kaggle_minst_cnn.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "kaggle-minst-cnn.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "[View in Colaboratory](https://colab.research.google.com/github/kbalu99/kaggle-minst-cnn/blob/master/kaggle_minst_cnn.ipynb)"
      ]
    },
    {
      "metadata": {
        "id": "06SBeAYrGdIP",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 448
        },
        "outputId": "2108e32a-506c-4ba5-9091-6b99b1686fc9"
      },
      "cell_type": "code",
      "source": [
        "# To import kaggle datasets\n",
        "!pip install kaggle"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting kaggle\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/c6/78/832b9a9ec6b3baf8ec566e1f0a695f2fd08d2c94a6797257a106304bfc3c/kaggle-1.4.7.1.tar.gz (52kB)\n",
            "\u001b[K    100% |████████████████████████████████| 61kB 3.9MB/s \n",
            "\u001b[?25hRequirement 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: python-dateutil in /usr/local/lib/python3.6/dist-packages (from kaggle) (2.5.3)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.6/dist-packages (from kaggle) (2.18.4)\n",
            "Requirement already satisfied: tqdm in /usr/local/lib/python3.6/dist-packages (from kaggle) (4.25.0)\n",
            "Collecting python-slugify (from kaggle)\n",
            "  Downloading https://files.pythonhosted.org/packages/00/ad/c778a6df614b6217c30fe80045b365bfa08b5dd3cb02e8b37a6d25126781/python-slugify-1.2.6.tar.gz\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: chardet<3.1.0,>=3.0.2 in /usr/local/lib/python3.6/dist-packages (from requests->kaggle) (3.0.4)\n",
            "Collecting Unidecode>=0.04.16 (from python-slugify->kaggle)\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/59/ef/67085e30e8bbcdd76e2f0a4ad8151c13a2c5bce77c85f8cad6e1f16fb141/Unidecode-1.0.22-py2.py3-none-any.whl (235kB)\n",
            "\u001b[K    100% |████████████████████████████████| 235kB 7.7MB/s \n",
            "\u001b[?25hBuilding wheels for collected packages: kaggle, python-slugify\n",
            "  Running setup.py bdist_wheel for kaggle ... \u001b[?25l-\b \b\\\b \bdone\n",
            "\u001b[?25h  Stored in directory: /root/.cache/pip/wheels/44/2c/df/22a6eeb780c36c28190faef6252b739fdc47145fd87a6642d4\n",
            "  Running setup.py bdist_wheel for python-slugify ... \u001b[?25l-\b \bdone\n",
            "\u001b[?25h  Stored in directory: /root/.cache/pip/wheels/e3/65/da/2045deea3098ed7471eca0e2460cfbd3fdfe8c1d6fa6fcac92\n",
            "Successfully built kaggle python-slugify\n",
            "Installing collected packages: Unidecode, python-slugify, kaggle\n",
            "Successfully installed Unidecode-1.0.22 kaggle-1.4.7.1 python-slugify-1.2.6\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "78l574d0FA0w",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Importing libraries\n",
        "import keras\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "import matplotlib.image as mpimg\n",
        "import seaborn as sns\n",
        "\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.metrics import confusion_matrix\n",
        "import itertools\n",
        "from keras.utils.np_utils import to_categorical \n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense, Dropout, Flatten, Conv2D, MaxPool2D\n",
        "from keras.optimizers import RMSprop\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras.callbacks import ReduceLROnPlateau, ModelCheckpoint"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "22X1sX4kGf8J",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# import colab libraries\n",
        "from google.colab import files\n",
        "\n",
        "# import kaggle json to connect to kaggle user account to download datsets\n",
        "files.upload()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "yEUASbROHUT4",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "9a8e0289-6a8b-440e-ff4f-a9d2ba43bccb"
      },
      "cell_type": "code",
      "source": [
        "# see if kaggle json exists\n",
        "!ls -lha kaggle.json"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "-rw-r--r-- 1 root root 63 Sep  8 02:57 kaggle.json\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "MISUOTzJHi5t",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# The Kaggle API client expects this file to be in ~/.kaggle,\n",
