CNN_modelling.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "9w1k3_pcht7Z"
   },
   "source": [
    "# Machine Learning Modelling with CNN\n",
    "**In this section taking we are going to code for  the image recognition system using the data prepared from previous section.**\n",
    "\n",
    "\n",
    "****\n",
    "****\n",
    "**IN this section we are going to use advanced CNN architectures**\n",
    "\n",
    "\n",
    "1.   Inception Model(self coded)\n",
    "2.   Fully connected with few convolutional layer\n",
    "\n",
    "\n",
    "****\n",
    "\n",
    "**we have divided our data in 10 batches so as to decrease the computational cost.\n",
    "we will train the model on 9 batches and test on 1 batch so as to confirm that model is working as expected.**\n",
    "**we'll save the model for further reference.**\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "collapsed": true,
    "id": "j2WlmGEBHsQA"
   },
   "outputs": [],
   "source": [
    "#Loading the dataset\n",
    "##loading the dependencies\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "##Upload the Kaggle credentials\n",
    "from google.colab import files\n",
    "files.upload()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 318
    },
    "colab_type": "code",
    "id": "-KqWpFO6H9YN",
    "outputId": "d84ecea9-2e50-4094-ede1-2487e2847376"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: kaggle in /usr/local/lib/python3.6/dist-packages (1.4.7.1)\n",
      "Requirement already satisfied: requests in /usr/local/lib/python3.6/dist-packages (from kaggle) (2.18.4)\n",
      "Requirement already satisfied: python-dateutil in /usr/local/lib/python3.6/dist-packages (from kaggle) (2.5.3)\n",
      "Requirement already satisfied: python-slugify in /usr/local/lib/python3.6/dist-packages (from kaggle) (1.2.6)\n",
      "Requirement already satisfied: tqdm in /usr/local/lib/python3.6/dist-packages (from kaggle) (4.26.0)\n",
      "Requirement already satisfied: urllib3<1.23.0,>=1.15 in /usr/local/lib/python3.6/dist-packages (from kaggle) (1.22)\n",
      "Requirement already satisfied: six>=1.10 in /usr/local/lib/python3.6/dist-packages (from kaggle) (1.11.0)\n",
      "Requirement already satisfied: certifi in /usr/local/lib/python3.6/dist-packages (from kaggle) (2018.8.24)\n",
      "Requirement already satisfied: chardet<3.1.0,>=3.0.2 in /usr/local/lib/python3.6/dist-packages (from requests->kaggle) (3.0.4)\n",
      "Requirement already satisfied: idna<2.7,>=2.5 in /usr/local/lib/python3.6/dist-packages (from requests->kaggle) (2.6)\n",
      "Requirement already satisfied: Unidecode>=0.04.16 in /usr/local/lib/python3.6/dist-packages (from python-slugify->kaggle) (1.0.22)\n",
      "az-handwritten-alphabets-in-csv-format.zip: Skipping, found more recently modified local copy (use --force to force download)\n",
      "Archive:  az-handwritten-alphabets-in-csv-format.zip\n",
      "replace A_Z Handwritten Data.zip? [y]es, [n]o, [A]ll, [N]one, [r]ename: Archive:  A_Z Handwritten Data.zip\n",
      "replace A_Z Handwritten Data.csv? [y]es, [n]o, [A]ll, [N]one, [r]ename: "
     ]
    }
   ],
   "source": [
    "##File configuration\n",
    "!mkdir -p ~/.kaggle\n",
    "!cp kaggle.json ~/.kaggle/\n",
    "\n",
    "!chmod 600 ~/.kaggle/kaggle.json\n",
    "\n",
    "##downloading the dataset\n",
    "!pip install kaggle\n",
    "!kaggle datasets download -d sachinpatel21/az-handwritten-alphabets-in-csv-format\n",
    "\n",
    "##unzipping\n",
    "!unzip az-handwritten-alphabets-in-csv-format.zip\n",
    "##unzippig further\n",
    "!unzip 'A_Z Handwritten Data.zip'\n",
    "\n",
    "##Loading the dataset\n",
    "df = pd.read_csv('A_Z Handwritten Data.csv')\n",
    "\n",
    "## its time to convert the dataframe in to images of size 28X28X1 as it is gray scaled\n",
    "from keras.utils import np_utils\n",
    "X = df.iloc[:,1:].values\n",
    "y = np_utils.to_categorical(df['0'],num_classes = 26)\n",
    "print(y)\n",
    "\n",
    "## resizing \n",
