CIFAR10_Extra.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 105
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 15465,
     "status": "ok",
     "timestamp": 1527607949448,
     "user": {
      "displayName": "Zeki Kuş",
      "photoUrl": "https://lh3.googleusercontent.com/a/default-user=s128",
      "userId": "113800785453624523681"
     },
     "user_tz": -180
    },
    "id": "eZ3xO8XDZBfy",
    "outputId": "cb8081e8-891c-4a52-b5d5-be2bcb01606f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Please, open the following URL in a web browser: https://accounts.google.com/o/oauth2/auth?client_id=32555940559.apps.googleusercontent.com&redirect_uri=urn%3Aietf%3Awg%3Aoauth%3A2.0%3Aoob&scope=https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdrive&response_type=code&access_type=offline&approval_prompt=force\r\n",
      "··········\n",
      "Please, open the following URL in a web browser: https://accounts.google.com/o/oauth2/auth?client_id=32555940559.apps.googleusercontent.com&redirect_uri=urn%3Aietf%3Awg%3Aoauth%3A2.0%3Aoob&scope=https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdrive&response_type=code&access_type=offline&approval_prompt=force\n",
      "Please enter the verification code: Access token retrieved correctly.\n"
     ]
    }
   ],
   "source": [
    "!apt-get install -y -qq software-properties-common python-software-properties module-init-tools\n",
    "!add-apt-repository -y ppa:alessandro-strada/ppa 2>&1 > /dev/null\n",
    "!apt-get update -qq 2>&1 > /dev/null\n",
    "!apt-get -y install -qq google-drive-ocamlfuse fuse\n",
    "from google.colab import auth\n",
    "auth.authenticate_user()\n",
    "from oauth2client.client import GoogleCredentials\n",
    "creds = GoogleCredentials.get_application_default()\n",
    "import getpass\n",
    "!google-drive-ocamlfuse -headless -id={creds.client_id} -secret={creds.client_secret} < /dev/null 2>&1 | grep URL\n",
    "vcode = getpass.getpass()\n",
    "!echo {vcode} | google-drive-ocamlfuse -headless -id={creds.client_id} -secret={creds.client_secret}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "D00Qk8m3ZF7N"
   },
   "outputs": [],
   "source": [
    "!mkdir -p drive\n",
    "!google-drive-ocamlfuse drive"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1691,
     "status": "ok",
     "timestamp": 1527607957128,
     "user": {
      "displayName": "Zeki Kuş",
      "photoUrl": "https://lh3.googleusercontent.com/a/default-user=s128",
      "userId": "113800785453624523681"
     },
     "user_tz": -180
    },
    "id": "F_AD0Kg7ZIav",
    "outputId": "72781e4c-dd67-4761-a0b0-702a851b8094"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import tensorflow as tf\n",
    "from sklearn.model_selection import train_test_split\n",
    "from keras.utils import to_categorical\n",
    "import matplotlib.pyplot as plt\n",
    "import keras\n",
    "from keras.datasets import cifar10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 85939,
     "status": "ok",
     "timestamp": 1527608157928,
     "user": {
      "displayName": "Zeki Kuş",
      "photoUrl": "https://lh3.googleusercontent.com/a/default-user=s128",
      "userId": "113800785453624523681"
     },
     "user_tz": -180
    },
    "id": "hDLtWk3YgZZ6",
    "outputId": "d43d72e3-5d7e-4610-c864-2787d7eb5875"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading data from https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz\n",
      "126066688/170498071 [=====================>........] - ETA: 23s170500096/170498071 [==============================] - 82s 0us/step\n"
     ]
    }
   ],
   "source": [
    "(x_train, y_train), (x_test, y_test) = cifar10.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "NLA2494IZPJX"
