Using our Network - Using Convolutional Neural Network to Identify Dogs vs Cats p. 4.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "https://www.youtube.com/user/sentdex"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "import cv2\n",
    "import numpy as np\n",
    "import os\n",
    "from random import shuffle\n",
    "from tqdm import tqdm\n",
    "\n",
    "TRAIN_DIR = \"C:/Users/Karam/Codes/Preprocessing_Sentdex/Kaggle_Data/train\"\n",
    "TEST_DIR =\"C:/Users/Karam/Codes/Preprocessing_Sentdex/Kaggle_Data/test\"\n",
    "IMG_SIZE = 50\n",
    "LR = 1e-3\n",
    "\n",
    "MODEL_NAME = \"dogsvscats-{}-{}.model\".format(LR,\"8conv-basic-video\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "def label_img(img):\n",
    "    word_label = img.split(\".\")[-3] # image name is dog.93.png so -3 is dog\n",
    "    if word_label == \"cat\": return [1,0]\n",
    "    elif word_label == \"dog\" : return [0,1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_train_data():\n",
    "    training_data = []\n",
    "    for img in tqdm(os.listdir(TRAIN_DIR)):\n",
    "        label = label_img(img)\n",
    "        path = os.path.join(TRAIN_DIR,img)\n",
    "        img = cv2.resize(cv2.imread(path, cv2.IMREAD_GRAYSCALE),(IMG_SIZE,IMG_SIZE),interpolation = cv2.INTER_CUBIC)\n",
    "        training_data.append([np.array(img),np.array(label)])\n",
    "        \n",
    "    shuffle(training_data)\n",
    "    np.save(\"train_data.npy\",training_data)\n",
    "    return training_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "def process_test_data():\n",
    "    testing_data=[]\n",
    "    for img in tqdm(os.listdir(TEST_DIR)):\n",
    "        path = os.path.join(TEST_DIR,img)\n",
    "        img_num = img.split(\".\")[0]\n",
    "        img = cv2.resize(cv2.imread(path, cv2.IMREAD_GRAYSCALE),(IMG_SIZE,IMG_SIZE),interpolation = cv2.INTER_CUBIC)\n",
    "        testing_data.append([np.array(img), img_num])\n",
    "    np.save(\"test_data.npy\",testing_data)\n",
    "    return testing_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "#train_data = create_train_data()\n",
    "# if you already have train data:\n",
    "train_data = np.load('train_data.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "#testing_data = process_test_data()\n",
    "test_data = np.load(\"test_data.npy\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tflearn\n",
    "from tflearn.layers.conv import conv_2d, max_pool_2d\n",
    "from tflearn.layers.core import input_data, dropout, fully_connected\n",
    "from tflearn.layers.estimator import regression\n",
    "\n",
    "import tensorflow as tf\n",
    "tf.reset_default_graph()\n",
    "\n",
    "convnet = input_data(shape=[None, IMG_SIZE, IMG_SIZE, 1], name='input')\n",
    "\n",
    "convnet = conv_2d(convnet, 32, 5, activation='relu')\n",
    "convnet = max_pool_2d(convnet, 5)\n",
    "\n",
    "convnet = conv_2d(convnet, 64, 5, activation='relu')\n",
    "convnet = max_pool_2d(convnet, 5)\n",
    "\n",
    "convnet = conv_2d(convnet, 32, 5, activation='relu')\n",
    "convnet = max_pool_2d(convnet, 5)\n",
    "\n",
    "convnet = conv_2d(convnet, 64, 5, activation='relu')\n",
    "convnet = max_pool_2d(convnet, 5)\n",
    "\n",
    "convnet = conv_2d(convnet, 32, 5, activation='relu')\n",
    "convnet = max_pool_2d(convnet, 5)\n",
    "\n",
    "convnet = conv_2d(convnet, 64, 5, activation='relu')\n",
    "convnet = max_pool_2d(convnet, 5)\n",
    "\n",
    "convnet = conv_2d(convnet, 32, 5, activation='relu')\n",
    "convnet = max_pool_2d(convnet, 5)\n",
    "\n",
    "convnet = conv_2d(convnet, 64, 5, activation='relu')\n",
    "convnet = max_pool_2d(convnet, 5)\n",
    "\n",
    "convnet = fully_connected(convnet, 1024, activation='relu')\n",
    "convnet = dropout(convnet, 0.8)\n",
    "\n",
    "convnet = fully_connected(convnet, 1024, activation='relu')\n",
    "convnet = dropout(convnet, 0.8)\n",
    "\n",