    "X = X.astype('float32')\n",
    "X = X/255.0\n",
    "\n",
    "X = X.reshape(-1,28,28,1)\n",
    "print(X.shape)\n",
    "\n",
    "\n",
    "##creating the batches of the data\n",
    "##We are going to create 10 batches\n",
    "##For inputs (9 training batch and 1 testing batch)\n",
    "batch_1,batch_2,batch_3,batch_4,batch_5,batch_6,batch_7,batch_8,batch_9,b10,b11,b12,b13,b14,b15,b16,b17,b18,b19,b20,b21,b22,b23,b24,b25,b26,b27,b28,b29,x_test = np.split(X,30)\n",
    "\n",
    "##For Outputs\n",
    "y1,y2,y3,y4,y5,y6,y7,y8,y9,y10,y11,y12,y13,y14,y15,y16,y17,y18,y19,y20,y21,y22,y23,y24,y25,y26,y27,y28,y29,y_test = np.split(y,30)\n",
    "\n",
    "del df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "colab_type": "code",
    "id": "YzyKybKfi60_",
    "outputId": "1c30258c-ab33-4128-ec1a-891544edc23d"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'\\nour only dependeny is tensorflow .\\n\\n'"
      ]
     },
     "execution_count": 3,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "##importing the necessary dependencies\n",
    "import tensorflow as tf\n",
    "import keras        # sorry \n",
    "\n",
    "'''\n",
    "our only dependeny is tensorflow .\n",
    "\n",
    "'''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "collapsed": true,
    "id": "GXuQME-Uqd3z"
   },
   "outputs": [],
   "source": [
    "# Remove previous weights, bias, inputs, etc..\n",
    "tf.reset_default_graph()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 111,
     "resources": {
      "http://localhost:8080/nbextensions/google.colab/files.js": {
       "data": "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",
       "headers": [
        [
         "content-type",
         "application/javascript"
        ]
       ],
       "ok": true,
       "status": 200,
       "status_text": ""
      }
     }
    },
    "colab_type": "code",
    "id": "HTCRC8gCj_4P",
    "outputId": "373777f4-9989-4a96-e2d1-444614c6df93"
   },
   "outputs": [],
   "source": [
    "#from google.colab import files\n",
    "#files.upload()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 334
    },
    "colab_type": "code",
    "id": "oODlYB2WkSGF",
    "outputId": "21b294cf-2364-4fd6-c913-eb5a96f74309"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA1IAAAE9CAMAAAA2xaOTAAAABGdBTUEAALGPC/xhBQAAAAFzUkdC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\n+392PxzFb9W6mQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 5,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from IPython.display import Image\n",
    "Image('inception2.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 111,
     "resources": {
      "http://localhost:8080/nbextensions/google.colab/files.js": {
       "data": "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",
       "headers": [
        [
         "content-type",
         "application/javascript"
        ]
       ],
       "ok": true,
       "status": 200,
       "status_text": ""
      }
     }
    },
    "colab_type": "code",
    "id": "HBbrduJUly58",
    "outputId": "821faa17-f915-45f1-8b09-7ea6467020a9"
   },
   "outputs": [],
   "source": [
    "#from google.colab import files\n",
    "#files.upload()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 354
    },
    "colab_type": "code",
    "id": "GrncB_Xal6Kw",
    "outputId": "2f69526e-abe0-433c-8519-f5a861545a1c"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAokAAAFRCAIAAACXMKCpAAAACXBIWXMAAAsTAAALEwEAmpwYAACA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\nAElFTkSuQmCC\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 7,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from IPython.display import Image\n",
    "Image('inception_implement.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "V9blXny3ksVo"
   },
   "source": [
    "# Background\n",
    "**This is the general architecture of Inception model**\n",
    "****\n",
    "**The self coded inception model which we are going to use will follow the same architecture\n",
