CvD_train.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Google Machinelearing Challenge\n",
    "#### Team 루프탑"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "from glob import glob\n",
    "\n",
    "import pandas as pd # Dataframe\n",
    "import numpy as np #Linear Algebra\n",
    "import matplotlib.pyplot as plt #Visualization\n",
    "%matplotlib inline\n",
    "from skimage import color, io \n",
    "from scipy.misc import imresize \n",
    "#from PIL import Image\n",
    "\n",
    "import tflearn\n",
    "from tflearn.data_utils import shuffle, to_categorical\n",
    "from tflearn.layers.core import input_data, dropout, fully_connected\n",
    "from tflearn.layers.conv import conv_2d, max_pool_2d\n",
    "from tflearn.layers.estimator import regression\n",
    "from tflearn.data_preprocessing import ImagePreprocessing\n",
    "from tflearn.data_augmentation import ImageAugmentation\n",
    "from tflearn.metrics import Accuracy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## EDA\n",
    "- 개 고양이 사진 분류\n",
    "- Train set\n",
    "    - cats : 7,854\n",
    "    - dogs : 7,872\n",
    "- Validation\n",
    "    - cats : 989\n",
    "    - dogs : 977\n",
    "- Test set\n",
    "    - 1,963\n",
    "   \n",
    "   \n",
    "## Data Import"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "files_path = './data'\n",
    "cat_files_path = os.path.join(files_path, 'train/cats/*.jpg')\n",
    "dog_files_path = os.path.join(files_path, 'train/dogs/*.jpg')\n",
    "cat_val_path = os.path.join(files_path, 'validation/cats/*.jpg')\n",
    "dog_val_path = os.path.join(files_path, 'validation/dogs/*.jpg')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "./data/train/dogs/*.jpg\n",
      "./data/validation/dogs/*.jpg\n"
     ]
    }
   ],
   "source": [
    "cat_files = sorted(glob(cat_files_path))\n",
    "dog_files = sorted(glob(dog_files_path))\n",
    "train_files = cat_files + dog_files\n",
    "cat_val = sorted(glob(cat_val_path))\n",
    "dog_val = sorted(glob(dog_val_path))\n",
    "val_files = cat_val + dog_val\n",
    "print(dog_files_path)\n",
    "print(dog_val_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "15724\n",
      "1964\n"
     ]
    }
   ],
   "source": [
    "n_files = len(train_files)\n",
    "n_val = len(val_files)\n",
    "print(n_files)\n",
    "print(n_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "image_size = 128\n",
    "X = np.zeros((n_files, image_size, image_size, 3), dtype='float64')\n",
    "y = np.zeros(n_files)\n",
    "X_val = np.zeros((n_val, image_size, image_size, 3), dtype='float64')\n",
    "y_val = np.zeros(n_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/mac/anaconda/lib/python3.6/site-packages/PIL/TiffImagePlugin.py:692: UserWarning: Possibly corrupt EXIF data.  Expecting to read 8589934590 bytes but only got 37573. Skipping tag 34855\n",
      "  \"Skipping tag %s\" % (size, len(data), tag))\n",
      "/Users/mac/anaconda/lib/python3.6/site-packages/PIL/TiffImagePlugin.py:692: UserWarning: Possibly corrupt EXIF data.  Expecting to read 8589934590 bytes but only got 28996. Skipping tag 34855\n",
      "  \"Skipping tag %s\" % (size, len(data), tag))\n",
      "/Users/mac/anaconda/lib/python3.6/site-packages/PIL/TiffImagePlugin.py:709: UserWarning: Corrupt EXIF data.  Expecting to read 2 bytes but only got 0. \n",
      "  warnings.warn(str(msg))\n",
      "/Users/mac/anaconda/lib/python3.6/site-packages/PIL/TiffImagePlugin.py:692: UserWarning: Possibly corrupt EXIF data.  Expecting to read 2818048 bytes but only got 0. Skipping tag 0\n",
      "  \"Skipping tag %s\" % (size, len(data), tag))\n",
      "/Users/mac/anaconda/lib/python3.6/site-packages/PIL/TiffImagePlugin.py:709: UserWarning: Corrupt EXIF data.  Expecting to read 12 bytes but only got 11. \n",
      "  warnings.warn(str(msg))\n",
      "/Users/mac/anaconda/lib/python3.6/site-packages/PIL/TiffImagePlugin.py:692: UserWarning: Possibly corrupt EXIF data.  Expecting to read 1114112 bytes but only got 0. Skipping tag 0\n",
      "  \"Skipping tag %s\" % (size, len(data), tag))\n",
      "/Users/mac/anaconda/lib/python3.6/site-packages/PIL/TiffImagePlugin.py:692: UserWarning: Possibly corrupt EXIF data.  Expecting to read 3473408 bytes but only got 0. Skipping tag 0\n",
      "  \"Skipping tag %s\" % (size, len(data), tag))\n"
     ]
    }
   ],
   "source": [
    "count = 0\n",
    "for f in cat_files:\n",
    "    try:\n",
    "        img = io.imread(f)\n",
    "        new_img = imresize(img, (image_size, image_size, 3))\n",
    "        X[count] = np.array(new_img)\n",
    "        y[count] = 0\n",
    "        count += 1\n",
    "    except:\n",
    "        continue\n",
