameters\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "name: \"GoogleNet\"\r\n",
      "input: \"data\"\r\n",
      "input_shape {\r\n",
      "  dim: 10\r\n",
      "  dim: 3\r\n",
      "  dim: 224\r\n",
      "  dim: 224\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"conv1/7x7_s2\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"data\"\r\n",
      "  top: \"conv1/7x7_s2\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 64\r\n",
      "    pad: 3\r\n",
      "    kernel_size: 7\r\n",
      "    stride: 2\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.1\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"conv1/relu_7x7\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"conv1/7x7_s2\"\r\n",
      "  top: \"conv1/7x7_s2\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"pool1/3x3_s2\"\r\n",
      "  type: \"Pooling\"\r\n",
      "  bottom: \"conv1/7x7_s2\"\r\n",
      "  top: \"pool1/3x3_s2\"\r\n",
      "  pooling_param {\r\n",
      "    pool: MAX\r\n",
      "    kernel_size: 3\r\n",
      "    stride: 2\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"pool1/norm1\"\r\n",
      "  type: \"LRN\"\r\n",
      "  bottom: \"pool1/3x3_s2\"\r\n",
      "  top: \"pool1/norm1\"\r\n",
      "  lrn_param {\r\n",
      "    local_size: 5\r\n",
      "    alpha: 0.0001\r\n",
      "    beta: 0.75\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"conv2/3x3_reduce\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"pool1/norm1\"\r\n",
      "  top: \"conv2/3x3_reduce\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 64\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.1\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"conv2/relu_3x3_reduce\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"conv2/3x3_reduce\"\r\n",
      "  top: \"conv2/3x3_reduce\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"conv2/3x3\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"conv2/3x3_reduce\"\r\n",
      "  top: \"conv2/3x3\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 192\r\n",
      "    pad: 1\r\n",
      "    kernel_size: 3\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"conv2/relu_3x3\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"conv2/3x3\"\r\n",
      "  top: \"conv2/3x3\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"conv2/norm2\"\r\n",
      "  type: \"LRN\"\r\n",
      "  bottom: \"conv2/3x3\"\r\n",
      "  top: \"conv2/norm2\"\r\n",
      "  lrn_param {\r\n",
      "    local_size: 5\r\n",
      "    alpha: 0.0001\r\n",
      "    beta: 0.75\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"pool2/3x3_s2\"\r\n",
      "  type: \"Pooling\"\r\n",
      "  bottom: \"conv2/norm2\"\r\n",
      "  top: \"pool2/3x3_s2\"\r\n",
      "  pooling_param {\r\n",
      "    pool: MAX\r\n",
      "    kernel_size: 3\r\n",
      "    stride: 2\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3a/1x1\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"pool2/3x3_s2\"\r\n",
      "  top: \"inception_3a/1x1\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 64\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3a/relu_1x1\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_3a/1x1\"\r\n",
      "  top: \"inception_3a/1x1\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3a/3x3_reduce\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"pool2/3x3_s2\"\r\n",
      "  top: \"inception_3a/3x3_reduce\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 96\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.09\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3a/relu_3x3_reduce\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_3a/3x3_reduce\"\r\n",
      "  top: \"inception_3a/3x3_reduce\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3a/3x3\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_3a/3x3_reduce\"\r\n",
      "  top: \"inception_3a/3x3\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 128\r\n",
      "    pad: 1\r\n",
      "    kernel_size: 3\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3a/relu_3x3\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_3a/3x3\"\r\n",
      "  top: \"inception_3a/3x3\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3a/5x5_reduce\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"pool2/3x3_s2\"\r\n",
      "  top: \"inception_3a/5x5_reduce\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 16\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.2\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3a/relu_5x5_reduce\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_3a/5x5_reduce\"\r\n",
      "  top: \"inception_3a/5x5_reduce\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3a/5x5\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_3a/5x5_reduce\"\r\n",
      "  top: \"inception_3a/5x5\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 32\r\n",
      "    pad: 2\r\n",
      "    kernel_size: 5\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3a/relu_5x5\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_3a/5x5\"\r\n",
      "  top: \"inception_3a/5x5\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3a/pool\"\r\n",
      "  type: \"Pooling\"\r\n",
      "  bottom: \"pool2/3x3_s2\"\r\n",
      "  top: \"inception_3a/pool\"\r\n",
      "  pooling_param {\r\n",
      "    pool: MAX\r\n",
      "    kernel_size: 3\r\n",
      "    stride: 1\r\n",
      "    pad: 1\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3a/pool_proj\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_3a/pool\"\r\n",
      "  top: \"inception_3a/pool_proj\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 32\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.1\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3a/relu_pool_proj\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_3a/pool_proj\"\r\n",
      "  top: \"inception_3a/pool_proj\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3a/output\"\r\n",
      "  type: \"Concat\"\r\n",
      "  bottom: \"inception_3a/1x1\"\r\n",
      "  bottom: \"inception_3a/3x3\"\r\n",
      "  bottom: \"inception_3a/5x5\"\r\n",
      "  bottom: \"inception_3a/pool_proj\"\r\n",
      "  top: \"inception_3a/output\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3b/1x1\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_3a/output\"\r\n",
      "  top: \"inception_3b/1x1\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 128\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3b/relu_1x1\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_3b/1x1\"\r\n",
      "  top: \"inception_3b/1x1\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3b/3x3_reduce\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_3a/output\"\r\n",
      "  top: \"inception_3b/3x3_reduce\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 128\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.09\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3b/relu_3x3_reduce\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_3b/3x3_reduce\"\r\n",
      "  top: \"inception_3b/3x3_reduce\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3b/3x3\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_3b/3x3_reduce\"\r\n",
      "  top: \"inception_3b/3x3\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 192\r\n",
      "    pad: 1\r\n",
      "    kernel_size: 3\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3b/relu_3x3\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_3b/3x3\"\r\n",
      "  top: \"inception_3b/3x3\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3b/5x5_reduce\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_3a/output\"\r\n",
      "  top: \"inception_3b/5x5_reduce\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 32\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.2\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3b/relu_5x5_reduce\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_3b/5x5_reduce\"\r\n",
      "  top: \"inception_3b/5x5_reduce\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3b/5x5\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_3b/5x5_reduce\"\r\n",
      "  top: \"inception_3b/5x5\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 96\r\n",
      "    pad: 2\r\n",
      "    kernel_size: 5\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3b/relu_5x5\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_3b/5x5\"\r\n",
      "  top: \"inception_3b/5x5\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3b/pool\"\r\n",
      "  type: \"Pooling\"\r\n",
      "  bottom: \"inception_3a/output\"\r\n",
      "  top: \"inception_3b/pool\"\r\n",
      "  pooling_param {\r\n",
      "    pool: MAX\r\n",
      "    kernel_size: 3\r\n",
      "    stride: 1\r\n",
      "    pad: 1\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3b/pool_proj\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_3b/pool\"\r\n",
      "  top: \"inception_3b/pool_proj\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 64\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.1\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3b/relu_pool_proj\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_3b/pool_proj\"\r\n",
      "  top: \"inception_3b/pool_proj\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_3b/output\"\r\n",
      "  type: \"Concat\"\r\n",
      "  bottom: \"inception_3b/1x1\"\r\n",
      "  bottom: \"inception_3b/3x3\"\r\n",
      "  bottom: \"inception_3b/5x5\"\r\n",
      "  bottom: \"inception_3b/pool_proj\"\r\n",
      "  top: \"inception_3b/output\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"pool3/3x3_s2\"\r\n",
      "  type: \"Pooling\"\r\n",
      "  bottom: \"inception_3b/output\"\r\n",
      "  top: \"pool3/3x3_s2\"\r\n",
      "  pooling_param {\r\n",
      "    pool: MAX\r\n",
      "    kernel_size: 3\r\n",
      "    stride: 2\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4a/1x1\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"pool3/3x3_s2\"\r\n",
      "  top: \"inception_4a/1x1\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 192\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4a/relu_1x1\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4a/1x1\"\r\n",