    "It will be having just 2 layers of inception.**\n",
    "\n",
    "\n",
    "1.   1X1 filters,3X3 filters ,5X5 filters,Avg_pooling layers\n",
    "2.   each subsequent filters are connected by 1X1 filters to reduce the computational cost.\n",
    "\n",
    "****\n",
    "Source Udacity:--\n",
    "**Notice that we get the variety of convolutions that we want; specifically, we will be using 1×1, 3×3, and 5×5 convolutions along with a 3×3 max pooling.  If you’re wondering what the max pooling is doing there with all the other convolutions, we’ve got an answer: pooling is added to the Inception module for no other reason than, historically, good networks having pooling.  The larger convolutions are more computationally expensive, so the paper suggests first doing a 1×1 convolution  reducing the dimensionality of its feature map, passing the resulting feature map through a relu, and then doing the larger convolution (in this case, 5×5 or 3×3). The 1×1 convolution is key because it will be used to reduce the dimensionality of its feature map. **\n",
    "\n",
    "****\n",
    "*If you want more clarity then you must visit deeplearning.ai or anderw ng's deep learning course.*\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "collapsed": true,
    "id": "LUBesGuLklcW"
   },
   "outputs": [],
   "source": [
    "##implementation of Imception Model\n",
    "\n",
    "def Inception_Model(X,keep_prob,second_inception_layer = False):\n",
    "\n",
    "\n",
    "    # one layer of inception\n",
    "    filter1 = tf.Variable(tf.truncated_normal(shape=[1, 1, 1, 96], mean=0, stddev=0.08))     # 1X1 filter having 96 channels\n",
    "    filter2 = tf.Variable(tf.truncated_normal(shape = [3,3,96,96],mean = 0,stddev  =0.08))   # 3X3 filter having 96 channels \n",
    "\n",
    "    in_layer1 = tf.nn.conv2d(X, filter1, strides=[1, 1, 1, 1], padding='SAME')               # same convolution\n",
    "    in_layer2 = tf.nn.conv2d(in_layer1, filter2, strides=[1, 1, 1, 1], padding='SAME')       # same convolution\n",
    "\n",
    "    in_layer2 = tf.nn.relu(in_layer2)                                                        # Relu Activation (Recommended)\n",
    "    inception1 = tf.layers.batch_normalization(in_layer2)                                    # batch normalization for better outcome\n",
    "\n",
    "    # 2nd parallel layer of inception\n",
    "    filter3 = tf.Variable(tf.truncated_normal(shape=[1, 1, 1, 32], mean=0, stddev=0.08))\n",
    "    filter4 = tf.Variable(tf.truncated_normal(shape=[5, 5,32, 32], mean=0, stddev=0.08))     # 5X5 filters having 32 channels\n",
    "\n",
    "    in_layer3 = tf.nn.conv2d(X, filter3, strides=[1, 1, 1, 1], padding='SAME')\n",
    "    in_layer4 = tf.nn.conv2d(in_layer3, filter4, strides=[1, 1, 1, 1], padding='SAME')\n",
    "\n",
    "    in_layer4 = tf.nn.relu(in_layer4)\n",
    "    inception2 = tf.layers.batch_normalization(in_layer4)\n",
    "\n",
    "    inception2 = tf.layers.dropout(inception2,keep_prob)                                     # Dropout used to reduce overfitting\n",
    "\n",
    "    # 3rd parallel layer of inception\n",
    "\n",
    "    filter5 = tf.Variable(tf.truncated_normal(shape=[1, 1, 1, 96], mean=0, stddev=0.08))\n",
    "\n",
    "    in_layer5 = tf.nn.conv2d(X, filter5, strides=[1, 1, 1, 1], padding='SAME')\n",
    "\n",
    "    in_layer5 = tf.nn.relu(in_layer5)\n",
    "    inception3 = tf.layers.batch_normalization(in_layer5)\n",
    "\n",
    "    inception3 = tf.layers.dropout(inception3,keep_prob)\n",
    "\n",
    "    # 4th parallel layer of max_pooling\n",
    "    inception4_pool = tf.nn.max_pool(X, ksize=[1, 3, 3, 1], strides=[1, 1, 1, 1], padding='SAME')  # we can use avg_pooling.\n",
    "    # Applying 1X1 convolution to the pooled version\n",
    "    filter6 = tf.Variable(tf.truncated_normal(shape=[1, 1, 1, 32], mean=0, stddev=0.08))\n",
    "\n",
    "    in_layer6 = tf.nn.conv2d(inception4_pool, filter6, strides=[1, 1, 1, 1], padding='SAME')\n",
