1

ypwhs · Mar 24, 2017 · a6a16ac · a6a16ac
commit a6a16ac
Show file tree

Hide file tree

Showing 5 changed files with 13,217 additions and 0 deletions.
diff --git a/.gitignore b/.gitignore
@@ -0,0 +1,7 @@
+train*
+test*
+*.h5
+*.zip
+*.png
+pred.csv
+.ipy*
diff --git a/Preprocessing train dataset gap.ipynb b/Preprocessing train dataset gap.ipynb
@@ -0,0 +1,94 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {
+    "collapsed": false,
+    "deletable": true,
+    "editable": true
+   },
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "import shutil"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {
+    "collapsed": false,
+    "deletable": true,
+    "editable": true
+   },
+   "outputs": [],
+   "source": [
+    "train_filenames = os.listdir('train')\n",
+    "train_cat = filter(lambda x:x[:3] == 'cat', train_filenames)\n",
+    "train_dog = filter(lambda x:x[:3] == 'dog', train_filenames)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {
+    "collapsed": false,
+    "deletable": true,
+    "editable": true
+   },
+   "outputs": [],
+   "source": [
+    "def rmrf_mkdir(dirname):\n",
+    "    if os.path.exists(dirname):\n",
+    "        shutil.rmtree(dirname)\n",
+    "    os.mkdir(dirname)\n",
+    "\n",
+    "rmrf_mkdir('train2')\n",
+    "os.mkdir('train2/cat')\n",
+    "os.mkdir('train2/dog')\n",
+    "\n",
+    "rmrf_mkdir('test2')\n",
+    "os.symlink('../test/', 'test2/test')\n",
+    "\n",
+    "for filename in train_cat:\n",
+    "    os.symlink('../../train/'+filename, 'train2/cat/'+filename)\n",
+    "\n",
+    "for filename in train_dog:\n",
+    "    os.symlink('../../train/'+filename, 'train2/dog/'+filename)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "collapsed": true,
+    "deletable": true,
+    "editable": true
+   },
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "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": "2.7.13"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 1
+}
diff --git a/gap.ipynb b/gap.ipynb
@@ -0,0 +1,205 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {
+    "collapsed": false,
+    "deletable": true,
+    "editable": true
+   },
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Using TensorFlow backend.\n"
+     ]
+    }
+   ],
+   "source": [
+    "from keras.models import *\n",
+    "from keras.layers import *\n",
+    "from keras.applications import *\n",
+    "from keras.preprocessing.image import *\n",
+    "\n",
+    "import h5py"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {
+    "collapsed": false,
+    "deletable": true,
+    "editable": true
+   },
+   "outputs": [],
+   "source": [
+    "def write_gap(MODEL, image_size, lambda_func=None):\n",
+    "    width = image_size[0]\n",
+    "    height = image_size[1]\n",
+    "    input_tensor = Input((height, width, 3))\n",
+    "    x = input_tensor\n",
+    "    if lambda_func:\n",
+    "        x = Lambda(lambda_func)(x)\n",
+    "    base_model = MODEL(input_tensor=x, weights='imagenet', include_top=False)\n",
+    "    model = Model(base_model.input, GlobalAveragePooling2D()(base_model.output))\n",
+    "\n",
+    "    gen = ImageDataGenerator()\n",
+    "    train_generator = gen.flow_from_directory(\"train2\", image_size, shuffle=False, \n",
+    "                                              batch_size=16)\n",
+    "    test_generator = gen.flow_from_directory(\"test2\", image_size, shuffle=False, \n",
+    "                                             batch_size=16, class_mode=None)\n",
+    "\n",
+    "    train = model.predict_generator(train_generator, train_generator.nb_sample)\n",
+    "    test = model.predict_generator(test_generator, test_generator.nb_sample)\n",
+    "    with h5py.File(\"gap_%s.h5\"%MODEL.func_name) as h:\n",
+    "        h.create_dataset(\"train\", data=train)\n",
+    "        h.create_dataset(\"test\", data=test)\n",
+    "        h.create_dataset(\"label\", data=train_generator.classes)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {
+    "collapsed": false,
+    "deletable": true,
+    "editable": true,
+    "scrolled": false
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Found 25000 images belonging to 2 classes.\n",
+      "Found 12500 images belonging to 1 classes.\n"
+     ]
+    }
+   ],
+   "source": [
+    "write_gap(ResNet50, (224, 224))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {
+    "collapsed": false,
+    "deletable": true,
+    "editable": true
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Found 25000 images belonging to 2 classes.\n",
+      "Found 12500 images belonging to 1 classes.\n"
+     ]
+    }
+   ],
+   "source": [
+    "write_gap(Xception, (299, 299), xception.preprocess_input)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {
+    "collapsed": false,
+    "deletable": true,
+    "editable": true
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Found 25000 images belonging to 2 classes.\n",
+      "Found 12500 images belonging to 1 classes.\n"
+     ]
+    }
+   ],
+   "source": [
+    "write_gap(InceptionV3, (299, 299), inception_v3.preprocess_input)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {
+    "collapsed": false,
+    "deletable": true,
+    "editable": true
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Found 25000 images belonging to 2 classes.\n",
+      "Found 12500 images belonging to 1 classes.\n"
+     ]
+    }
+   ],
+   "source": [
+    "write_gap(VGG16, (224, 224))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "metadata": {
+    "collapsed": false,
+    "deletable": true,
+    "editable": true
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Found 25000 images belonging to 2 classes.\n",
+      "Found 12500 images belonging to 1 classes.\n"
+     ]
+    }
+   ],
+   "source": [
+    "write_gap(VGG19, (224, 224))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "collapsed": true
+   },
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "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": "2.7.13"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}