      "  top: \"inception_4a/1x1\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4a/3x3_reduce\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"pool3/3x3_s2\"\r\n",
      "  top: \"inception_4a/3x3_reduce\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 96\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.09\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4a/relu_3x3_reduce\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4a/3x3_reduce\"\r\n",
      "  top: \"inception_4a/3x3_reduce\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4a/3x3\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4a/3x3_reduce\"\r\n",
      "  top: \"inception_4a/3x3\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 208\r\n",
      "    pad: 1\r\n",
      "    kernel_size: 3\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4a/relu_3x3\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4a/3x3\"\r\n",
      "  top: \"inception_4a/3x3\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4a/5x5_reduce\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"pool3/3x3_s2\"\r\n",
      "  top: \"inception_4a/5x5_reduce\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 16\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.2\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4a/relu_5x5_reduce\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4a/5x5_reduce\"\r\n",
      "  top: \"inception_4a/5x5_reduce\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4a/5x5\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4a/5x5_reduce\"\r\n",
      "  top: \"inception_4a/5x5\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 48\r\n",
      "    pad: 2\r\n",
      "    kernel_size: 5\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4a/relu_5x5\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4a/5x5\"\r\n",
      "  top: \"inception_4a/5x5\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4a/pool\"\r\n",
      "  type: \"Pooling\"\r\n",
      "  bottom: \"pool3/3x3_s2\"\r\n",
      "  top: \"inception_4a/pool\"\r\n",
      "  pooling_param {\r\n",
      "    pool: MAX\r\n",
      "    kernel_size: 3\r\n",
      "    stride: 1\r\n",
      "    pad: 1\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4a/pool_proj\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4a/pool\"\r\n",
      "  top: \"inception_4a/pool_proj\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 64\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.1\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4a/relu_pool_proj\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4a/pool_proj\"\r\n",
      "  top: \"inception_4a/pool_proj\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4a/output\"\r\n",
      "  type: \"Concat\"\r\n",
      "  bottom: \"inception_4a/1x1\"\r\n",
      "  bottom: \"inception_4a/3x3\"\r\n",
      "  bottom: \"inception_4a/5x5\"\r\n",
      "  bottom: \"inception_4a/pool_proj\"\r\n",
      "  top: \"inception_4a/output\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4b/1x1\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4a/output\"\r\n",
      "  top: \"inception_4b/1x1\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 160\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4b/relu_1x1\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4b/1x1\"\r\n",
      "  top: \"inception_4b/1x1\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4b/3x3_reduce\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4a/output\"\r\n",
      "  top: \"inception_4b/3x3_reduce\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 112\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.09\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4b/relu_3x3_reduce\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4b/3x3_reduce\"\r\n",
      "  top: \"inception_4b/3x3_reduce\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4b/3x3\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4b/3x3_reduce\"\r\n",
      "  top: \"inception_4b/3x3\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 224\r\n",
      "    pad: 1\r\n",
      "    kernel_size: 3\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4b/relu_3x3\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4b/3x3\"\r\n",
      "  top: \"inception_4b/3x3\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4b/5x5_reduce\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4a/output\"\r\n",
      "  top: \"inception_4b/5x5_reduce\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 24\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.2\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4b/relu_5x5_reduce\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4b/5x5_reduce\"\r\n",
      "  top: \"inception_4b/5x5_reduce\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4b/5x5\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4b/5x5_reduce\"\r\n",
      "  top: \"inception_4b/5x5\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 64\r\n",
      "    pad: 2\r\n",
      "    kernel_size: 5\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4b/relu_5x5\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4b/5x5\"\r\n",
      "  top: \"inception_4b/5x5\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4b/pool\"\r\n",
      "  type: \"Pooling\"\r\n",
      "  bottom: \"inception_4a/output\"\r\n",
      "  top: \"inception_4b/pool\"\r\n",
      "  pooling_param {\r\n",
      "    pool: MAX\r\n",
      "    kernel_size: 3\r\n",
      "    stride: 1\r\n",
      "    pad: 1\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4b/pool_proj\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4b/pool\"\r\n",
      "  top: \"inception_4b/pool_proj\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 64\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.1\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4b/relu_pool_proj\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4b/pool_proj\"\r\n",
      "  top: \"inception_4b/pool_proj\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4b/output\"\r\n",
      "  type: \"Concat\"\r\n",
      "  bottom: \"inception_4b/1x1\"\r\n",
      "  bottom: \"inception_4b/3x3\"\r\n",
      "  bottom: \"inception_4b/5x5\"\r\n",
      "  bottom: \"inception_4b/pool_proj\"\r\n",
      "  top: \"inception_4b/output\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4c/1x1\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4b/output\"\r\n",
      "  top: \"inception_4c/1x1\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 128\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4c/relu_1x1\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4c/1x1\"\r\n",
      "  top: \"inception_4c/1x1\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4c/3x3_reduce\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4b/output\"\r\n",
      "  top: \"inception_4c/3x3_reduce\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 128\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.09\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4c/relu_3x3_reduce\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4c/3x3_reduce\"\r\n",
      "  top: \"inception_4c/3x3_reduce\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4c/3x3\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4c/3x3_reduce\"\r\n",
      "  top: \"inception_4c/3x3\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 256\r\n",
      "    pad: 1\r\n",
      "    kernel_size: 3\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4c/relu_3x3\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4c/3x3\"\r\n",
      "  top: \"inception_4c/3x3\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4c/5x5_reduce\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4b/output\"\r\n",
      "  top: \"inception_4c/5x5_reduce\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 24\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.2\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4c/relu_5x5_reduce\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4c/5x5_reduce\"\r\n",
      "  top: \"inception_4c/5x5_reduce\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4c/5x5\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4c/5x5_reduce\"\r\n",
      "  top: \"inception_4c/5x5\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 64\r\n",
      "    pad: 2\r\n",
      "    kernel_size: 5\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4c/relu_5x5\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4c/5x5\"\r\n",
      "  top: \"inception_4c/5x5\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4c/pool\"\r\n",
      "  type: \"Pooling\"\r\n",
      "  bottom: \"inception_4b/output\"\r\n",
      "  top: \"inception_4c/pool\"\r\n",
      "  pooling_param {\r\n",
      "    pool: MAX\r\n",
      "    kernel_size: 3\r\n",
      "    stride: 1\r\n",
      "    pad: 1\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4c/pool_proj\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4c/pool\"\r\n",
      "  top: \"inception_4c/pool_proj\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 64\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.1\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4c/relu_pool_proj\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4c/pool_proj\"\r\n",
      "  top: \"inception_4c/pool_proj\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4c/output\"\r\n",
      "  type: \"Concat\"\r\n",
      "  bottom: \"inception_4c/1x1\"\r\n",
      "  bottom: \"inception_4c/3x3\"\r\n",
      "  bottom: \"inception_4c/5x5\"\r\n",
      "  bottom: \"inception_4c/pool_proj\"\r\n",
      "  top: \"inception_4c/output\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4d/1x1\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4c/output\"\r\n",
      "  top: \"inception_4d/1x1\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 112\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4d/relu_1x1\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4d/1x1\"\r\n",