    "    # Relu activation\n",
    "    in_layer6 = tf.nn.relu(in_layer6)\n",
    "    # batch normalization\n",
    "    inception4 = tf.layers.batch_normalization(in_layer6)\n",
    "\n",
    "    # Concatenation of the result\n",
    "\n",
    "    inception = keras.layers.concatenate([inception1,inception2,inception3,inception4],axis = 3)\n",
    "    print(inception.shape)\n",
    "\n",
    "\n",
    "    '''\n",
    "    2nd inception layer\n",
    "    \n",
    "    following same architecture\n",
    "    '''\n",
    "    \n",
    "    # Incase we are just ok with the only layer(In that case we need to modify fully_connected function)\n",
    "    \n",
    "    if second_inception_layer == True:\n",
    "\n",
    "        filter21 = tf.Variable(tf.truncated_normal(shape = [1,1,256,512],mean = 0,stddev = 0.08))\n",
    "        filter22 = tf.Variable(tf.truncated_normal(shape = [3,3,512,512],mean = 0,stddev=0.09))\n",
    "\n",
    "        in2_layer1 = tf.nn.conv2d(inception,filter21,strides = [1,1,1,1],padding = 'SAME')\n",
    "        in2_layer2 = tf.nn.conv2d(in2_layer1,filter22,strides = [1,1,1,1],padding = 'SAME')\n",
    "\n",
    "        in2_layer2 = tf.nn.relu(in2_layer2)\n",
    "\n",
    "        inception21 = tf.layers.batch_normalization(in2_layer2)\n",
    "\n",
    "        inception21 = tf.layers.dropout(inception21,keep_prob)\n",
    "\n",
    "        # 2nd parallel layer\n",
    "        filter23 = tf.Variable(tf.truncated_normal(shape = [1,1,256,256],mean = 0,stddev = 0.08))\n",
    "        filter24 = tf.Variable(tf.truncated_normal(shape = [5,5,256,256],mean = 0,stddev= 0.08))\n",
    "\n",
    "        in2_layer3 = tf.nn.conv2d(inception,filter23,strides = [1,1,1,1],padding = 'SAME')\n",
    "        in2_layer4 = tf.nn.conv2d(in2_layer3,filter24,strides = [1,1,1,1],padding = 'SAME')\n",
    "\n",
    "        in2_layer4 =  tf.nn.relu(in2_layer4)\n",
    "\n",
    "        inception22 = tf.layers.batch_normalization(in2_layer4)\n",
    "\n",
    "        inception22 = tf.layers.dropout(inception22,keep_prob)\n",
    "\n",
    "        # 3rd parallel layer\n",
    "        filter25 = tf.Variable(tf.truncated_normal(shape = [1,1,256,256],mean = 0,stddev = 0.08))\n",
    "\n",
    "        in2_layer5 = tf.nn.conv2d(inception,filter25, strides  =[1,1,1,1],padding = 'SAME')\n",
    "\n",
    "        in2_layer5 = tf.nn.relu(in2_layer5)\n",
    "\n",
    "        inception23 = tf.layers.batch_normalization(in2_layer5)\n",
    "\n",
    "        inception23 = tf.layers.dropout(inception23,keep_prob)\n",
    "\n",
    "        # 4th parallel avg pooling layer\n",
    "        inception24_pool = tf.nn.max_pool(inception, ksize=[1, 3, 3, 1], strides=[1, 1, 1, 1], padding='SAME')\n",
    "        # Applying 1X1 convolution to the pooled version\n",
    "        filter26 = tf.Variable(tf.truncated_normal(shape=[1, 1, 256, 256], mean=0, stddev=0.08))\n",
    "\n",
    "        in2_layer6 = tf.nn.conv2d(inception24_pool, filter26, strides=[1, 1, 1, 1], padding='SAME')\n",
    "        # Relu activation\n",
    "        in2_layer6 = tf.nn.relu(in2_layer6)\n",
    "        # batch normalization\n",
    "        inception24 = tf.layers.batch_normalization(in2_layer6)\n",
    "\n",
    "\n",
    "        # concatenation of the layers\n",
    "        inception_layer = keras.layers.concatenate([inception21,inception22,inception23,inception24],axis = 3)\n",
    "\n",
    "        print(inception_layer.shape)\n",
    "\n",
    "        output = fully_connected(inception_layer,keep_prob)\n",
    "        \n",
    "        return output\n",
    "        \n",
    "\n",
    "    '''\n",
    "    our dimension is 28X28X1280 \n",
    "    and our original dataset has dimensions of 28X28X1\n",
    "    '''\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "collapsed": true,
    "id": "s7Rj9BgXnpxG"
   },
   "outputs": [],
   "source": [
    "##Implementation of fully connected layers \n",
    "def fully_connected(inception_layers,keep_prob):\n",