   },
   "outputs": [],
   "source": [
    "input_shape = [None, 32, 32, 3]\n",
    "number_of_classes = 10\n",
    "\n",
    "#Conver output label to one hot vector\n",
    "y_train = to_categorical(y_train, number_of_classes)\n",
    "y_test = to_categorical(y_test, number_of_classes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 85
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1166,
     "status": "ok",
     "timestamp": 1527608160452,
     "user": {
      "displayName": "Zeki Kuş",
      "photoUrl": "https://lh3.googleusercontent.com/a/default-user=s128",
      "userId": "113800785453624523681"
     },
     "user_tz": -180
    },
    "id": "0X-ZefV4hi1z",
    "outputId": "0fd0a196-065c-46a1-9955-06ba5d34a700"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of training features: (50000, 32, 32, 3)\n",
      "Shape of training lables: (50000, 10)\n",
      "Shape of testing features: (10000, 32, 32, 3)\n",
      "Shape of testing lables: (10000, 10)\n"
     ]
    }
   ],
   "source": [
    "print(\"Shape of training features: {}\".format(x_train.shape))\n",
    "print(\"Shape of training lables: {}\".format(y_train.shape))\n",
    "print(\"Shape of testing features: {}\".format(x_test.shape))\n",
    "print(\"Shape of testing lables: {}\".format(y_test.shape))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "dusSPSGEhQ0v"
   },
   "outputs": [],
   "source": [
    "#Hyper parameters\n",
    "learning_rate = 0.01\n",
    "epoch = 10\n",
    "batch_size = 32"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 364
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 767,
     "status": "ok",
     "timestamp": 1527417860442,
     "user": {
      "displayName": "Zeki Kuş",
      "photoUrl": "https://lh3.googleusercontent.com/a/default-user=s128",
      "userId": "113800785453624523681"
     },
     "user_tz": -180
    },
    "id": "rtF8G7YZZ8IJ",
    "outputId": "806d44ef-7efd-440a-e868-07bc3d808958"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f8545c35e48>"
      ]
     },
     "execution_count": 8,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUsAAAFKCAYAAACU6307AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nmvAvz70cAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8584eb5f98>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(x_train[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1629,
     "status": "ok",
     "timestamp": 1527608176660,
     "user": {
      "displayName": "Zeki Kuş",
      "photoUrl": "https://lh3.googleusercontent.com/a/default-user=s128",
      "userId": "113800785453624523681"
     },
     "user_tz": -180
    },
    "id": "5An8SAuhaQF7",
    "outputId": "8c652b02-0722-4a07-d27c-1ba3d15a2d31"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of examples in training set: 45000\n",
      "Number of examples in cross validation set: 5000\n"
     ]
    }
   ],
   "source": [
    "x_train, cv_x, y_train, cv_y = train_test_split(x_train, y_train, \n",
    "                                                test_size = 5000, random_state = 42)\n",
    "\n",
    "print(\"Number of examples in training set: {}\".format(x_train.shape[0]))\n",
    "print(\"Number of examples in cross validation set: {}\".format(cv_x.shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "6wuLKvcRkamg"
   },
   "outputs": [],
   "source": [
    "x_train = x_train.astype('float32')\n",
    "x_test = x_test.astype('float32')\n",
    "x_train /= 255\n",
    "x_test /= 255"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "wjUQYHBSiPik"
   },
   "outputs": [],
   "source": [
    "#GRAPH\n",
    "\n",
    "#Resetting graph\n",