      "  top: \"inception_4d/1x1\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4d/3x3_reduce\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4c/output\"\r\n",
      "  top: \"inception_4d/3x3_reduce\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 144\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.09\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4d/relu_3x3_reduce\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4d/3x3_reduce\"\r\n",
      "  top: \"inception_4d/3x3_reduce\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4d/3x3\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4d/3x3_reduce\"\r\n",
      "  top: \"inception_4d/3x3\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 288\r\n",
      "    pad: 1\r\n",
      "    kernel_size: 3\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4d/relu_3x3\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4d/3x3\"\r\n",
      "  top: \"inception_4d/3x3\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4d/5x5_reduce\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4c/output\"\r\n",
      "  top: \"inception_4d/5x5_reduce\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 32\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.2\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4d/relu_5x5_reduce\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4d/5x5_reduce\"\r\n",
      "  top: \"inception_4d/5x5_reduce\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4d/5x5\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4d/5x5_reduce\"\r\n",
      "  top: \"inception_4d/5x5\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 64\r\n",
      "    pad: 2\r\n",
      "    kernel_size: 5\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4d/relu_5x5\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4d/5x5\"\r\n",
      "  top: \"inception_4d/5x5\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4d/pool\"\r\n",
      "  type: \"Pooling\"\r\n",
      "  bottom: \"inception_4c/output\"\r\n",
      "  top: \"inception_4d/pool\"\r\n",
      "  pooling_param {\r\n",
      "    pool: MAX\r\n",
      "    kernel_size: 3\r\n",
      "    stride: 1\r\n",
      "    pad: 1\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4d/pool_proj\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4d/pool\"\r\n",
      "  top: \"inception_4d/pool_proj\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 64\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.1\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4d/relu_pool_proj\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4d/pool_proj\"\r\n",
      "  top: \"inception_4d/pool_proj\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4d/output\"\r\n",
      "  type: \"Concat\"\r\n",
      "  bottom: \"inception_4d/1x1\"\r\n",
      "  bottom: \"inception_4d/3x3\"\r\n",
      "  bottom: \"inception_4d/5x5\"\r\n",
      "  bottom: \"inception_4d/pool_proj\"\r\n",
      "  top: \"inception_4d/output\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4e/1x1\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4d/output\"\r\n",
      "  top: \"inception_4e/1x1\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 256\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4e/relu_1x1\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4e/1x1\"\r\n",
      "  top: \"inception_4e/1x1\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4e/3x3_reduce\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4d/output\"\r\n",
      "  top: \"inception_4e/3x3_reduce\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 160\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.09\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4e/relu_3x3_reduce\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4e/3x3_reduce\"\r\n",
      "  top: \"inception_4e/3x3_reduce\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4e/3x3\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4e/3x3_reduce\"\r\n",
      "  top: \"inception_4e/3x3\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 320\r\n",
      "    pad: 1\r\n",
      "    kernel_size: 3\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4e/relu_3x3\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4e/3x3\"\r\n",
      "  top: \"inception_4e/3x3\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4e/5x5_reduce\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4d/output\"\r\n",
      "  top: \"inception_4e/5x5_reduce\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 32\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.2\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4e/relu_5x5_reduce\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4e/5x5_reduce\"\r\n",
      "  top: \"inception_4e/5x5_reduce\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4e/5x5\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4e/5x5_reduce\"\r\n",
      "  top: \"inception_4e/5x5\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 128\r\n",
      "    pad: 2\r\n",
      "    kernel_size: 5\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4e/relu_5x5\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4e/5x5\"\r\n",
      "  top: \"inception_4e/5x5\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4e/pool\"\r\n",
      "  type: \"Pooling\"\r\n",
      "  bottom: \"inception_4d/output\"\r\n",
      "  top: \"inception_4e/pool\"\r\n",
      "  pooling_param {\r\n",
      "    pool: MAX\r\n",
      "    kernel_size: 3\r\n",
      "    stride: 1\r\n",
      "    pad: 1\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4e/pool_proj\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_4e/pool\"\r\n",
      "  top: \"inception_4e/pool_proj\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 128\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.1\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4e/relu_pool_proj\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_4e/pool_proj\"\r\n",
      "  top: \"inception_4e/pool_proj\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_4e/output\"\r\n",
      "  type: \"Concat\"\r\n",
      "  bottom: \"inception_4e/1x1\"\r\n",
      "  bottom: \"inception_4e/3x3\"\r\n",
      "  bottom: \"inception_4e/5x5\"\r\n",
      "  bottom: \"inception_4e/pool_proj\"\r\n",
      "  top: \"inception_4e/output\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"pool4/3x3_s2\"\r\n",
      "  type: \"Pooling\"\r\n",
      "  bottom: \"inception_4e/output\"\r\n",
      "  top: \"pool4/3x3_s2\"\r\n",
      "  pooling_param {\r\n",
      "    pool: MAX\r\n",
      "    kernel_size: 3\r\n",
      "    stride: 2\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5a/1x1\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"pool4/3x3_s2\"\r\n",
      "  top: \"inception_5a/1x1\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 256\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5a/relu_1x1\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_5a/1x1\"\r\n",
      "  top: \"inception_5a/1x1\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5a/3x3_reduce\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"pool4/3x3_s2\"\r\n",
      "  top: \"inception_5a/3x3_reduce\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 160\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.09\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5a/relu_3x3_reduce\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_5a/3x3_reduce\"\r\n",
      "  top: \"inception_5a/3x3_reduce\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5a/3x3\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_5a/3x3_reduce\"\r\n",
      "  top: \"inception_5a/3x3\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 320\r\n",
      "    pad: 1\r\n",
      "    kernel_size: 3\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5a/relu_3x3\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_5a/3x3\"\r\n",
      "  top: \"inception_5a/3x3\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5a/5x5_reduce\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"pool4/3x3_s2\"\r\n",
      "  top: \"inception_5a/5x5_reduce\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 32\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.2\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5a/relu_5x5_reduce\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_5a/5x5_reduce\"\r\n",
      "  top: \"inception_5a/5x5_reduce\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5a/5x5\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_5a/5x5_reduce\"\r\n",
      "  top: \"inception_5a/5x5\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 128\r\n",
      "    pad: 2\r\n",
      "    kernel_size: 5\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5a/relu_5x5\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_5a/5x5\"\r\n",
      "  top: \"inception_5a/5x5\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5a/pool\"\r\n",
      "  type: \"Pooling\"\r\n",
      "  bottom: \"pool4/3x3_s2\"\r\n",
      "  top: \"inception_5a/pool\"\r\n",
      "  pooling_param {\r\n",
      "    pool: MAX\r\n",
      "    kernel_size: 3\r\n",
      "    stride: 1\r\n",
      "    pad: 1\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5a/pool_proj\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_5a/pool\"\r\n",
      "  top: \"inception_5a/pool_proj\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 128\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.1\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5a/relu_pool_proj\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_5a/pool_proj\"\r\n",
      "  top: \"inception_5a/pool_proj\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5a/output\"\r\n",
      "  type: \"Concat\"\r\n",
      "  bottom: \"inception_5a/1x1\"\r\n",
      "  bottom: \"inception_5a/3x3\"\r\n",
      "  bottom: \"inception_5a/5x5\"\r\n",
      "  bottom: \"inception_5a/pool_proj\"\r\n",
      "  top: \"inception_5a/output\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5b/1x1\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_5a/output\"\r\n",