        "# so lets move it there.\n",
        "!mkdir -p ~/.kaggle\n",
        "!cp kaggle.json ~/.kaggle/"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "Rp6tthP-Hk8B",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# This permissions change avoids a warning on Kaggle tool startup.\n",
        "!chmod 600 ~/.kaggle/kaggle.json"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "clDjWrxWHpIe",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 171
        },
        "outputId": "a3609901-9534-4de8-c2cc-a0b1243164f2"
      },
      "cell_type": "code",
      "source": [
        "#lets now download our dataset\n",
        "!kaggle competitions download -c digit-recognizer"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading train.csv to /content\n",
            " 70% 51.0M/73.2M [00:00<00:00, 67.6MB/s]\n",
            "100% 73.2M/73.2M [00:00<00:00, 158MB/s] \n",
            "Downloading test.csv to /content\n",
            " 76% 37.0M/48.8M [00:00<00:00, 68.6MB/s]\n",
            "100% 48.8M/48.8M [00:00<00:00, 148MB/s] \n",
            "Downloading sample_submission.csv to /content\n",
            "  0% 0.00/235k [00:00<?, ?B/s]\n",
            "100% 235k/235k [00:00<00:00, 60.2MB/s]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "Zn7x_RFzH6St",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "X_train = pd.read_csv('train.csv')\n",
        "X_test = pd.read_csv('test.csv')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "BpVIDSDFICfX",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# separate out the labels \n",
        "Y_train = X_train['label']"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "M4tsonQiIPNi",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "X_train = X_train.drop(labels = ['label'], axis = 1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "THkE6RjIIUv9",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "eac5ed97-fbef-437f-a987-ea6aaf65a650"
      },
      "cell_type": "code",
      "source": [
        "# check the distribution of label values\n",
        "Y_train.value_counts()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "1    4684\n",
              "7    4401\n",
              "3    4351\n",
              "9    4188\n",
              "2    4177\n",
              "6    4137\n",
              "0    4132\n",
              "4    4072\n",
              "8    4063\n",
              "5    3795\n",
              "Name: label, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 20
        }
      ]
    },
    {
      "metadata": {
        "id": "TlMrfqCIIXHj",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 103
        },
        "outputId": "fe674696-7d21-4001-ba50-5e9c5559388a"
      },
      "cell_type": "code",
      "source": [
        "X_train.isnull().any().describe()\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "count       784\n",
              "unique        1\n",
              "top       False\n",
              "freq        784\n",
              "dtype: object"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 21
        }
      ]
    },
    {
      "metadata": {
        "id": "48yjbXXJIbe_",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 103
        },
        "outputId": "702b5658-4f09-42f9-86b8-ab3921337dcd"
      },
      "cell_type": "code",
      "source": [
        "X_test.isnull().any().describe()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "count       784\n",
              "unique        1\n",
              "top       False\n",
              "freq        784\n",
              "dtype: object"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 22
        }
      ]
    },
    {
      "metadata": {
        "id": "J4Q0KyexIhHp",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "X_train = X_train /255.0\n",
        "X_test = X_test / 255.0"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "mh4RSEvTIlml",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# reshape the labels to fit the input format for our CNN\n",
        "X_train = X_train.values.reshape(-1,28,28,1)\n",
        "X_test = X_test.values.reshape(-1,28,28,1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "axB-VijkImu6",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# label encoding\n",