    "tf.reset_default_graph()\n",
    "\n",
    "#Defining Placeholders\n",
    "x = tf.placeholder(tf.float32,shape=[None,32,32,3])\n",
    "y_true = tf.placeholder(tf.float32,shape=[None,10])\n",
    "train_mode = tf.placeholder(tf.bool)\n",
    "\n",
    "config = tf.ConfigProto(allow_soft_placement=True)\n",
    "config.gpu_options.allow_growth = True\n",
    "config.gpu_options.allocator_type = 'BFC'\n",
    "\n",
    "filter_size = 3\n",
    "\n",
    "with tf.name_scope('Model'):\n",
    "    #Convolutional Layer 1 + RELU\n",
    "    convo_1 = tf.layers.conv2d(inputs=x,filters=32, kernel_size = [filter_size,filter_size], padding=\"same\", activation=tf.nn.relu)\n",
    "    #Convolutional Layer 2 + RELU\n",
    "    convo_2 = tf.layers.conv2d(inputs=convo_1,filters=64, kernel_size = [filter_size,filter_size], activation=tf.nn.relu)\n",
    "    #Pooling Layer 1\n",
    "    pooling_1 = tf.layers.max_pooling2d(inputs=convo_2, pool_size=[2, 2], strides=2)\n",
    "    dropout_1 = tf.layers.dropout(pooling_1,rate=0.25, training= train_mode)\n",
    "    \n",
    "    #Convolutional Layer 3 + RELU\n",
    "    convo_3 = tf.layers.conv2d(inputs=dropout_1,filters=64, kernel_size = [filter_size,filter_size], padding=\"same\", activation=tf.nn.relu)\n",
    "    #Convolutional Layer 4 + RELU\n",
    "    convo_4 = tf.layers.conv2d(inputs=convo_3,filters=128, kernel_size = [filter_size,filter_size], activation=tf.nn.relu)\n",
    "    #Pooling Layer 2\n",
    "    pooling_2 = tf.layers.max_pooling2d(inputs=convo_4, pool_size=[2, 2], strides=2)\n",
    "    dropout_2 = tf.layers.dropout(pooling_2,rate=0.25, training= train_mode)\n",
    "    \n",
    "    #Flattening\n",
    "    flatten = tf.reshape(dropout_2,[-1,6*6*128])\n",
    "    #Fully Connected 1 + RELU\n",
    "    full_layer_one = tf.layers.dense(inputs=flatten, units = 512, activation=tf.nn.relu)\n",
    "    #Dropout Layer 1\n",
    "    full_one_dropout = tf.layers.dropout(full_layer_one,rate=0.5, training= train_mode)\n",
    "    #Output Layer,containing 2 output nodes.\n",
    "    y_pred = tf.layers.dense(inputs=full_one_dropout,units=10)\n",
    "\n",
    "#Loss function\n",
    "with tf.name_scope('Loss'):\n",
    "    loss = tf.losses.softmax_cross_entropy(y_true, y_pred)\n",
    "\n",
    "#Optimizer Function\n",
    "with tf.name_scope('ADAM'):\n",
    "     train_op = tf.train.AdamOptimizer(learning_rate=0.0001).minimize(loss)\n",
    "\n",
    "#Defining Accuracy\n",
    "with tf.name_scope('Accuracy'):\n",
    "    accuracy = tf.reduce_mean(tf.cast(tf.equal(tf.argmax(y_pred, 1), tf.argmax(y_true, 1)), tf.float32))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 238
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 278046,
     "status": "ok",
     "timestamp": 1527274372656,
     "user": {
      "displayName": "Zeki Kuş",
      "photoUrl": "https://lh3.googleusercontent.com/a/default-user=s128",
      "userId": "113800785453624523681"
     },
     "user_tz": -180
    },
    "id": "hyTMyp27korl",
    "outputId": "2b096c99-2a13-47fd-f7d0-fdeae39df00c"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training:Start\n",
      "Epoch: 1 Cost: 1.717341889527919 accuracy: 0.3974991115849325 \n",
      "Epoch: 2 Cost: 1.3639636663759755 accuracy: 0.542866027007818 \n",
      "Epoch: 3 Cost: 1.2148108573505922 accuracy: 0.6045664534470505 \n",
      "Epoch: 4 Cost: 1.1090097642724916 accuracy: 0.6480987917555082 \n",
      "Epoch: 5 Cost: 1.021113065279395 accuracy: 0.681703091684435 \n",
      "Epoch: 6 Cost: 0.9508058969487846 accuracy: 0.709155117270789 \n",
      "Epoch: 7 Cost: 0.8878147993581224 accuracy: 0.7331867448471926 \n",