      "  top: \"inception_5b/1x1\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 384\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5b/relu_1x1\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_5b/1x1\"\r\n",
      "  top: \"inception_5b/1x1\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5b/3x3_reduce\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_5a/output\"\r\n",
      "  top: \"inception_5b/3x3_reduce\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 192\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.09\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5b/relu_3x3_reduce\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_5b/3x3_reduce\"\r\n",
      "  top: \"inception_5b/3x3_reduce\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5b/3x3\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_5b/3x3_reduce\"\r\n",
      "  top: \"inception_5b/3x3\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 384\r\n",
      "    pad: 1\r\n",
      "    kernel_size: 3\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5b/relu_3x3\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_5b/3x3\"\r\n",
      "  top: \"inception_5b/3x3\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5b/5x5_reduce\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_5a/output\"\r\n",
      "  top: \"inception_5b/5x5_reduce\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 48\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.2\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5b/relu_5x5_reduce\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_5b/5x5_reduce\"\r\n",
      "  top: \"inception_5b/5x5_reduce\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5b/5x5\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_5b/5x5_reduce\"\r\n",
      "  top: \"inception_5b/5x5\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 128\r\n",
      "    pad: 2\r\n",
      "    kernel_size: 5\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.03\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5b/relu_5x5\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_5b/5x5\"\r\n",
      "  top: \"inception_5b/5x5\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5b/pool\"\r\n",
      "  type: \"Pooling\"\r\n",
      "  bottom: \"inception_5a/output\"\r\n",
      "  top: \"inception_5b/pool\"\r\n",
      "  pooling_param {\r\n",
      "    pool: MAX\r\n",
      "    kernel_size: 3\r\n",
      "    stride: 1\r\n",
      "    pad: 1\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5b/pool_proj\"\r\n",
      "  type: \"Convolution\"\r\n",
      "  bottom: \"inception_5b/pool\"\r\n",
      "  top: \"inception_5b/pool_proj\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  convolution_param {\r\n",
      "    num_output: 128\r\n",
      "    kernel_size: 1\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "      std: 0.1\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0.2\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5b/relu_pool_proj\"\r\n",
      "  type: \"ReLU\"\r\n",
      "  bottom: \"inception_5b/pool_proj\"\r\n",
      "  top: \"inception_5b/pool_proj\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"inception_5b/output\"\r\n",
      "  type: \"Concat\"\r\n",
      "  bottom: \"inception_5b/1x1\"\r\n",
      "  bottom: \"inception_5b/3x3\"\r\n",
      "  bottom: \"inception_5b/5x5\"\r\n",
      "  bottom: \"inception_5b/pool_proj\"\r\n",
      "  top: \"inception_5b/output\"\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"pool5/7x7_s1\"\r\n",
      "  type: \"Pooling\"\r\n",
      "  bottom: \"inception_5b/output\"\r\n",
      "  top: \"pool5/7x7_s1\"\r\n",
      "  pooling_param {\r\n",
      "    pool: AVE\r\n",
      "    kernel_size: 7\r\n",
      "    stride: 1\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"pool5/drop_7x7_s1\"\r\n",
      "  type: \"Dropout\"\r\n",
      "  bottom: \"pool5/7x7_s1\"\r\n",
      "  top: \"pool5/7x7_s1\"\r\n",
      "  dropout_param {\r\n",
      "    dropout_ratio: 0.4\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"loss3/classifier\"\r\n",
      "  type: \"InnerProduct\"\r\n",
      "  bottom: \"pool5/7x7_s1\"\r\n",
      "  top: \"loss3/classifier\"\r\n",
      "  param {\r\n",
      "    lr_mult: 1\r\n",
      "    decay_mult: 1\r\n",
      "  }\r\n",
      "  param {\r\n",
      "    lr_mult: 2\r\n",
      "    decay_mult: 0\r\n",
      "  }\r\n",
      "  inner_product_param {\r\n",
      "    num_output: 1000\r\n",
      "    weight_filler {\r\n",
      "      type: \"xavier\"\r\n",
      "    }\r\n",
      "    bias_filler {\r\n",
      "      type: \"constant\"\r\n",
      "      value: 0\r\n",
      "    }\r\n",
      "  }\r\n",
      "}\r\n",
      "layer {\r\n",
      "  name: \"prob\"\r\n",
      "  type: \"Softmax\"\r\n",
      "  bottom: \"loss3/classifier\"\r\n",
      "  top: \"prob\"\r\n",
      "}\r\n"
     ]
    }
   ],
   "source": [
    "!cat ~/caffe/models/bvlc_googlenet/deploy.prototxt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "\n",
    "caffe_root = '/home/ubuntu/caffe/'\n",
    "model_root = '/home/ubuntu/caffe/models/bvlc_googlenet/'\n",
    "# Make sure that caffe is on the python path:\n",
    "import sys\n",
    "sys.path.insert(0, caffe_root + 'python')\n",
    "import caffe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "caffe.set_mode_gpu()\n",
    "net = caffe.Net(model_root + 'deploy.prototxt', \n",
    "                model_root + 'bvlc_googlenet.caffemodel',\n",
    "                caffe.TEST)\n",
    "\n",
    "# input preprocessing: 'data' is the name of the input blob == net.inputs[0]\n",
    "transformer = caffe.io.Transformer({'data': net.blobs['data'].data.shape})\n",
    "transformer.set_transpose('data', (2,0,1))\n",
    "transformer.set_mean('data', np.load(caffe_root + 'python/caffe/imagenet/ilsvrc_2012_mean.npy').mean(1).mean(1)) # mean pixel\n",
    "transformer.set_raw_scale('data', 255)  # the reference model operates on images in [0,255] range instead of [0,1]\n",
    "transformer.set_channel_swap('data', (2,1,0))  # the reference model has channels in BGR order instead of RGB (opencv)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We want to classify this image of a norwegian elkhound:\n",
    "\n",
    "![](./images/norwegian_elkhound_small.jpg)\n",
    "\n",
    "Feed in the image (with some preprocessing) and classify with a forward pass."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Predicted class is #174.\n"
     ]
    }
   ],
   "source": [
    "im = caffe.io.load_image('./images/norwegian_elkhound_small.jpg')\n",
    "#im = caffe.io.load_image('./images/ant_small.jpg')\n",
    "#im = caffe.io.load_image('./images/strawberry_small.jpg')\n",
    "im = transformer.preprocess('data', im)\n",
    "net.blobs['data'].data[...] = im\n",
    "out = net.forward()\n",
    "out['prob']\n",
    "print(\"Predicted class is #{}.\".format(out['prob'].argmax()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Norwegian elkhound : 0.999874.3\n",
      "malamute : 0.000060.3\n",
      "keeshond : 0.000015.3\n",
      "German shepherd : 0.000014.3\n",
      "malinois : 0.000013.3\n"
     ]
    }
   ],
   "source": [
    "# load labels\n",
    "imagenet_labels_filename = caffe_root + 'data/ilsvrc12/synset_words.txt'\n",
    "labels = np.loadtxt(imagenet_labels_filename, str, delimiter='\\t')\n",
    "\n",
    "# sort top k predictions from softmax output\n",
    "top_k = net.blobs['prob'].data[0].flatten().argsort()[-1:-6:-1]\n",
    "#print labels[top_k]\n",
    "for k in top_k:\n",
    "    label_string = labels[k][10:].split(\",\")[0]\n",
    "    print \"%s : %f.3\" % (label_string, net.blobs['prob'].data[0].flatten()[k])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Another image, a magpie:\n",
    "\n",
    "![](./images/magpie_small.jpg)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "magpie : 0.999774.3\n",
      "water ouzel : 0.000080.3\n",
      "oystercatcher : 0.000045.3\n",
      "jay : 0.000022.3\n",
      "toucan : 0.000020.3\n"
     ]
    }
   ],
   "source": [
    "im = caffe.io.load_image('./images/magpie_small.jpg')\n",
    "im = transformer.preprocess('data', im)\n",
    "net.blobs['data'].data[...] = im\n",
    "out = net.forward()\n",
    "top_k = net.blobs['prob'].data[0].flatten().argsort()[-1:-6:-1]\n",
    "#print labels[top_k]\n",
    "for k in top_k:\n",
    "    label_string = labels[k][10:].split(\",\")[0]\n",
    "    print \"%s : %f.3\" % (label_string, net.blobs['prob'].data[0].flatten()[k])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A castle:\n",
    "\n",
    "![](./images/castle_small.jpg)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "castle : 0.964401.3\n",
      "palace : 0.024251.3\n",
      "monastery : 0.010008.3\n",
      "church : 0.000948.3\n",
      "bell cote : 0.000097.3\n"
     ]
    }
   ],
   "source": [
    "im = caffe.io.load_image('./images/castle_small.jpg')\n",
    "im = transformer.preprocess('data', im)\n",
    "net.blobs['data'].data[...] = im\n",
    "out = net.forward()\n",
    "top_k = net.blobs['prob'].data[0].flatten().argsort()[-1:-6:-1]\n",
    "#print labels[top_k]\n",
    "for k in top_k:\n",
    "    label_string = labels[k][10:].split(\",\")[0]\n",
    "    print \"%s : %f.3\" % (label_string, net.blobs['prob'].data[0].flatten()[k])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 6.51 s, sys: 78.9 ms, total: 6.58 s\n",
      "Wall time: 6.58 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "# CPU mode\n",
    "caffe.set_mode_cpu()\n",
    "net.forward()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 51.8 ms, sys: 24.1 ms, total: 75.8 ms\n",
      "Wall time: 75.5 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "# GPU mode\n",
    "caffe.set_mode_gpu()\n",
    "net.forward()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['n01440764 tench, Tinca tinca' 'n01443537 goldfish, Carassius auratus'\n",
      " 'n01484850 great white shark, white shark, man-eater, man-eating shark, Carcharodon carcharias'\n",
      " 'n01491361 tiger shark, Galeocerdo cuvieri'\n",