    "\n",
    "for f in dog_files:\n",
    "    try:\n",
    "        img = io.imread(f)\n",
    "        new_img = imresize(img, (image_size, image_size, 3))\n",
    "        X[count] = np.array(new_img)\n",
    "        y[count] = 1\n",
    "        count += 1\n",
    "    except:\n",
    "        continue"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "count = 0\n",
    "for f in cat_val:\n",
    "    try:\n",
    "        img = io.imread(f)\n",
    "        new_img = imresize(img, (image_size, image_size, 3))\n",
    "        X_val[count] = np.array(new_img)\n",
    "        y_val[count] = 0\n",
    "        count += 1\n",
    "    except:\n",
    "        continue\n",
    "\n",
    "\n",
    "for f in dog_val:\n",
    "    try:\n",
    "        img = io.imread(f)\n",
    "        new_img = imresize(img, (image_size, image_size, 3))\n",
    "        X_val[count] = np.array(new_img)\n",
    "        y_val[count] = 1\n",
    "        count += 1\n",
    "    except:\n",
    "        continue"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "perm = np.arange(len(X))\n",
    "np.random.shuffle(perm)\n",
    "X = X[perm]\n",
    "y = y[perm]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "perm = np.arange(len(X_val))\n",
    "np.random.shuffle(perm)\n",
    "X_val = X_val[perm]\n",
    "y_val = y_val[perm]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "Y = to_categorical(y, 2)\n",
    "Y_val = to_categorical(y_val, 2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "img_prep = ImagePreprocessing()\n",
    "img_prep.add_featurewise_zero_center()\n",
    "img_prep.add_featurewise_stdnorm()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "img_aug = ImageAugmentation()\n",
    "img_aug.add_random_flip_leftright()\n",
    "img_aug.add_random_rotation(max_angle=25.)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "network = input_data(shape=[None, image_size, image_size, 3],\n",
    "                     data_preprocessing=img_prep,\n",
    "                     data_augmentation=img_aug)\n",
    "network = conv_2d(network, 32, 3, activation='relu')\n",
    "network = max_pool_2d(network, 2)\n",
    "network = conv_2d(network, 64, 3, activation='relu')\n",
    "network = conv_2d(network, 64, 3, activation='relu')\n",
    "network = max_pool_2d(network, 2)\n",
    "network = fully_connected(network, 512, activation='relu')\n",
    "network = dropout(network, 0.5)\n",
    "network = fully_connected(network, 2, activation='softmax')\n",
    "acc = Accuracy(name=\"Accuracy\")\n",
    "network = regression(network, optimizer='adam',\n",
    "                     loss='categorical_crossentropy',\n",
    "                     learning_rate=0.0005, metric=acc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# 서버에서 학습시킨 코드\n",
    "model = tflearn.DNN(network, checkpoint_path='MLCrooftop.tflearn', max_checkpoints = 3,\n",
    "                    tensorboard_verbose = 3, tensorboard_dir='tmp/tflearn_logs/')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Step: 31  | total loss: \u001b[1m\u001b[32m0.68143\u001b[0m\u001b[0m | time: 650.827s\n",
      "| Adam | epoch: 001 | loss: 0.68143 - Accuracy: 0.5561 -- iter: 15500/15724\n",
      "Training Step: 32  | total loss: \u001b[1m\u001b[32m0.68094\u001b[0m\u001b[0m | time: 678.876s\n",
      "| Adam | epoch: 001 | loss: 0.68094 - Accuracy: 0.5610 | val_loss: 0.67645 - val_acc: 0.5820 -- iter: 15724/15724\n",
      "--\n"
     ]
    }
   ],
   "source": [
    "model.fit(X, Y, validation_set=(X_val, Y_val), batch_size=500,\n",
    "      n_epoch=1, run_id='MLCrooftop', show_metric=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- 실제 서버에서는 100 epoch 실행"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "model.save('MLCrooftop_final.tflearn')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAQUAAAEICAYAAABWCOFPAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f74733e6eb8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(X[1])\n",
    "plt.axis('off')\n",
    "plt.title('Sample image with label {}'.format(Y[1]))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.42874637,  0.5712536 ],\n",
       "       [ 0.60955197,  0.390448  ],\n",
       "       [ 0.52209407,  0.47790587],\n",
       "       [ 0.53380615,  0.46619383],\n",
       "       [ 0.50968468,  0.49031526],\n",
       "       [ 0.49538428,  0.50461572],\n",
       "       [ 0.56830901,  0.43169096],\n",
       "       [ 0.50978982,  0.49021012],\n",
       "       [ 0.43235639,  0.56764364],\n",
       "       [ 0.48608565,  0.51391435]], dtype=float32)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.predict(X_val[:10])"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}