        "Y_train = to_categorical(Y_train, num_classes = 10)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "K0UhDTVFIr2z",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Set the random seed\n",
        "random_seed = 2"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "RgEdgpobIuXk",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Split the train and the validation set for the fitting\n",
        "X_train, X_val, Y_train, Y_val = train_test_split(X_train, Y_train, test_size = 0.1, random_state=random_seed)\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "RpfhMmGBIyYe",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Define CNN model\n",
        "model = Sequential()\n",
        "model.add(Conv2D(filters = 32, kernel_size = (5,5),padding = 'Same', \n",
        "                 activation ='relu', input_shape = (28,28,1)))\n",
        "model.add(Conv2D(filters = 32, kernel_size = (5,5),padding = 'Same', \n",
        "                 activation ='relu'))\n",
        "model.add(MaxPool2D(pool_size=(2,2)))\n",
        "model.add(Dropout(0.25))\n",
        "model.add(Conv2D(filters = 64, kernel_size = (3,3),padding = 'Same', \n",
        "                 activation ='relu'))\n",
        "model.add(Conv2D(filters = 64, kernel_size = (3,3),padding = 'Same', \n",
        "                 activation ='relu'))\n",
        "model.add(MaxPool2D(pool_size=(2,2), strides=(2,2)))\n",
        "model.add(Dropout(0.25))\n",
        "model.add(Flatten())\n",
        "model.add(Dense(256, activation = \"relu\"))\n",
        "model.add(Dropout(0.5))\n",
        "model.add(Dense(10, activation = \"softmax\"))\n",
        "\n",
        "# Define the optimizer\n",
        "optimizer = RMSprop(lr=0.001, rho=0.9, epsilon=1e-08, decay=0.0)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "rOtiSd-DJCYd",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "\n",
        "# Compile the model\n",
        "model.compile(optimizer = optimizer , loss = \"categorical_crossentropy\", metrics=[\"accuracy\"])\n",
        "\n",
        "# Set a learning rate annealer and model checkpoint\n",
        "learning_rate_reduction = [ReduceLROnPlateau(monitor='val_acc', \n",
        "                                            patience=3, \n",
        "                                            verbose=1, \n",
        "                                            factor=0.5, \n",
        "                                            min_lr=0.00001),\n",
        "                           ModelCheckpoint('model-cnn-minst.h5', verbose=1, save_best_only=True, save_weights_only=True)]\n",
        "\n",
        "\n",
        "epochs = 25 # Turn epochs to 30 to get 0.9967 accuracy\n",
        "batch_size = 86"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "ZpfYc-9UJRLW",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 548
        },
        "outputId": "63d0ce06-b81b-4086-a987-451e68691245"
      },
      "cell_type": "code",
      "source": [
        "# print model summary\n",
        "model.summary()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d_1 (Conv2D)            (None, 28, 28, 32)        832       \n",
            "_________________________________________________________________\n",
            "conv2d_2 (Conv2D)            (None, 28, 28, 32)        25632     \n",
            "_________________________________________________________________\n",
            "max_pooling2d_1 (MaxPooling2 (None, 14, 14, 32)        0         \n",
            "_________________________________________________________________\n",
            "dropout_1 (Dropout)          (None, 14, 14, 32)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_3 (Conv2D)            (None, 14, 14, 64)        18496     \n",
            "_________________________________________________________________\n",
            "conv2d_4 (Conv2D)            (None, 14, 14, 64)        36928     \n",
            "_________________________________________________________________\n",
            "max_pooling2d_2 (MaxPooling2 (None, 7, 7, 64)          0         \n",
            "_________________________________________________________________\n",
            "dropout_2 (Dropout)          (None, 7, 7, 64)          0         \n",
            "_________________________________________________________________\n",