      " 'n01494475 hammerhead, hammerhead shark'\n",
      " 'n01496331 electric ray, crampfish, numbfish, torpedo'\n",
      " 'n01498041 stingray' 'n01514668 cock' 'n01514859 hen'\n",
      " 'n01518878 ostrich, Struthio camelus'\n",
      " 'n01530575 brambling, Fringilla montifringilla'\n",
      " 'n01531178 goldfinch, Carduelis carduelis'\n",
      " 'n01532829 house finch, linnet, Carpodacus mexicanus'\n",
      " 'n01534433 junco, snowbird'\n",
      " 'n01537544 indigo bunting, indigo finch, indigo bird, Passerina cyanea'\n",
      " 'n01558993 robin, American robin, Turdus migratorius' 'n01560419 bulbul'\n",
      " 'n01580077 jay' 'n01582220 magpie' 'n01592084 chickadee'\n",
      " 'n01601694 water ouzel, dipper' 'n01608432 kite'\n",
      " 'n01614925 bald eagle, American eagle, Haliaeetus leucocephalus'\n",
      " 'n01616318 vulture'\n",
      " 'n01622779 great grey owl, great gray owl, Strix nebulosa'\n",
      " 'n01629819 European fire salamander, Salamandra salamandra'\n",
      " 'n01630670 common newt, Triturus vulgaris' 'n01631663 eft'\n",
      " 'n01632458 spotted salamander, Ambystoma maculatum'\n",
      " 'n01632777 axolotl, mud puppy, Ambystoma mexicanum'\n",
      " 'n01641577 bullfrog, Rana catesbeiana' 'n01644373 tree frog, tree-frog'\n",
      " 'n01644900 tailed frog, bell toad, ribbed toad, tailed toad, Ascaphus trui'\n",
      " 'n01664065 loggerhead, loggerhead turtle, Caretta caretta'\n",
      " 'n01665541 leatherback turtle, leatherback, leathery turtle, Dermochelys coriacea'\n",
      " 'n01667114 mud turtle' 'n01667778 terrapin'\n",
      " 'n01669191 box turtle, box tortoise' 'n01675722 banded gecko'\n",
      " 'n01677366 common iguana, iguana, Iguana iguana'\n",
      " 'n01682714 American chameleon, anole, Anolis carolinensis'\n",
      " 'n01685808 whiptail, whiptail lizard' 'n01687978 agama'\n",
      " 'n01688243 frilled lizard, Chlamydosaurus kingi'\n",
      " 'n01689811 alligator lizard' 'n01692333 Gila monster, Heloderma suspectum'\n",
      " 'n01693334 green lizard, Lacerta viridis'\n",
      " 'n01694178 African chameleon, Chamaeleo chamaeleon'\n",
      " 'n01695060 Komodo dragon, Komodo lizard, dragon lizard, giant lizard, Varanus komodoensis'\n",
      " 'n01697457 African crocodile, Nile crocodile, Crocodylus niloticus'\n",
      " 'n01698640 American alligator, Alligator mississipiensis'\n",
      " 'n01704323 triceratops'\n",
      " 'n01728572 thunder snake, worm snake, Carphophis amoenus'\n",
      " 'n01728920 ringneck snake, ring-necked snake, ring snake'\n",
      " 'n01729322 hognose snake, puff adder, sand viper'\n",
      " 'n01729977 green snake, grass snake' 'n01734418 king snake, kingsnake'\n",
      " 'n01735189 garter snake, grass snake' 'n01737021 water snake'\n",
      " 'n01739381 vine snake' 'n01740131 night snake, Hypsiglena torquata'\n",
      " 'n01742172 boa constrictor, Constrictor constrictor'\n",
      " 'n01744401 rock python, rock snake, Python sebae'\n",
      " 'n01748264 Indian cobra, Naja naja' 'n01749939 green mamba'\n",
      " 'n01751748 sea snake'\n",
      " 'n01753488 horned viper, cerastes, sand viper, horned asp, Cerastes cornutus'\n",
      " 'n01755581 diamondback, diamondback rattlesnake, Crotalus adamanteus'\n",
      " 'n01756291 sidewinder, horned rattlesnake, Crotalus cerastes'\n",
      " 'n01768244 trilobite'\n",
      " 'n01770081 harvestman, daddy longlegs, Phalangium opilio'\n",
      " 'n01770393 scorpion'\n",
      " 'n01773157 black and gold garden spider, Argiope aurantia'\n",
      " 'n01773549 barn spider, Araneus cavaticus'\n",
      " 'n01773797 garden spider, Aranea diademata'\n",
      " 'n01774384 black widow, Latrodectus mactans' 'n01774750 tarantula'\n",
      " 'n01775062 wolf spider, hunting spider' 'n01776313 tick'\n",
      " 'n01784675 centipede' 'n01795545 black grouse' 'n01796340 ptarmigan'\n",
      " 'n01797886 ruffed grouse, partridge, Bonasa umbellus'\n",
      " 'n01798484 prairie chicken, prairie grouse, prairie fowl'\n",
      " 'n01806143 peacock' 'n01806567 quail' 'n01807496 partridge'\n",
      " 'n01817953 African grey, African gray, Psittacus erithacus'\n",
      " 'n01818515 macaw'\n",
      " 'n01819313 sulphur-crested cockatoo, Kakatoe galerita, Cacatua galerita'\n",
      " 'n01820546 lorikeet' 'n01824575 coucal' 'n01828970 bee eater'\n",
      " 'n01829413 hornbill' 'n01833805 hummingbird' 'n01843065 jacamar'\n",
      " 'n01843383 toucan' 'n01847000 drake'\n",
      " 'n01855032 red-breasted merganser, Mergus serrator' 'n01855672 goose'\n",
      " 'n01860187 black swan, Cygnus atratus' 'n01871265 tusker'\n",
      " 'n01872401 echidna, spiny anteater, anteater'\n",
      " 'n01873310 platypus, duckbill, duckbilled platypus, duck-billed platypus, Ornithorhynchus anatinus'\n",
      " 'n01877812 wallaby, brush kangaroo'\n",
      " 'n01882714 koala, koala bear, kangaroo bear, native bear, Phascolarctos cinereus'\n",
      " 'n01883070 wombat' 'n01910747 jellyfish' 'n01914609 sea anemone, anemone'\n",
      " 'n01917289 brain coral' 'n01924916 flatworm, platyhelminth'\n",
      " 'n01930112 nematode, nematode worm, roundworm' 'n01943899 conch'\n",
      " 'n01944390 snail' 'n01945685 slug' 'n01950731 sea slug, nudibranch'\n",
      " 'n01955084 chiton, coat-of-mail shell, sea cradle, polyplacophore'\n",
      " 'n01968897 chambered nautilus, pearly nautilus, nautilus'\n",
      " 'n01978287 Dungeness crab, Cancer magister'\n",
      " 'n01978455 rock crab, Cancer irroratus' 'n01980166 fiddler crab'\n",
      " 'n01981276 king crab, Alaska crab, Alaskan king crab, Alaska king crab, Paralithodes camtschatica'\n",
      " 'n01983481 American lobster, Northern lobster, Maine lobster, Homarus americanus'\n",
      " 'n01984695 spiny lobster, langouste, rock lobster, crawfish, crayfish, sea crawfish'\n",
      " 'n01985128 crayfish, crawfish, crawdad, crawdaddy' 'n01986214 hermit crab'\n",
      " 'n01990800 isopod' 'n02002556 white stork, Ciconia ciconia'\n",
      " 'n02002724 black stork, Ciconia nigra' 'n02006656 spoonbill'\n",
      " 'n02007558 flamingo' 'n02009229 little blue heron, Egretta caerulea'\n",
      " 'n02009912 American egret, great white heron, Egretta albus'\n",
      " 'n02011460 bittern' 'n02012849 crane' 'n02013706 limpkin, Aramus pictus'\n",
      " 'n02017213 European gallinule, Porphyrio porphyrio'\n",
      " 'n02018207 American coot, marsh hen, mud hen, water hen, Fulica americana'\n",
      " 'n02018795 bustard' 'n02025239 ruddy turnstone, Arenaria interpres'\n",
      " 'n02027492 red-backed sandpiper, dunlin, Erolia alpina'\n",
      " 'n02028035 redshank, Tringa totanus' 'n02033041 dowitcher'\n",
      " 'n02037110 oystercatcher, oyster catcher' 'n02051845 pelican'\n",
      " 'n02056570 king penguin, Aptenodytes patagonica'\n",
      " 'n02058221 albatross, mollymawk'\n",
      " 'n02066245 grey whale, gray whale, devilfish, Eschrichtius gibbosus, Eschrichtius robustus'\n",
      " 'n02071294 killer whale, killer, orca, grampus, sea wolf, Orcinus orca'\n",
      " 'n02074367 dugong, Dugong dugon' 'n02077923 sea lion'\n",
      " 'n02085620 Chihuahua' 'n02085782 Japanese spaniel'\n",
      " 'n02085936 Maltese dog, Maltese terrier, Maltese'\n",
      " 'n02086079 Pekinese, Pekingese, Peke' 'n02086240 Shih-Tzu'\n",
      " 'n02086646 Blenheim spaniel' 'n02086910 papillon' 'n02087046 toy terrier'\n",
      " 'n02087394 Rhodesian ridgeback' 'n02088094 Afghan hound, Afghan'\n",
      " 'n02088238 basset, basset hound' 'n02088364 beagle'\n",
      " 'n02088466 bloodhound, sleuthhound' 'n02088632 bluetick'\n",
      " 'n02089078 black-and-tan coonhound'\n",
      " 'n02089867 Walker hound, Walker foxhound' 'n02089973 English foxhound'\n",
      " 'n02090379 redbone' 'n02090622 borzoi, Russian wolfhound'\n",
      " 'n02090721 Irish wolfhound' 'n02091032 Italian greyhound'\n",
      " 'n02091134 whippet' 'n02091244 Ibizan hound, Ibizan Podenco'\n",
      " 'n02091467 Norwegian elkhound, elkhound'\n",
      " 'n02091635 otterhound, otter hound' 'n02091831 Saluki, gazelle hound'\n",
      " 'n02092002 Scottish deerhound, deerhound' 'n02092339 Weimaraner'\n",
      " 'n02093256 Staffordshire bullterrier, Staffordshire bull terrier'\n",
      " 'n02093428 American Staffordshire terrier, Staffordshire terrier, American pit bull terrier, pit bull terrier'\n",
      " 'n02093647 Bedlington terrier' 'n02093754 Border terrier'\n",
      " 'n02093859 Kerry blue terrier' 'n02093991 Irish terrier'\n",
      " 'n02094114 Norfolk terrier' 'n02094258 Norwich terrier'\n",
      " 'n02094433 Yorkshire terrier' 'n02095314 wire-haired fox terrier'\n",
      " 'n02095570 Lakeland terrier' 'n02095889 Sealyham terrier, Sealyham'\n",
      " 'n02096051 Airedale, Airedale terrier' 'n02096177 cairn, cairn terrier'\n",
      " 'n02096294 Australian terrier'\n",
      " 'n02096437 Dandie Dinmont, Dandie Dinmont terrier'\n",
      " 'n02096585 Boston bull, Boston terrier' 'n02097047 miniature schnauzer'\n",
      " 'n02097130 giant schnauzer' 'n02097209 standard schnauzer'\n",
      " 'n02097298 Scotch terrier, Scottish terrier, Scottie'\n",
      " 'n02097474 Tibetan terrier, chrysanthemum dog'\n",
      " 'n02097658 silky terrier, Sydney silky'\n",
      " 'n02098105 soft-coated wheaten terrier'\n",
      " 'n02098286 West Highland white terrier' 'n02098413 Lhasa, Lhasa apso'\n",
      " 'n02099267 flat-coated retriever' 'n02099429 curly-coated retriever'\n",
      " 'n02099601 golden retriever' 'n02099712 Labrador retriever'\n",
      " 'n02099849 Chesapeake Bay retriever'\n",
      " 'n02100236 German short-haired pointer'\n",
      " 'n02100583 vizsla, Hungarian pointer' 'n02100735 English setter'\n",
      " 'n02100877 Irish setter, red setter' 'n02101006 Gordon setter'\n",
      " 'n02101388 Brittany spaniel' 'n02101556 clumber, clumber spaniel'\n",
      " 'n02102040 English springer, English springer spaniel'\n",
      " 'n02102177 Welsh springer spaniel'\n",
      " 'n02102318 cocker spaniel, English cocker spaniel, cocker'\n",
      " 'n02102480 Sussex spaniel' 'n02102973 Irish water spaniel'\n",
      " 'n02104029 kuvasz' 'n02104365 schipperke' 'n02105056 groenendael'\n",