    "\n",
    "    # Moving with general convolutions\n",
    "    gen_filter1 = tf.Variable(tf.truncated_normal(shape = [3,3,1280,1024],mean = 0,stddev=0.08))\n",
    "    gen_filter2 = tf.Variable(tf.truncated_normal(shape = [3,3,1024,512],mean = 0,stddev=0.08))\n",
    "    gen_filter3 = tf.Variable(tf.truncated_normal(shape=[3, 3,512,256], mean=0, stddev=0.08))\n",
    "\n",
    "    # general convolution layer -1\n",
    "    gen_conv1 = tf.nn.conv2d(inception_layers,gen_filter1,strides = [1,1,1,1],padding = 'SAME')\n",
    "    gen_conv1 = tf.nn.relu(gen_conv1)\n",
    "    pool_conv1 = tf.nn.max_pool(gen_conv1,ksize = [1,2,2,1],strides = [1,2,2,1],padding  = 'SAME')\n",
    "    conv1 = tf.layers.batch_normalization(pool_conv1)\n",
    "\n",
    "    # general convolution layer-2\n",
    "    gen_conv2 = tf.nn.conv2d(conv1, gen_filter2, strides=[1, 1, 1, 1], padding='SAME')\n",
    "    gen_conv2 = tf.nn.relu(gen_conv2)\n",
    "    pool_conv2 = tf.nn.max_pool(gen_conv2, ksize=[1, 3, 3, 1], strides=[1, 2, 2, 1], padding='SAME')\n",
    "    conv2 = tf.layers.batch_normalization(pool_conv2)\n",
    "    print(conv2.shape)\n",
    "\n",
    "    # general convolution layer-3\n",
    "    gen_conv3 = tf.nn.conv2d(conv2, gen_filter3, strides=[1, 1, 1, 1], padding='SAME')\n",
    "    gen_conv3 = tf.nn.relu(gen_conv3)\n",
    "    pool_conv3 = tf.nn.max_pool(gen_conv3, ksize=[1, 3, 3, 1], strides=[1, 2, 2, 1], padding='SAME')\n",
    "    conv3 = tf.layers.batch_normalization(pool_conv3)\n",
    "    print(conv3.shape)\n",
    "\n",
    "\n",
    "    # fully connected layers\n",
    "    flat =  tf.contrib.layers.flatten(conv3)\n",
    "    print(flat.shape)\n",
    "\n",
    "    # connection 1\n",
    "    full1 = tf.contrib.layers.fully_connected(inputs = flat,\n",
    "                                              num_outputs = 1024,\n",
    "                                              activation_fn = tf.nn.relu)\n",
    "    full1 = tf.nn.dropout(full1,keep_prob = keep_prob)\n",
    "    full1 = tf.layers.batch_normalization(full1)\n",
    "\n",
    "    # conncetion 2\n",
    "    full2 = tf.contrib.layers.fully_connected(inputs=full1,\n",
    "                                              num_outputs=512,\n",
    "                                              activation_fn=tf.nn.relu)\n",
    "    full2 = tf.nn.dropout(full2, keep_prob=keep_prob)\n",
    "    full2 = tf.layers.batch_normalization(full2)\n",
    "\n",
    "    # connection 3\n",
    "    full3 = tf.contrib.layers.fully_connected(inputs=full2,\n",
    "                                              num_outputs=128,\n",
    "                                              activation_fn=tf.nn.relu)\n",
    "    full3 = tf.nn.dropout(full3, keep_prob=keep_prob)\n",
    "    full3 = tf.layers.batch_normalization(full3)\n",
    "\n",
    "    # output layer\n",
    "    output = tf.contrib.layers.fully_connected(inputs  = full3,\n",
    "                                               num_outputs =  26,\n",
    "                                               activation_fn = tf.nn.softmax)\n",
    "\n",
    "    return output\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 107
    },
    "colab_type": "code",
    "id": "T5Wvjbfhn-o9",
    "outputId": "0d2723d7-6e0b-4dc8-df9a-c44d041d110a"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(?, 28, 28, 256)\n",
      "(?, 28, 28, 1280)\n",
      "(?, 7, 7, 512)\n",
      "(?, 4, 4, 256)\n",
      "(?, 4096)\n"
     ]
    }
   ],
   "source": [
    "## Hyperparameters\n",
    "\n",
    "epochs = 10\n",
    "keep_probability = 0.7\n",
    "learning_rate = 0.01\n",
    "\n",
    "## cost function and optimizers\n",
    "logits = Inception_Model(x,keep_prob = keep_probability,second_inception_layer = True)\n",
    "cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits_v2(logits = logits,labels = y))\n",
    "optimizer = tf.train.AdamOptimizer(learning_rate = learning_rate).minimize(cost)\n",
    "\n",
    "##Accuracy\n",