      "Epoch: 8 Cost: 0.8376032652208849 accuracy: 0.7542643923240938 \n",
      "Epoch: 9 Cost: 0.7900279221544904 accuracy: 0.7730765813788202 \n",
      "Epoch: 10 Cost: 0.7431189785810826 accuracy: 0.7914001421464109 \n",
      "[3]\n",
      "Model saved in path: drive/Colab Notebooks/CIFAR-10/model.ckpt\n"
     ]
    }
   ],
   "source": [
    "saver = tf.train.Saver()\n",
    "\n",
    "with tf.Session(config=config) as sess:\n",
    "    sess.run(tf.global_variables_initializer())\n",
    "    \n",
    "    \n",
    "    total_size = x_train.shape[0]\n",
    "    number_of_batches = int(total_size/batch_size)\n",
    "    \n",
    "    print(\"Training:Start\")\n",
    "    for e in range(epoch):\n",
    "        epoch_cost = 0\n",
    "        epoch_accuracy = 0\n",
    "        for i in range(number_of_batches):\n",
    "            mini_x = x_train[i*batch_size:(i+1)*batch_size, :, :, :]\n",
    "            mini_y = y_train[i*batch_size:(i+1)*batch_size, :]\n",
    "            _, cost = sess.run([train_op,loss], feed_dict={x:mini_x, y_true:mini_y,train_mode:True})\n",
    "            train_accuracy = sess.run(accuracy, feed_dict={x:mini_x, y_true:mini_y,train_mode:False})\n",
    "            epoch_cost += cost\n",
    "            epoch_accuracy += train_accuracy\n",
    "        epoch_cost /= number_of_batches\n",
    "        \n",
    "        if total_size % batch_size != 0:\n",
    "            epoch_accuracy /= (number_of_batches+1)\n",
    "        else:\n",
    "            epoch_accuracy /= number_of_batches\n",
    "        print(\"Epoch: {} Cost: {} accuracy: {} \".format(e+1, np.squeeze(epoch_cost), epoch_accuracy))\n",
    "    \n",
    "    a = tf.argmax(y_pred, 1)\n",
    "    print(a.eval(feed_dict={x:x_train[0:1,:,:,:],train_mode:False}))\n",
    "    save_path = saver.save(sess, \"drive/Colab Notebooks/CIFAR-10/model.ckpt\")\n",
    "    print(\"Model saved in path: %s\" % save_path)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Eğitilen Ağın Performansının Daha Doğru Bir Şekilde Ölçülebilmesi İçin Farklı Random-State Değerleri ile Ağı 5 Defa Daha Sıfırdan Eğitip Elde Edilen Doğruluk Oranlarının Ortalamasını Hesaplıyorum.\n",
    "\n",
    " Böylece bir önceki kısımdaki gibi ağı sürekli aynı test ve train verileri ile eğitmek yerine her çalıştırmada(run) farklı train ve test verilerini kullandım."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "pho-TTW8Qddv"
   },
   "outputs": [],
   "source": [
    "X = np.concatenate((x_train[:,:,:,:],x_test[:,:,:]),axis=0)\n",
    "y = np.concatenate((y_train[:,:],y_test[:,:]),axis=0)\n",
    "\n",
    "from  sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 1224
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1236278,
     "status": "ok",
     "timestamp": 1527421057693,
     "user": {
      "displayName": "Zeki Kuş",
      "photoUrl": "https://lh3.googleusercontent.com/a/default-user=s128",
      "userId": "113800785453624523681"
     },
     "user_tz": -180
    },
    "id": "54BGoTB_1Z3h",
    "outputId": "bab2ae4e-c514-45ee-996f-9e93ece198e9"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random = 0\n",
      "Training:Start\n",
      "Epoch: 1 Cost: 1.7284195787256413 accuracy: 0.3945 \n",
      "Epoch: 2 Cost: 1.397501811764457 accuracy: 0.5297045454545455 \n",
      "Epoch: 3 Cost: 1.252150617686185 accuracy: 0.5888181818181818 \n",
      "Epoch: 4 Cost: 1.1481414057558232 accuracy: 0.6297272727272727 \n",
      "Epoch: 5 Cost: 1.0646370544000106 accuracy: 0.6659772727272727 \n",
      "Epoch: 6 Cost: 0.992348838372664 accuracy: 0.69575 \n",
      "Epoch: 7 Cost: 0.9291473836898804 accuracy: 0.72075 \n",
      "Epoch: 8 Cost: 0.8761960230740634 accuracy: 0.7414318181818181 \n",