      " 'n02105162 malinois' 'n02105251 briard' 'n02105412 kelpie'\n",
      " 'n02105505 komondor' 'n02105641 Old English sheepdog, bobtail'\n",
      " 'n02105855 Shetland sheepdog, Shetland sheep dog, Shetland'\n",
      " 'n02106030 collie' 'n02106166 Border collie'\n",
      " 'n02106382 Bouvier des Flandres, Bouviers des Flandres'\n",
      " 'n02106550 Rottweiler'\n",
      " 'n02106662 German shepherd, German shepherd dog, German police dog, alsatian'\n",
      " 'n02107142 Doberman, Doberman pinscher' 'n02107312 miniature pinscher'\n",
      " 'n02107574 Greater Swiss Mountain dog' 'n02107683 Bernese mountain dog'\n",
      " 'n02107908 Appenzeller' 'n02108000 EntleBucher' 'n02108089 boxer'\n",
      " 'n02108422 bull mastiff' 'n02108551 Tibetan mastiff'\n",
      " 'n02108915 French bulldog' 'n02109047 Great Dane'\n",
      " 'n02109525 Saint Bernard, St Bernard' 'n02109961 Eskimo dog, husky'\n",
      " 'n02110063 malamute, malemute, Alaskan malamute'\n",
      " 'n02110185 Siberian husky' 'n02110341 dalmatian, coach dog, carriage dog'\n",
      " 'n02110627 affenpinscher, monkey pinscher, monkey dog' 'n02110806 basenji'\n",
      " 'n02110958 pug, pug-dog' 'n02111129 Leonberg'\n",
      " 'n02111277 Newfoundland, Newfoundland dog' 'n02111500 Great Pyrenees'\n",
      " 'n02111889 Samoyed, Samoyede' 'n02112018 Pomeranian'\n",
      " 'n02112137 chow, chow chow' 'n02112350 keeshond'\n",
      " 'n02112706 Brabancon griffon' 'n02113023 Pembroke, Pembroke Welsh corgi'\n",
      " 'n02113186 Cardigan, Cardigan Welsh corgi' 'n02113624 toy poodle'\n",
      " 'n02113712 miniature poodle' 'n02113799 standard poodle'\n",
      " 'n02113978 Mexican hairless'\n",
      " 'n02114367 timber wolf, grey wolf, gray wolf, Canis lupus'\n",
      " 'n02114548 white wolf, Arctic wolf, Canis lupus tundrarum'\n",
      " 'n02114712 red wolf, maned wolf, Canis rufus, Canis niger'\n",
      " 'n02114855 coyote, prairie wolf, brush wolf, Canis latrans'\n",
      " 'n02115641 dingo, warrigal, warragal, Canis dingo'\n",
      " 'n02115913 dhole, Cuon alpinus'\n",
      " 'n02116738 African hunting dog, hyena dog, Cape hunting dog, Lycaon pictus'\n",
      " 'n02117135 hyena, hyaena' 'n02119022 red fox, Vulpes vulpes'\n",
      " 'n02119789 kit fox, Vulpes macrotis'\n",
      " 'n02120079 Arctic fox, white fox, Alopex lagopus'\n",
      " 'n02120505 grey fox, gray fox, Urocyon cinereoargenteus'\n",
      " 'n02123045 tabby, tabby cat' 'n02123159 tiger cat' 'n02123394 Persian cat'\n",
      " 'n02123597 Siamese cat, Siamese' 'n02124075 Egyptian cat'\n",
      " 'n02125311 cougar, puma, catamount, mountain lion, painter, panther, Felis concolor'\n",
      " 'n02127052 lynx, catamount' 'n02128385 leopard, Panthera pardus'\n",
      " 'n02128757 snow leopard, ounce, Panthera uncia'\n",
      " 'n02128925 jaguar, panther, Panthera onca, Felis onca'\n",
      " 'n02129165 lion, king of beasts, Panthera leo'\n",
      " 'n02129604 tiger, Panthera tigris'\n",
      " 'n02130308 cheetah, chetah, Acinonyx jubatus'\n",
      " 'n02132136 brown bear, bruin, Ursus arctos'\n",
      " 'n02133161 American black bear, black bear, Ursus americanus, Euarctos americanus'\n",
      " 'n02134084 ice bear, polar bear, Ursus Maritimus, Thalarctos maritimus'\n",
      " 'n02134418 sloth bear, Melursus ursinus, Ursus ursinus'\n",
      " 'n02137549 mongoose' 'n02138441 meerkat, mierkat' 'n02165105 tiger beetle'\n",
      " 'n02165456 ladybug, ladybeetle, lady beetle, ladybird, ladybird beetle'\n",
      " 'n02167151 ground beetle, carabid beetle'\n",
      " 'n02168699 long-horned beetle, longicorn, longicorn beetle'\n",
      " 'n02169497 leaf beetle, chrysomelid' 'n02172182 dung beetle'\n",
      " 'n02174001 rhinoceros beetle' 'n02177972 weevil' 'n02190166 fly'\n",
      " 'n02206856 bee' 'n02219486 ant, emmet, pismire'\n",
      " 'n02226429 grasshopper, hopper' 'n02229544 cricket'\n",
      " 'n02231487 walking stick, walkingstick, stick insect'\n",
      " 'n02233338 cockroach, roach' 'n02236044 mantis, mantid'\n",
      " 'n02256656 cicada, cicala' 'n02259212 leafhopper'\n",
      " 'n02264363 lacewing, lacewing fly'\n",
      " \"n02268443 dragonfly, darning needle, devil's darning needle, sewing needle, snake feeder, snake doctor, mosquito hawk, skeeter hawk\"\n",
      " 'n02268853 damselfly' 'n02276258 admiral'\n",
      " 'n02277742 ringlet, ringlet butterfly'\n",
      " 'n02279972 monarch, monarch butterfly, milkweed butterfly, Danaus plexippus'\n",
      " 'n02280649 cabbage butterfly'\n",
      " 'n02281406 sulphur butterfly, sulfur butterfly'\n",
      " 'n02281787 lycaenid, lycaenid butterfly' 'n02317335 starfish, sea star'\n",
      " 'n02319095 sea urchin' 'n02321529 sea cucumber, holothurian'\n",
      " 'n02325366 wood rabbit, cottontail, cottontail rabbit' 'n02326432 hare'\n",
      " 'n02328150 Angora, Angora rabbit' 'n02342885 hamster'\n",
      " 'n02346627 porcupine, hedgehog'\n",
      " 'n02356798 fox squirrel, eastern fox squirrel, Sciurus niger'\n",
      " 'n02361337 marmot' 'n02363005 beaver' 'n02364673 guinea pig, Cavia cobaya'\n",
      " 'n02389026 sorrel' 'n02391049 zebra'\n",
      " 'n02395406 hog, pig, grunter, squealer, Sus scrofa'\n",
      " 'n02396427 wild boar, boar, Sus scrofa' 'n02397096 warthog'\n",
      " 'n02398521 hippopotamus, hippo, river horse, Hippopotamus amphibius'\n",
      " 'n02403003 ox'\n",
      " 'n02408429 water buffalo, water ox, Asiatic buffalo, Bubalus bubalis'\n",
      " 'n02410509 bison' 'n02412080 ram, tup'\n",
      " 'n02415577 bighorn, bighorn sheep, cimarron, Rocky Mountain bighorn, Rocky Mountain sheep, Ovis canadensis'\n",
      " 'n02417914 ibex, Capra ibex' 'n02422106 hartebeest'\n",
      " 'n02422699 impala, Aepyceros melampus' 'n02423022 gazelle'\n",
      " 'n02437312 Arabian camel, dromedary, Camelus dromedarius'\n",
      " 'n02437616 llama' 'n02441942 weasel' 'n02442845 mink'\n",
      " 'n02443114 polecat, fitch, foulmart, foumart, Mustela putorius'\n",
      " 'n02443484 black-footed ferret, ferret, Mustela nigripes'\n",
      " 'n02444819 otter' 'n02445715 skunk, polecat, wood pussy'\n",
      " 'n02447366 badger' 'n02454379 armadillo'\n",
      " 'n02457408 three-toed sloth, ai, Bradypus tridactylus'\n",
      " 'n02480495 orangutan, orang, orangutang, Pongo pygmaeus'\n",
      " 'n02480855 gorilla, Gorilla gorilla'\n",
      " 'n02481823 chimpanzee, chimp, Pan troglodytes'\n",
      " 'n02483362 gibbon, Hylobates lar'\n",
      " 'n02483708 siamang, Hylobates syndactylus, Symphalangus syndactylus'\n",
      " 'n02484975 guenon, guenon monkey'\n",
      " 'n02486261 patas, hussar monkey, Erythrocebus patas' 'n02486410 baboon'\n",
      " 'n02487347 macaque' 'n02488291 langur' 'n02488702 colobus, colobus monkey'\n",
      " 'n02489166 proboscis monkey, Nasalis larvatus' 'n02490219 marmoset'\n",
      " 'n02492035 capuchin, ringtail, Cebus capucinus'\n",
      " 'n02492660 howler monkey, howler' 'n02493509 titi, titi monkey'\n",
      " 'n02493793 spider monkey, Ateles geoffroyi'\n",
      " 'n02494079 squirrel monkey, Saimiri sciureus'\n",
      " 'n02497673 Madagascar cat, ring-tailed lemur, Lemur catta'\n",
      " 'n02500267 indri, indris, Indri indri, Indri brevicaudatus'\n",
      " 'n02504013 Indian elephant, Elephas maximus'\n",
      " 'n02504458 African elephant, Loxodonta africana'\n",
      " 'n02509815 lesser panda, red panda, panda, bear cat, cat bear, Ailurus fulgens'\n",
      " 'n02510455 giant panda, panda, panda bear, coon bear, Ailuropoda melanoleuca'\n",
      " 'n02514041 barracouta, snoek' 'n02526121 eel'\n",
      " 'n02536864 coho, cohoe, coho salmon, blue jack, silver salmon, Oncorhynchus kisutch'\n",
      " 'n02606052 rock beauty, Holocanthus tricolor' 'n02607072 anemone fish'\n",
      " 'n02640242 sturgeon'\n",
      " 'n02641379 gar, garfish, garpike, billfish, Lepisosteus osseus'\n",
      " 'n02643566 lionfish' 'n02655020 puffer, pufferfish, blowfish, globefish'\n",
      " 'n02666196 abacus' 'n02667093 abaya'\n",
      " \"n02669723 academic gown, academic robe, judge's robe\"\n",
      " 'n02672831 accordion, piano accordion, squeeze box'\n",
      " 'n02676566 acoustic guitar'\n",
      " 'n02687172 aircraft carrier, carrier, flattop, attack aircraft carrier'\n",
      " 'n02690373 airliner' 'n02692877 airship, dirigible' 'n02699494 altar'\n",
      " 'n02701002 ambulance' 'n02704792 amphibian, amphibious vehicle'\n",
      " 'n02708093 analog clock' 'n02727426 apiary, bee house' 'n02730930 apron'\n",
      " 'n02747177 ashcan, trash can, garbage can, wastebin, ash bin, ash-bin, ashbin, dustbin, trash barrel, trash bin'\n",
      " 'n02749479 assault rifle, assault gun'\n",
      " 'n02769748 backpack, back pack, knapsack, packsack, rucksack, haversack'\n",
      " 'n02776631 bakery, bakeshop, bakehouse' 'n02777292 balance beam, beam'\n",
      " 'n02782093 balloon' 'n02783161 ballpoint, ballpoint pen, ballpen, Biro'\n",
      " 'n02786058 Band Aid' 'n02787622 banjo'\n",
      " 'n02788148 bannister, banister, balustrade, balusters, handrail'\n",
      " 'n02790996 barbell' 'n02791124 barber chair' 'n02791270 barbershop'\n",
      " 'n02793495 barn' 'n02794156 barometer' 'n02795169 barrel, cask'\n",
      " 'n02797295 barrow, garden cart, lawn cart, wheelbarrow'\n",
      " 'n02799071 baseball' 'n02802426 basketball' 'n02804414 bassinet'\n",
      " 'n02804610 bassoon' 'n02807133 bathing cap, swimming cap'\n",
      " 'n02808304 bath towel' 'n02808440 bathtub, bathing tub, bath, tub'\n",
      " 'n02814533 beach wagon, station wagon, wagon, estate car, beach waggon, station waggon, waggon'\n",
      " 'n02814860 beacon, lighthouse, beacon light, pharos' 'n02815834 beaker'\n",
      " 'n02817516 bearskin, busby, shako' 'n02823428 beer bottle'\n",
      " 'n02823750 beer glass' 'n02825657 bell cote, bell cot' 'n02834397 bib'\n",
      " 'n02835271 bicycle-built-for-two, tandem bicycle, tandem'\n",
      " 'n02837789 bikini, two-piece' 'n02840245 binder, ring-binder'\n",
      " 'n02841315 binoculars, field glasses, opera glasses' 'n02843684 birdhouse'\n",