    "convnet = fully_connected(convnet, 2, activation='softmax')\n",
    "convnet = regression(convnet, optimizer='adam', learning_rate=LR, loss='categorical_crossentropy', name='targets')\n",
    "\n",
    "model = tflearn.DNN(convnet, tensorboard_dir='log')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "if os.path.exists(\"{}.meta\".format(MODEL_NAME)):\n",
    "    model.load(MODEL_NAME)\n",
    "    print(\"model loaded!\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "train = train_data[:-500] #the data except the last 500\n",
    "test = train_data[-500:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = np.array([i[0] for i in train]).reshape(-1, IMG_SIZE,IMG_SIZE, 1) # features set, [0] is the image data\n",
    "Y = [i[1] for i in train] # Labels set\n",
    "\n",
    "test_x = np.array([i[0] for i in train]).reshape(-1, IMG_SIZE,IMG_SIZE, 1) # for testing accuracy\n",
    "test_y = [i[1] for i in train]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Step: 5893  | total loss: 0.22807 | time: 25.304s\n",
      "| Adam | epoch: 016 | loss: 0.22807 - acc: 0.9237 -- iter: 24448/24500\n",
      "Training Step: 5894  | total loss: 0.21555 | time: 29.058s\n",
      "| Adam | epoch: 016 | loss: 0.21555 - acc: 0.9298 | val_loss: 0.15795 - val_acc: 0.9404 -- iter: 24500/24500\n",
      "--\n"
     ]
    }
   ],
   "source": [
    "# train the model\n",
    "model.fit({'input': X}, {'targets': Y}, n_epoch=15, validation_set=({'input': test_x}, {'targets': test_y}), \n",
    "    snapshot_step=500, show_metric=True, run_id=MODEL_NAME)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "#tensorboard --logdir=foo:C:\\Users\\Karam\\Codes\\Preprocessing_Sentdex\\log"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:C:\\Users\\Karam\\Codes\\Preprocessing_Sentdex\\dogsvscats-0.001-8conv-basic-video.model is not in all_model_checkpoint_paths. Manually adding it.\n"
     ]
    }
   ],
   "source": [
    "#to save the model\n",
    "model.save(MODEL_NAME)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#to recall the model\n",
    "model.load(MODEL_NAME)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Test the data before Submission"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# if we don't have the data\n",
    "#test_data = preprocess_test_data\n",
    "# if we already have it\n",
    "test_data = np.load(\"test_data.npy\")\n",
    "\n",
    "fig = plt.figure() \n",
    "for num,data in enumerate(test_data[:12]): # iterate the first 12 testing data\n",
    "    # cat:[1,0] , dog: [0,1]\n",
    "    \n",
    "    img_num = data[1]\n",
    "    img_data = data[0]\n",
    "    \n",
    "    y = fig.add_subplot(3,4,num+1)\n",
    "    orig = img_data.reshape\n",
    "    data = img_data.reshape(IMG_SIZE,IMG_SIZE,1)\n",
    "    \n",
    "    model_out = model.predict([data])[0]\n",
    "    \n",
    "    if np.argmax(model_out) == 1: str_label=\"Dog\"\n",
    "    else: str_label = \"Cat\"\n",
    "        \n",
    "    y.imshow(orig,cmap=\"gray\")\n",
    "    plt.title(str_label)\n",
    "    y.axes.get_xaxis().set_visible(False)\n",
    "    y.axes.get_yaxis().set_visible(False)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"submission-file.csv\",\"w\") as f:\n",
    "    f.write(\"id,label\\n\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████████████| 12500/12500 [00:23<00:00, 543.13it/s]\n"
     ]
    }
   ],
   "source": [
    "with open(\"submission-file.csv\",\"a\") as f:\n",
    "    for data in tqdm(test_data):\n",
    "        img_num = data[1]\n",
    "        img_data = data[0]\n",
    "        orig = img_data\n",
    "        data = img_data.reshape(IMG_SIZE,IMG_SIZE,1)\n",
    "        model_out = model.predict([data])[0]\n",
    "        f.write('{},{}\\n'.format(img_num,model_out[1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0.96610725 0.03389278]]\n"
     ]
    }
   ],
   "source": [
    "print(model.predict([data]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}