    "correct_pred = tf.equal(tf.argmax(logits, 1), tf.argmax(y, 1))\n",
    "accuracy = tf.reduce_mean(tf.cast(correct_pred, tf.float32), name='accuracy')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "collapsed": true,
    "id": "R3f9YWKIJAnH"
   },
   "outputs": [],
   "source": [
    "x_batch = [batch_1,batch_2,batch_3,batch_4,batch_5,batch_6,batch_7,batch_8,batch_9,b10,b11,b12,b13,b14,b15,b16,b17,b18,b19,b20,b21,b22,b23,b24,b25,b26,b27,b28,b29,]\n",
    "y_batch = [y1,y2,y3,y4,y5,y6,y7,y8,y9,y10,y11,y12,y13,y14,y15,y16,y17,y18,y19,y20,y21,y22,y23,y24,y25,y26,y27,y28,y29]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "collapsed": true,
    "id": "7Q3Zxk8nqoF2"
   },
   "outputs": [],
   "source": [
    "def train_neural_network(session, optimizer, keep_probability, feature_batch, label_batch):\n",
    "    session.run(optimizer, \n",
    "                feed_dict={\n",
    "                    x: feature_batch,\n",
    "                    y: label_batch,\n",
    "                    keep_prob: keep_probability\n",
    "                })\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "collapsed": true,
    "id": "AuQQ8nN0ufFF"
   },
   "outputs": [],
   "source": [
    "def print_stats(session, feature_batch, label_batch, cost, accuracy):\n",
    "    loss = sess.run(cost, \n",
    "                    feed_dict={\n",
    "                        x: feature_batch,\n",
    "                        y: label_batch,\n",
    "                        keep_prob: 1.\n",
    "                    })\n",
    "    valid_acc = sess.run(accuracy, \n",
    "                         feed_dict={\n",
    "                             x: valid_features,\n",
    "                             y: valid_labels,\n",
    "                             keep_prob: 1.\n",
    "                         })\n",
    "    \n",
    "    print('Loss: {:>10.4f} Validation Accuracy: {:.6f}'.format(loss, valid_acc))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "collapsed": true,
    "id": "WUBY6kPQq6FF"
   },
   "outputs": [],
   "source": [
    "with tf.Graph().as_default():\n",
    "  # Inputs\n",
    "  x = tf.placeholder(tf.float32, shape=(None, 28,28,1), name='input_x')\n",
    "  y =  tf.placeholder(tf.float32, shape=(None, 26), name='output_y')\n",
    "  keep_prob = tf.placeholder(tf.float32, name='keep_prob') \n",
    "  ## cost function and optimizers\n",
    "  logits = Inception_Model(x,keep_prob = keep_probability,second_inception_layer = True)\n",
    "  cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits_v2(logits = logits,labels = y))\n",
    "  optimizer = tf.train.AdamOptimizer(learning_rate = learning_rate).minimize(cost)\n",
    "\n",
    "  ##Accuracy\n",
    "  correct_pred = tf.equal(tf.argmax(logits, 1), tf.argmax(y, 1))\n",
    "  accuracy = tf.reduce_mean(tf.cast(correct_pred, tf.float32), name='accuracy')\n",
    "  init = tf.initialize_all_variables()\n",
    "\n",
    "  with tf.Session() as sess:\n",
    "      # Initializing the variables\n",
    "      sess.run(init)\n",
    "\n",
    "      # Training cycle\n",
    "      for epoch in range(epochs):\n",
    "          # Loop over all batches\n",
    "          #n_batches = 9\n",
    "          for batch_features, batch_labels in zip(x_batch,y_batch):\n",
    "\n",
    "            train_neural_network(sess, optimizer, keep_probability, batch_features, batch_labels)\n",
    "\n",
    "            print('Epoch {:>2}, data batch {}:  '.format(epoch + 1, batch_i), end='')\n",
    "            print_stats(sess, batch_features, batch_labels, cost, accuracy)\n",
    "\n",
    "            "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "collapsed": true,
    "id": "kIVlHzhcrOVa"
   },
   "outputs": [],
   "source": [
    "    \n",
    "  "
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "name": "CNN modelling",
   "provenance": [],
   "version": "0.3.2"
  },
  "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": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}