            "flatten_1 (Flatten)          (None, 3136)              0         \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 256)               803072    \n",
            "_________________________________________________________________\n",
            "dropout_3 (Dropout)          (None, 256)               0         \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 10)                2570      \n",
            "=================================================================\n",
            "Total params: 887,530\n",
            "Trainable params: 887,530\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "GjlKOFk4JYgz",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "\n",
        "# With data augmentation to prevent overfitting (accuracy 0.99286)\n",
        "\n",
        "datagen = ImageDataGenerator(\n",
        "        featurewise_center=False,  # set input mean to 0 over the dataset\n",
        "        samplewise_center=False,  # set each sample mean to 0\n",
        "        featurewise_std_normalization=False,  # divide inputs by std of the dataset\n",
        "        samplewise_std_normalization=False,  # divide each input by its std\n",
        "        zca_whitening=False,  # apply ZCA whitening\n",
        "        rotation_range=10,  # randomly rotate images in the range (degrees, 0 to 180)\n",
        "        zoom_range = 0.1, # Randomly zoom image \n",
        "        width_shift_range=0.1,  # randomly shift images horizontally (fraction of total width)\n",
        "        height_shift_range=0.1,  # randomly shift images vertically (fraction of total height)\n",
        "        horizontal_flip=False,  # randomly flip images\n",
        "        vertical_flip=False)  # randomly flip images\n",
        "\n",
        "\n",
        "datagen.fit(X_train)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "6EtLw5TWJcMT",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1868
        },
        "outputId": "463a5ef4-e39c-418f-a301-bd72cec81bce"
      },
      "cell_type": "code",
      "source": [
        "# Fit the model\n",
        "history = model.fit_generator(datagen.flow(X_train,Y_train, batch_size=batch_size),\n",
        "                              epochs = epochs, validation_data = (X_val,Y_val),\n",
        "                              verbose = 2, steps_per_epoch=X_train.shape[0] // batch_size\n",
        "                              , callbacks=learning_rate_reduction)\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/25\n",
            " - 15s - loss: 0.4124 - acc: 0.8681 - val_loss: 0.1411 - val_acc: 0.9555\n",
            "\n",
            "Epoch 00001: val_loss improved from inf to 0.14107, saving model to model-cnn-minst.h5\n",
            "Epoch 2/25\n",
            " - 14s - loss: 0.1246 - acc: 0.9639 - val_loss: 0.0369 - val_acc: 0.9883\n",
            "\n",
            "Epoch 00002: val_loss improved from 0.14107 to 0.03691, saving model to model-cnn-minst.h5\n",
            "Epoch 3/25\n",
            " - 14s - loss: 0.0953 - acc: 0.9719 - val_loss: 0.0464 - val_acc: 0.9876\n",
            "\n",
            "Epoch 00003: val_loss did not improve from 0.03691\n",
            "Epoch 4/25\n",
            " - 14s - loss: 0.0770 - acc: 0.9776 - val_loss: 0.0272 - val_acc: 0.9929\n",
            "\n",
            "Epoch 00004: val_loss improved from 0.03691 to 0.02724, saving model to model-cnn-minst.h5\n",
            "Epoch 5/25\n",
            " - 14s - loss: 0.0719 - acc: 0.9789 - val_loss: 0.0314 - val_acc: 0.9921\n",
            "\n",
            "Epoch 00005: val_loss did not improve from 0.02724\n",
            "Epoch 6/25\n",
            " - 14s - loss: 0.0678 - acc: 0.9814 - val_loss: 0.0378 - val_acc: 0.9902\n",
            "\n",
            "Epoch 00006: val_loss did not improve from 0.02724\n",
            "Epoch 7/25\n",
            " - 14s - loss: 0.0647 - acc: 0.9808 - val_loss: 0.0313 - val_acc: 0.9919\n",
            "\n",
            "Epoch 00007: ReduceLROnPlateau reducing learning rate to 0.0005000000237487257.\n",
            "\n",
            "Epoch 00007: val_loss did not improve from 0.02724\n",
            "Epoch 8/25\n",
            " - 14s - loss: 0.0488 - acc: 0.9862 - val_loss: 0.0292 - val_acc: 0.9924\n",
            "\n",
            "Epoch 00008: val_loss did not improve from 0.02724\n",
            "Epoch 9/25\n",