      "Epoch: 9 Cost: 0.8261348401849921 accuracy: 0.7615681818181819 \n",
      "Epoch: 10 Cost: 0.7804373869245702 accuracy: 0.7794318181818182 \n",
      "Test Accuracy: 0.7170239825581395\n",
      "**************************************************\n",
      "Random = 1\n",
      "Training:Start\n",
      "Epoch: 1 Cost: 1.7023617825508117 accuracy: 0.4027045454545455 \n",
      "Epoch: 2 Cost: 1.3719317113269458 accuracy: 0.5357727272727273 \n",
      "Epoch: 3 Cost: 1.229983094215393 accuracy: 0.5961136363636363 \n",
      "Epoch: 4 Cost: 1.1141033100214872 accuracy: 0.6421136363636364 \n",
      "Epoch: 5 Cost: 1.029150279933756 accuracy: 0.6755454545454546 \n",
      "Epoch: 6 Cost: 0.9558794853470542 accuracy: 0.7026136363636364 \n",
      "Epoch: 7 Cost: 0.8951661304343831 accuracy: 0.7279318181818182 \n",
      "Epoch: 8 Cost: 0.8439058328541842 accuracy: 0.7492727272727273 \n",
      "Epoch: 9 Cost: 0.8007489295222543 accuracy: 0.7673409090909091 \n",
      "Epoch: 10 Cost: 0.7572150414640253 accuracy: 0.7865454545454545 \n",
      "Test Accuracy: 0.7298328488372093\n",
      "**************************************************\n",
      "Random = 2\n",
      "Training:Start\n",
      "Epoch: 1 Cost: 1.7531994571685792 accuracy: 0.38729545454545455 \n",
      "Epoch: 2 Cost: 1.4109971478635615 accuracy: 0.5222727272727272 \n",
      "Epoch: 3 Cost: 1.2598372336300936 accuracy: 0.5832954545454545 \n",
      "Epoch: 4 Cost: 1.1538263282342391 accuracy: 0.6298409090909091 \n",
      "Epoch: 5 Cost: 1.0740471850308504 accuracy: 0.6631363636363636 \n",
      "Epoch: 6 Cost: 0.9985068596493114 accuracy: 0.6914772727272728 \n",
      "Epoch: 7 Cost: 0.9386087618741122 accuracy: 0.7147272727272728 \n",
      "Epoch: 8 Cost: 0.8939875820550052 accuracy: 0.7354318181818181 \n",
      "Epoch: 9 Cost: 0.844056420174512 accuracy: 0.7535 \n",
      "Epoch: 10 Cost: 0.7997596893960779 accuracy: 0.7721363636363636 \n",
      "Test Accuracy: 0.7198401162790697\n",
      "**************************************************\n",
      "Random = 3\n",
      "Training:Start\n",
      "Epoch: 1 Cost: 1.7060114365491 accuracy: 0.40493181818181817 \n",
      "Epoch: 2 Cost: 1.3580776779868386 accuracy: 0.5435227272727273 \n",
      "Epoch: 3 Cost: 1.2176482822678305 accuracy: 0.6011136363636363 \n",
      "Epoch: 4 Cost: 1.1043910621729764 accuracy: 0.6458181818181818 \n",
      "Epoch: 5 Cost: 1.0229510558301753 accuracy: 0.6810454545454545 \n",
      "Epoch: 6 Cost: 0.9530781009630723 accuracy: 0.7100909090909091 \n",
      "Epoch: 7 Cost: 0.8927466751011935 accuracy: 0.7333863636363637 \n",
      "Epoch: 8 Cost: 0.8416699346845801 accuracy: 0.7540681818181818 \n",
      "Epoch: 9 Cost: 0.7967303344769912 accuracy: 0.7721363636363636 \n",
      "Epoch: 10 Cost: 0.7516635195341977 accuracy: 0.7902045454545454 \n",
      "Test Accuracy: 0.7356468023255814\n",
      "**************************************************\n",
      "Random = 4\n",
      "Training:Start\n",
      "Epoch: 1 Cost: 1.6897235353643243 accuracy: 0.40925 \n",
      "Epoch: 2 Cost: 1.3600561467084018 accuracy: 0.5448863636363637 \n",
      "Epoch: 3 Cost: 1.2070882118398494 accuracy: 0.6078636363636364 \n",
      "Epoch: 4 Cost: 1.1023640940839594 accuracy: 0.6496590909090909 \n",
      "Epoch: 5 Cost: 1.020008610118519 accuracy: 0.68175 \n",
      "Epoch: 6 Cost: 0.9503960934552279 accuracy: 0.7099545454545455 \n",
      "Epoch: 7 Cost: 0.8914544835524125 accuracy: 0.7341363636363636 \n",
      "Epoch: 8 Cost: 0.8414965019442818 accuracy: 0.7541363636363636 \n",
      "Epoch: 9 Cost: 0.7956912612264807 accuracy: 0.7718863636363636 \n",