      " 'n02859443 boathouse' 'n02860847 bobsled, bobsleigh, bob'\n",
      " 'n02865351 bolo tie, bolo, bola tie, bola' 'n02869837 bonnet, poke bonnet'\n",
      " 'n02870880 bookcase' 'n02871525 bookshop, bookstore, bookstall'\n",
      " 'n02877765 bottlecap' 'n02879718 bow' 'n02883205 bow tie, bow-tie, bowtie'\n",
      " 'n02892201 brass, memorial tablet, plaque'\n",
      " 'n02892767 brassiere, bra, bandeau'\n",
      " 'n02894605 breakwater, groin, groyne, mole, bulwark, seawall, jetty'\n",
      " 'n02895154 breastplate, aegis, egis' 'n02906734 broom'\n",
      " 'n02909870 bucket, pail' 'n02910353 buckle' 'n02916936 bulletproof vest'\n",
      " 'n02917067 bullet train, bullet' 'n02927161 butcher shop, meat market'\n",
      " 'n02930766 cab, hack, taxi, taxicab' 'n02939185 caldron, cauldron'\n",
      " 'n02948072 candle, taper, wax light' 'n02950826 cannon' 'n02951358 canoe'\n",
      " 'n02951585 can opener, tin opener' 'n02963159 cardigan'\n",
      " 'n02965783 car mirror'\n",
      " 'n02966193 carousel, carrousel, merry-go-round, roundabout, whirligig'\n",
      " \"n02966687 carpenter's kit, tool kit\" 'n02971356 carton'\n",
      " 'n02974003 car wheel'\n",
      " 'n02977058 cash machine, cash dispenser, automated teller machine, automatic teller machine, automated teller, automatic teller, ATM'\n",
      " 'n02978881 cassette' 'n02979186 cassette player' 'n02980441 castle'\n",
      " 'n02981792 catamaran' 'n02988304 CD player' 'n02992211 cello, violoncello'\n",
      " 'n02992529 cellular telephone, cellular phone, cellphone, cell, mobile phone'\n",
      " 'n02999410 chain' 'n03000134 chainlink fence'\n",
      " 'n03000247 chain mail, ring mail, mail, chain armor, chain armour, ring armor, ring armour'\n",
      " 'n03000684 chain saw, chainsaw' 'n03014705 chest'\n",
      " 'n03016953 chiffonier, commode' 'n03017168 chime, bell, gong'\n",
      " 'n03018349 china cabinet, china closet' 'n03026506 Christmas stocking'\n",
      " 'n03028079 church, church building'\n",
      " 'n03032252 cinema, movie theater, movie theatre, movie house, picture palace'\n",
      " 'n03041632 cleaver, meat cleaver, chopper' 'n03042490 cliff dwelling'\n",
      " 'n03045698 cloak' 'n03047690 clog, geta, patten, sabot'\n",
      " 'n03062245 cocktail shaker' 'n03063599 coffee mug' 'n03063689 coffeepot'\n",
      " 'n03065424 coil, spiral, volute, whorl, helix'\n",
      " 'n03075370 combination lock' 'n03085013 computer keyboard, keypad'\n",
      " 'n03089624 confectionery, confectionary, candy store'\n",
      " 'n03095699 container ship, containership, container vessel'\n",
      " 'n03100240 convertible' 'n03109150 corkscrew, bottle screw'\n",
      " 'n03110669 cornet, horn, trumpet, trump' 'n03124043 cowboy boot'\n",
      " 'n03124170 cowboy hat, ten-gallon hat' 'n03125729 cradle'\n",
      " 'n03126707 crane' 'n03127747 crash helmet' 'n03127925 crate'\n",
      " 'n03131574 crib, cot' 'n03133878 Crock Pot' 'n03134739 croquet ball'\n",
      " 'n03141823 crutch' 'n03146219 cuirass' 'n03160309 dam, dike, dyke'\n",
      " 'n03179701 desk' 'n03180011 desktop computer'\n",
      " 'n03187595 dial telephone, dial phone' 'n03188531 diaper, nappy, napkin'\n",
      " 'n03196217 digital clock' 'n03197337 digital watch'\n",
      " 'n03201208 dining table, board' 'n03207743 dishrag, dishcloth'\n",
      " 'n03207941 dishwasher, dish washer, dishwashing machine'\n",
      " 'n03208938 disk brake, disc brake'\n",
      " 'n03216828 dock, dockage, docking facility'\n",
      " 'n03218198 dogsled, dog sled, dog sleigh' 'n03220513 dome'\n",
      " 'n03223299 doormat, welcome mat'\n",
      " 'n03240683 drilling platform, offshore rig'\n",
      " 'n03249569 drum, membranophone, tympan' 'n03250847 drumstick'\n",
      " 'n03255030 dumbbell' 'n03259280 Dutch oven'\n",
      " 'n03271574 electric fan, blower' 'n03272010 electric guitar'\n",
      " 'n03272562 electric locomotive' 'n03290653 entertainment center'\n",
      " 'n03291819 envelope' 'n03297495 espresso maker' 'n03314780 face powder'\n",
      " 'n03325584 feather boa, boa'\n",
      " 'n03337140 file, file cabinet, filing cabinet' 'n03344393 fireboat'\n",
      " 'n03345487 fire engine, fire truck' 'n03347037 fire screen, fireguard'\n",
      " 'n03355925 flagpole, flagstaff' 'n03372029 flute, transverse flute'\n",
      " 'n03376595 folding chair' 'n03379051 football helmet' 'n03384352 forklift'\n",
      " 'n03388043 fountain' 'n03388183 fountain pen' 'n03388549 four-poster'\n",
      " 'n03393912 freight car' 'n03394916 French horn, horn'\n",
      " 'n03400231 frying pan, frypan, skillet' 'n03404251 fur coat'\n",
      " 'n03417042 garbage truck, dustcart'\n",
      " 'n03424325 gasmask, respirator, gas helmet'\n",
      " 'n03425413 gas pump, gasoline pump, petrol pump, island dispenser'\n",
      " 'n03443371 goblet' 'n03444034 go-kart' 'n03445777 golf ball'\n",
      " 'n03445924 golfcart, golf cart' 'n03447447 gondola'\n",
      " 'n03447721 gong, tam-tam' 'n03450230 gown' 'n03452741 grand piano, grand'\n",
      " 'n03457902 greenhouse, nursery, glasshouse'\n",
      " 'n03459775 grille, radiator grille'\n",
      " 'n03461385 grocery store, grocery, food market, market'\n",
      " 'n03467068 guillotine' 'n03476684 hair slide' 'n03476991 hair spray'\n",
      " 'n03478589 half track' 'n03481172 hammer' 'n03482405 hamper'\n",
      " 'n03483316 hand blower, blow dryer, blow drier, hair dryer, hair drier'\n",
      " 'n03485407 hand-held computer, hand-held microcomputer'\n",
      " 'n03485794 handkerchief, hankie, hanky, hankey'\n",
      " 'n03492542 hard disc, hard disk, fixed disk'\n",
      " 'n03494278 harmonica, mouth organ, harp, mouth harp' 'n03495258 harp'\n",
      " 'n03496892 harvester, reaper' 'n03498962 hatchet' 'n03527444 holster'\n",
      " 'n03529860 home theater, home theatre' 'n03530642 honeycomb'\n",
      " 'n03532672 hook, claw' 'n03534580 hoopskirt, crinoline'\n",
      " 'n03535780 horizontal bar, high bar' 'n03538406 horse cart, horse-cart'\n",
      " 'n03544143 hourglass' 'n03584254 iPod' 'n03584829 iron, smoothing iron'\n",
      " \"n03590841 jack-o'-lantern\" 'n03594734 jean, blue jean, denim'\n",
      " 'n03594945 jeep, landrover' 'n03595614 jersey, T-shirt, tee shirt'\n",
      " 'n03598930 jigsaw puzzle' 'n03599486 jinrikisha, ricksha, rickshaw'\n",
      " 'n03602883 joystick' 'n03617480 kimono' 'n03623198 knee pad'\n",
      " 'n03627232 knot' 'n03630383 lab coat, laboratory coat' 'n03633091 ladle'\n",
      " 'n03637318 lampshade, lamp shade' 'n03642806 laptop, laptop computer'\n",
      " 'n03649909 lawn mower, mower' 'n03657121 lens cap, lens cover'\n",
      " 'n03658185 letter opener, paper knife, paperknife' 'n03661043 library'\n",
      " 'n03662601 lifeboat' 'n03666591 lighter, light, igniter, ignitor'\n",
      " 'n03670208 limousine, limo' 'n03673027 liner, ocean liner'\n",
      " 'n03676483 lipstick, lip rouge' 'n03680355 Loafer' 'n03690938 lotion'\n",
      " 'n03691459 loudspeaker, speaker, speaker unit, loudspeaker system, speaker system'\n",
      " \"n03692522 loupe, jeweler's loupe\" 'n03697007 lumbermill, sawmill'\n",
      " 'n03706229 magnetic compass' 'n03709823 mailbag, postbag'\n",
      " 'n03710193 mailbox, letter box' 'n03710637 maillot'\n",
      " 'n03710721 maillot, tank suit' 'n03717622 manhole cover'\n",
      " 'n03720891 maraca' 'n03721384 marimba, xylophone' 'n03724870 mask'\n",
      " 'n03729826 matchstick' 'n03733131 maypole' 'n03733281 maze, labyrinth'\n",
      " 'n03733805 measuring cup' 'n03742115 medicine chest, medicine cabinet'\n",
      " 'n03743016 megalith, megalithic structure' 'n03759954 microphone, mike'\n",
      " 'n03761084 microwave, microwave oven' 'n03763968 military uniform'\n",
      " 'n03764736 milk can' 'n03769881 minibus' 'n03770439 miniskirt, mini'\n",
      " 'n03770679 minivan' 'n03773504 missile' 'n03775071 mitten'\n",
      " 'n03775546 mixing bowl' 'n03776460 mobile home, manufactured home'\n",
      " 'n03777568 Model T' 'n03777754 modem' 'n03781244 monastery'\n",
      " 'n03782006 monitor' 'n03785016 moped' 'n03786901 mortar'\n",
      " 'n03787032 mortarboard' 'n03788195 mosque' 'n03788365 mosquito net'\n",
      " 'n03791053 motor scooter, scooter'\n",
      " 'n03792782 mountain bike, all-terrain bike, off-roader'\n",
      " 'n03792972 mountain tent' 'n03793489 mouse, computer mouse'\n",
      " 'n03794056 mousetrap' 'n03796401 moving van' 'n03803284 muzzle'\n",
      " 'n03804744 nail' 'n03814639 neck brace' 'n03814906 necklace'\n",
      " 'n03825788 nipple' 'n03832673 notebook, notebook computer'\n",
      " 'n03837869 obelisk' 'n03838899 oboe, hautboy, hautbois'\n",
      " 'n03840681 ocarina, sweet potato'\n",
      " 'n03841143 odometer, hodometer, mileometer, milometer'\n",
      " 'n03843555 oil filter' 'n03854065 organ, pipe organ'\n",
      " 'n03857828 oscilloscope, scope, cathode-ray oscilloscope, CRO'\n",
      " 'n03866082 overskirt' 'n03868242 oxcart' 'n03868863 oxygen mask'\n",
      " 'n03871628 packet' 'n03873416 paddle, boat paddle'\n",
      " 'n03874293 paddlewheel, paddle wheel' 'n03874599 padlock'\n",
      " 'n03876231 paintbrush' \"n03877472 pajama, pyjama, pj's, jammies\"\n",
      " 'n03877845 palace' 'n03884397 panpipe, pandean pipe, syrinx'\n",
      " 'n03887697 paper towel' 'n03888257 parachute, chute'\n",
      " 'n03888605 parallel bars, bars' 'n03891251 park bench'\n",
      " 'n03891332 parking meter' 'n03895866 passenger car, coach, carriage'\n",
      " 'n03899768 patio, terrace' 'n03902125 pay-phone, pay-station'\n",
      " 'n03903868 pedestal, plinth, footstall'\n",
      " 'n03908618 pencil box, pencil case' 'n03908714 pencil sharpener'\n",
      " 'n03916031 perfume, essence' 'n03920288 Petri dish'\n",
      " 'n03924679 photocopier' 'n03929660 pick, plectrum, plectron'\n",
      " 'n03929855 pickelhaube' 'n03930313 picket fence, paling'\n",
      " 'n03930630 pickup, pickup truck' 'n03933933 pier'\n",
      " 'n03935335 piggy bank, penny bank' 'n03937543 pill bottle'\n",
      " 'n03938244 pillow' 'n03942813 ping-pong ball' 'n03944341 pinwheel'\n",