      "Epoch: 10 Cost: 0.7548686705069109 accuracy: 0.7886136363636363 \n",
      "Test Accuracy: 0.7274709302325582\n",
      "**************************************************\n",
      "Mean Accuracy: 0.7259629360465117\n"
     ]
    }
   ],
   "source": [
    "# VERİ SETİNİN FARKLI RANDOM-STATE DEĞERLERİ İÇİN TEKRAR EĞİTİLMESİ VE TEST EDİLMESİ\n",
    "\n",
    "with tf.Session(config=config) as sess:\n",
    "    \n",
    "   \n",
    "  random_states = [10, 42, 83, 143, 251]\n",
    "  test_accuracies = []\n",
    "  batch_test_acc = []\n",
    "\n",
    "  \n",
    "  for i in range(5):\n",
    "    sess.run(tf.global_variables_initializer())\n",
    "    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.20, random_state=random_states[i])\n",
    "\n",
    "    total_size = X_train.shape[0]\n",
    "    number_of_batches = int(total_size/batch_size)\n",
    "    print(\"Random =\", str(i))\n",
    "    \n",
    "    print(\"Training:Start\")\n",
    "    # Training Epoch\n",
    "    for e in range(epoch):\n",
    "        # Training Batch\n",
    "        epoch_cost = 0\n",
    "        epoch_accuracy = 0\n",
    "        for i in range(number_of_batches):\n",
    "            mini_x = X_train[i*batch_size:(i+1)*batch_size, :, :, :]\n",
    "            mini_y = y_train[i*batch_size:(i+1)*batch_size, :]\n",
    "            _, cost = sess.run([train_op,loss], feed_dict={x:mini_x, y_true:mini_y,train_mode:True})\n",
    "            train_accuracy = sess.run(accuracy, feed_dict={x:mini_x, y_true:mini_y,train_mode:False})\n",
    "            epoch_cost += cost\n",
    "            epoch_accuracy += train_accuracy\n",
    "        epoch_cost /= number_of_batches\n",
    "\n",
    "        if total_size % batch_size != 0:\n",
    "            epoch_accuracy /= (number_of_batches+1)\n",
    "        else:\n",
    "            epoch_accuracy /= number_of_batches\n",
    "        print(\"Epoch: {} Cost: {} accuracy: {} \".format(e+1, np.squeeze(epoch_cost), epoch_accuracy))\n",
    "    \n",
    "    # Test Batch\n",
    "    total_size = y_test.shape[0]\n",
    "    number_of_batches = int(total_size/batch_size)\n",
    "    test_accuracy = 0\n",
    "    for i in range(number_of_batches):\n",
    "      mini_x = X_test[i*batch_size:(i+1)*batch_size, :, :, :]\n",
    "      mini_y = y_test[i*batch_size:(i+1)*batch_size, :]\n",
    "      batch_accuracy = sess.run(accuracy, feed_dict={x:mini_x, y_true:mini_y,train_mode:False})\n",
    "      test_accuracy += batch_accuracy\n",
    "\n",
    "    if total_size % batch_size != 0:\n",
    "        test_accuracy /= (number_of_batches+1)\n",
    "    else:\n",
    "        test_accuracy /= number_of_batches\n",
    "    print(\"Test Accuracy:\", test_accuracy)\n",
    "\n",
    "    test_accuracies.append(test_accuracy)\n",
    "    print(\"*\"*50)\n",
    "  print(\"Mean Accuracy:\", np.mean(test_accuracies))\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 68
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 3210,
     "status": "ok",
     "timestamp": 1527608186671,
     "user": {
      "displayName": "Zeki Kuş",
      "photoUrl": "https://lh3.googleusercontent.com/a/default-user=s128",
      "userId": "113800785453624523681"
     },
     "user_tz": -180
    },
    "id": "wRb1u_lTgr27",
    "outputId": "913bcc97-3cc3-4491-fa1d-810e042a4277"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Restoring parameters from drive/Colab Notebooks/CIFAR-10/model.ckpt\n",
      "Actual Class: [[0. 0. 0. 0. 0. 0. 0. 1. 0. 0.]]\n",
      "Predict Class: [7]\n"
     ]
    }
   ],
   "source": [
    "# Eğitilen modeli istenilen isim ile kaydedebilirsiniz.\n",
    "\n",