      " 'n03947888 pirate, pirate ship' 'n03950228 pitcher, ewer'\n",
      " \"n03954731 plane, carpenter's plane, woodworking plane\"\n",
      " 'n03956157 planetarium' 'n03958227 plastic bag' 'n03961711 plate rack'\n",
      " 'n03967562 plow, plough' \"n03970156 plunger, plumber's helper\"\n",
      " 'n03976467 Polaroid camera, Polaroid Land camera' 'n03976657 pole'\n",
      " 'n03977966 police van, police wagon, paddy wagon, patrol wagon, wagon, black Maria'\n",
      " 'n03980874 poncho' 'n03982430 pool table, billiard table, snooker table'\n",
      " 'n03983396 pop bottle, soda bottle' 'n03991062 pot, flowerpot'\n",
      " \"n03992509 potter's wheel\" 'n03995372 power drill'\n",
      " 'n03998194 prayer rug, prayer mat' 'n04004767 printer'\n",
      " 'n04005630 prison, prison house' 'n04008634 projectile, missile'\n",
      " 'n04009552 projector' 'n04019541 puck, hockey puck'\n",
      " 'n04023962 punching bag, punch bag, punching ball, punchball'\n",
      " 'n04026417 purse' 'n04033901 quill, quill pen'\n",
      " 'n04033995 quilt, comforter, comfort, puff'\n",
      " 'n04037443 racer, race car, racing car' 'n04039381 racket, racquet'\n",
      " 'n04040759 radiator' 'n04041544 radio, wireless'\n",
      " 'n04044716 radio telescope, radio reflector' 'n04049303 rain barrel'\n",
      " 'n04065272 recreational vehicle, RV, R.V.' 'n04067472 reel'\n",
      " 'n04069434 reflex camera' 'n04070727 refrigerator, icebox'\n",
      " 'n04074963 remote control, remote'\n",
      " 'n04081281 restaurant, eating house, eating place, eatery'\n",
      " 'n04086273 revolver, six-gun, six-shooter' 'n04090263 rifle'\n",
      " 'n04099969 rocking chair, rocker' 'n04111531 rotisserie'\n",
      " 'n04116512 rubber eraser, rubber, pencil eraser' 'n04118538 rugby ball'\n",
      " 'n04118776 rule, ruler' 'n04120489 running shoe' 'n04125021 safe'\n",
      " 'n04127249 safety pin' 'n04131690 saltshaker, salt shaker'\n",
      " 'n04133789 sandal' 'n04136333 sarong' 'n04141076 sax, saxophone'\n",
      " 'n04141327 scabbard' 'n04141975 scale, weighing machine'\n",
      " 'n04146614 school bus' 'n04147183 schooner' 'n04149813 scoreboard'\n",
      " 'n04152593 screen, CRT screen' 'n04153751 screw' 'n04154565 screwdriver'\n",
      " 'n04162706 seat belt, seatbelt' 'n04179913 sewing machine'\n",
      " 'n04192698 shield, buckler' 'n04200800 shoe shop, shoe-shop, shoe store'\n",
      " 'n04201297 shoji' 'n04204238 shopping basket' 'n04204347 shopping cart'\n",
      " 'n04208210 shovel' 'n04209133 shower cap' 'n04209239 shower curtain'\n",
      " 'n04228054 ski' 'n04229816 ski mask' 'n04235860 sleeping bag'\n",
      " 'n04238763 slide rule, slipstick' 'n04239074 sliding door'\n",
      " 'n04243546 slot, one-armed bandit' 'n04251144 snorkel'\n",
      " 'n04252077 snowmobile' 'n04252225 snowplow, snowplough'\n",
      " 'n04254120 soap dispenser' 'n04254680 soccer ball' 'n04254777 sock'\n",
      " 'n04258138 solar dish, solar collector, solar furnace'\n",
      " 'n04259630 sombrero' 'n04263257 soup bowl' 'n04264628 space bar'\n",
      " 'n04265275 space heater' 'n04266014 space shuttle' 'n04270147 spatula'\n",
      " 'n04273569 speedboat' \"n04275548 spider web, spider's web\"\n",
      " 'n04277352 spindle' 'n04285008 sports car, sport car'\n",
      " 'n04286575 spotlight, spot' 'n04296562 stage' 'n04310018 steam locomotive'\n",
      " 'n04311004 steel arch bridge' 'n04311174 steel drum'\n",
      " 'n04317175 stethoscope' 'n04325704 stole' 'n04326547 stone wall'\n",
      " 'n04328186 stopwatch, stop watch' 'n04330267 stove' 'n04332243 strainer'\n",
      " 'n04335435 streetcar, tram, tramcar, trolley, trolley car'\n",
      " 'n04336792 stretcher' 'n04344873 studio couch, day bed'\n",
      " 'n04346328 stupa, tope' 'n04347754 submarine, pigboat, sub, U-boat'\n",
      " 'n04350905 suit, suit of clothes' 'n04355338 sundial' 'n04355933 sunglass'\n",
      " 'n04356056 sunglasses, dark glasses, shades'\n",
      " 'n04357314 sunscreen, sunblock, sun blocker' 'n04366367 suspension bridge'\n",
      " 'n04367480 swab, swob, mop' 'n04370456 sweatshirt'\n",
      " 'n04371430 swimming trunks, bathing trunks' 'n04371774 swing'\n",
      " 'n04372370 switch, electric switch, electrical switch' 'n04376876 syringe'\n",
      " 'n04380533 table lamp'\n",
      " 'n04389033 tank, army tank, armored combat vehicle, armoured combat vehicle'\n",
      " 'n04392985 tape player' 'n04398044 teapot' 'n04399382 teddy, teddy bear'\n",
      " 'n04404412 television, television system' 'n04409515 tennis ball'\n",
      " 'n04417672 thatch, thatched roof'\n",
      " 'n04418357 theater curtain, theatre curtain' 'n04423845 thimble'\n",
      " 'n04428191 thresher, thrasher, threshing machine' 'n04429376 throne'\n",
      " 'n04435653 tile roof' 'n04442312 toaster'\n",
      " 'n04443257 tobacco shop, tobacconist shop, tobacconist'\n",
      " 'n04447861 toilet seat' 'n04456115 torch' 'n04458633 totem pole'\n",
      " 'n04461696 tow truck, tow car, wrecker' 'n04462240 toyshop'\n",
      " 'n04465501 tractor'\n",
      " 'n04467665 trailer truck, tractor trailer, trucking rig, rig, articulated lorry, semi'\n",
      " 'n04476259 tray' 'n04479046 trench coat'\n",
      " 'n04482393 tricycle, trike, velocipede' 'n04483307 trimaran'\n",
      " 'n04485082 tripod' 'n04486054 triumphal arch'\n",
      " 'n04487081 trolleybus, trolley coach, trackless trolley'\n",
      " 'n04487394 trombone' 'n04493381 tub, vat' 'n04501370 turnstile'\n",
      " 'n04505470 typewriter keyboard' 'n04507155 umbrella'\n",
      " 'n04509417 unicycle, monocycle' 'n04515003 upright, upright piano'\n",
      " 'n04517823 vacuum, vacuum cleaner' 'n04522168 vase' 'n04523525 vault'\n",
      " 'n04525038 velvet' 'n04525305 vending machine' 'n04532106 vestment'\n",
      " 'n04532670 viaduct' 'n04536866 violin, fiddle' 'n04540053 volleyball'\n",
      " 'n04542943 waffle iron' 'n04548280 wall clock'\n",
      " 'n04548362 wallet, billfold, notecase, pocketbook'\n",
      " 'n04550184 wardrobe, closet, press' 'n04552348 warplane, military plane'\n",
      " 'n04553703 washbasin, handbasin, washbowl, lavabo, wash-hand basin'\n",
      " 'n04554684 washer, automatic washer, washing machine'\n",
      " 'n04557648 water bottle' 'n04560804 water jug' 'n04562935 water tower'\n",
      " 'n04579145 whiskey jug' 'n04579432 whistle' 'n04584207 wig'\n",
      " 'n04589890 window screen' 'n04590129 window shade' 'n04591157 Windsor tie'\n",
      " 'n04591713 wine bottle' 'n04592741 wing' 'n04596742 wok'\n",
      " 'n04597913 wooden spoon' 'n04599235 wool, woolen, woollen'\n",
      " 'n04604644 worm fence, snake fence, snake-rail fence, Virginia fence'\n",
      " 'n04606251 wreck' 'n04612504 yawl' 'n04613696 yurt'\n",
      " 'n06359193 web site, website, internet site, site' 'n06596364 comic book'\n",
      " 'n06785654 crossword puzzle, crossword' 'n06794110 street sign'\n",
      " 'n06874185 traffic light, traffic signal, stoplight'\n",
      " 'n07248320 book jacket, dust cover, dust jacket, dust wrapper'\n",
      " 'n07565083 menu' 'n07579787 plate' 'n07583066 guacamole'\n",
      " 'n07584110 consomme' 'n07590611 hot pot, hotpot' 'n07613480 trifle'\n",
      " 'n07614500 ice cream, icecream'\n",
      " 'n07615774 ice lolly, lolly, lollipop, popsicle' 'n07684084 French loaf'\n",
      " 'n07693725 bagel, beigel' 'n07695742 pretzel' 'n07697313 cheeseburger'\n",
      " 'n07697537 hotdog, hot dog, red hot' 'n07711569 mashed potato'\n",
      " 'n07714571 head cabbage' 'n07714990 broccoli' 'n07715103 cauliflower'\n",
      " 'n07716358 zucchini, courgette' 'n07716906 spaghetti squash'\n",
      " 'n07717410 acorn squash' 'n07717556 butternut squash'\n",
      " 'n07718472 cucumber, cuke' 'n07718747 artichoke, globe artichoke'\n",
      " 'n07720875 bell pepper' 'n07730033 cardoon' 'n07734744 mushroom'\n",
      " 'n07742313 Granny Smith' 'n07745940 strawberry' 'n07747607 orange'\n",
      " 'n07749582 lemon' 'n07753113 fig' 'n07753275 pineapple, ananas'\n",
      " 'n07753592 banana' 'n07754684 jackfruit, jak, jack'\n",
      " 'n07760859 custard apple' 'n07768694 pomegranate' 'n07802026 hay'\n",
      " 'n07831146 carbonara' 'n07836838 chocolate sauce, chocolate syrup'\n",
      " 'n07860988 dough' 'n07871810 meat loaf, meatloaf'\n",
      " 'n07873807 pizza, pizza pie' 'n07875152 potpie' 'n07880968 burrito'\n",
      " 'n07892512 red wine' 'n07920052 espresso' 'n07930864 cup'\n",
      " 'n07932039 eggnog' 'n09193705 alp' 'n09229709 bubble'\n",
      " 'n09246464 cliff, drop, drop-off' 'n09256479 coral reef'\n",
      " 'n09288635 geyser' 'n09332890 lakeside, lakeshore'\n",
      " 'n09399592 promontory, headland, head, foreland'\n",
      " 'n09421951 sandbar, sand bar'\n",
      " 'n09428293 seashore, coast, seacoast, sea-coast' 'n09468604 valley, vale'\n",
      " 'n09472597 volcano' 'n09835506 ballplayer, baseball player'\n",
      " 'n10148035 groom, bridegroom' 'n10565667 scuba diver' 'n11879895 rapeseed'\n",
      " 'n11939491 daisy'\n",
      " \"n12057211 yellow lady's slipper, yellow lady-slipper, Cypripedium calceolus, Cypripedium parviflorum\"\n",
      " 'n12144580 corn' 'n12267677 acorn' 'n12620546 hip, rose hip, rosehip'\n",
      " 'n12768682 buckeye, horse chestnut, conker' 'n12985857 coral fungus'\n",
      " 'n12998815 agaric' 'n13037406 gyromitra'\n",
      " 'n13040303 stinkhorn, carrion fungus' 'n13044778 earthstar'\n",
      " 'n13052670 hen-of-the-woods, hen of the woods, Polyporus frondosus, Grifola frondosa'\n",
      " 'n13054560 bolete' 'n13133613 ear, spike, capitulum'\n",
      " 'n15075141 toilet tissue, toilet paper, bathroom tissue']\n"
     ]
    }
   ],
   "source": [
    "print labels[0:1000]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "description": "Instant recognition with a pre-trained model and a tour of the net interface for visualizing features and parameters layer-by-layer.",
  "example_name": "Image Classification and Filter Visualization",
  "include_in_docs": true,
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.6"
  },
  "priority": 1
 },
 "nbformat": 4,
 "nbformat_minor": 0
}