    "# Eğitilen model ile tahmin işleminin gerçekleştirilmesi\n",
    "saver = tf.train.Saver()\n",
    "with tf.Session(config=config) as sess:\n",
    "    saver.restore(sess, \"drive/Colab Notebooks/CIFAR-10/model.ckpt\")\n",
    "    prediction = tf.argmax(y_pred, 1)\n",
    "    print(\"Actual Class:\",y_train[100:101,:])\n",
    "    print(\"Predict Class:\",prediction.eval(feed_dict={x:x_train[100:101,:,:,:],train_mode:False}))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "Tv99gFncZ5W7"
   },
   "outputs": [],
   "source": [
    "import os\n",
    "import cv2\n",
    "from tqdm import tqdm\n",
    "\n",
    "TEST_DIR = 'drive/Colab Notebooks/CIFAR-10/Test_Images'\n",
    "IMG_SIZE = 32\n",
    "\n",
    "# Bu kısmı dışarıdan kullanıcı resim verdiğinde direkt olarak ağın anlayabileceği sayısal değerlere nasıl dönüştürüleceğini anlatabilmek adına yazdım.\n",
    "# Verilen klasördeki resimler okunur ve ağın anlayabileceği sayısal verilere dönüştürülür. OpenCV kütüphanesi ile kullanılır.\n",
    "def process_test_data():\n",
    "    testing_data = []\n",
    "    for img in tqdm(os.listdir(TEST_DIR)):\n",
    "        label = img\n",
    "        path = TEST_DIR + '/' + img\n",
    "        img = cv2.imread(path,1)\n",
    "        img = cv2.resize(img, (IMG_SIZE,IMG_SIZE))\n",
    "        testing_data.append([np.array(img),label])\n",
    "      \n",
    "    # Elde edilen matris istenirse .npy formatında kaydedilip. Paylaşılabilir.\n",
    "    #shuffle(testing_data)\n",
    "    #np.save('test_data_bi.npy', testing_data)\n",
    "    return testing_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "xwit8Y7Xi1ih"
   },
   "outputs": [],
   "source": [
    "# Load Data\n",
    "my_images = process_test_data()\n",
    "\n",
    "# Reshape Data\n",
    "test_img = np.array([i[0] for i in my_images]).reshape(-1,IMG_SIZE,IMG_SIZE,3)\n",
    "labels = np.array([i[1] for i in my_images])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 398
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1854,
     "status": "ok",
     "timestamp": 1527610384311,
     "user": {
      "displayName": "Zeki Kuş",
      "photoUrl": "https://lh3.googleusercontent.com/a/default-user=s128",
      "userId": "113800785453624523681"
     },
     "user_tz": -180
    },
    "id": "S1mHnZbulH2C",
    "outputId": "fa1d0d17-abec-4aeb-cf76-0f5a2a2e65ee"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Restoring parameters from drive/Colab Notebooks/CIFAR-10/model.ckpt\n",
      "Actual Class: ucak.jpg\n",
      "Predict Class: Uçak\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUsAAAFKCAYAAACU6307AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f1dc469cf60>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "saver = tf.train.Saver()\n",
    "cifar_classes = {0:'Uçak',1:'Otomobil',2:'Kuş',3:'Kedi',4:'Geyik',5:'Köpek',6:'Kurbağa',7:'At',8:'Gemi',9:'Tır'}\n",
    "\n",
    "# Resimlerden tahminlere.\n",
    "with tf.Session(config=config) as sess:\n",
    "    saver.restore(sess, \"drive/Colab Notebooks/CIFAR-10/model.ckpt\")\n",
    "    prediction = tf.argmax(y_pred, 1)\n",
    "    index = 0\n",
    "    pred_Val = prediction.eval(feed_dict={x:test_img[index:index+1,:,:,:],train_mode:False})\n",
    "    print(\"Actual Class:\",labels[index])\n",
    "    print(\"Predict Class:\",cifar_classes[pred_Val[0]])\n",
    "    plt.imshow(test_img[index])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "EBw7aWGZlYPM"
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "default_view": {},
   "name": "CIFAR10_Parameter_Test.ipynb",
   "provenance": [],
   "version": "0.3.2",
   "views": {}
  },
  "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": 1
}