            " - 14s - loss: 0.0475 - acc: 0.9864 - val_loss: 0.0209 - val_acc: 0.9929\n",
            "\n",
            "Epoch 00009: val_loss improved from 0.02724 to 0.02089, saving model to model-cnn-minst.h5\n",
            "Epoch 10/25\n",
            " - 13s - loss: 0.0439 - acc: 0.9870 - val_loss: 0.0243 - val_acc: 0.9938\n",
            "\n",
            "Epoch 00010: val_loss did not improve from 0.02089\n",
            "Epoch 11/25\n",
            " - 14s - loss: 0.0453 - acc: 0.9870 - val_loss: 0.0180 - val_acc: 0.9940\n",
            "\n",
            "Epoch 00011: val_loss improved from 0.02089 to 0.01799, saving model to model-cnn-minst.h5\n",
            "Epoch 12/25\n",
            " - 14s - loss: 0.0465 - acc: 0.9870 - val_loss: 0.0269 - val_acc: 0.9919\n",
            "\n",
            "Epoch 00012: val_loss did not improve from 0.01799\n",
            "Epoch 13/25\n",
            " - 13s - loss: 0.0432 - acc: 0.9876 - val_loss: 0.0230 - val_acc: 0.9936\n",
            "\n",
            "Epoch 00013: val_loss did not improve from 0.01799\n",
            "Epoch 14/25\n",
            " - 13s - loss: 0.0452 - acc: 0.9868 - val_loss: 0.0196 - val_acc: 0.9952\n",
            "\n",
            "Epoch 00014: val_loss did not improve from 0.01799\n",
            "Epoch 15/25\n",
            " - 14s - loss: 0.0459 - acc: 0.9876 - val_loss: 0.0198 - val_acc: 0.9938\n",
            "\n",
            "Epoch 00015: val_loss did not improve from 0.01799\n",
            "Epoch 16/25\n",
            " - 14s - loss: 0.0472 - acc: 0.9870 - val_loss: 0.0299 - val_acc: 0.9914\n",
            "\n",
            "Epoch 00016: val_loss did not improve from 0.01799\n",
            "Epoch 17/25\n",
            " - 14s - loss: 0.0476 - acc: 0.9865 - val_loss: 0.0220 - val_acc: 0.9940\n",
            "\n",
            "Epoch 00017: ReduceLROnPlateau reducing learning rate to 0.0002500000118743628.\n",
            "\n",
            "Epoch 00017: val_loss did not improve from 0.01799\n",
            "Epoch 18/25\n",
            " - 14s - loss: 0.0371 - acc: 0.9892 - val_loss: 0.0182 - val_acc: 0.9950\n",
            "\n",
            "Epoch 00018: val_loss did not improve from 0.01799\n",
            "Epoch 19/25\n",
            " - 14s - loss: 0.0355 - acc: 0.9895 - val_loss: 0.0201 - val_acc: 0.9931\n",
            "\n",
            "Epoch 00019: val_loss did not improve from 0.01799\n",
            "Epoch 20/25\n",
            " - 14s - loss: 0.0358 - acc: 0.9896 - val_loss: 0.0200 - val_acc: 0.9952\n",
            "\n",
            "Epoch 00020: ReduceLROnPlateau reducing learning rate to 0.0001250000059371814.\n",
            "\n",
            "Epoch 00020: val_loss did not improve from 0.01799\n",
            "Epoch 21/25\n",
            " - 15s - loss: 0.0329 - acc: 0.9902 - val_loss: 0.0221 - val_acc: 0.9950\n",
            "\n",
            "Epoch 00021: val_loss did not improve from 0.01799\n",
            "Epoch 22/25\n",
            " - 14s - loss: 0.0330 - acc: 0.9902 - val_loss: 0.0168 - val_acc: 0.9955\n",
            "\n",
            "Epoch 00022: val_loss improved from 0.01799 to 0.01678, saving model to model-cnn-minst.h5\n",
            "Epoch 23/25\n",
            " - 13s - loss: 0.0316 - acc: 0.9910 - val_loss: 0.0176 - val_acc: 0.9948\n",
            "\n",
            "Epoch 00023: val_loss did not improve from 0.01678\n",
            "Epoch 24/25\n",
            " - 14s - loss: 0.0315 - acc: 0.9909 - val_loss: 0.0205 - val_acc: 0.9945\n",
            "\n",
            "Epoch 00024: val_loss did not improve from 0.01678\n",
            "Epoch 25/25\n",
            " - 13s - loss: 0.0317 - acc: 0.9910 - val_loss: 0.0174 - val_acc: 0.9948\n",
            "\n",
            "Epoch 00025: ReduceLROnPlateau reducing learning rate to 6.25000029685907e-05.\n",
            "\n",
            "Epoch 00025: val_loss did not improve from 0.01678\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "H6b4yN9YML_e",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# function to plot confusion matrix\n",
        "def plot_confusion_matrix(cm, classes,\n",
        "                          normalize=False,\n",
        "                          title='Confusion matrix',\n",
        "                          cmap=plt.cm.Blues):\n",
        "    \"\"\"\n",
        "    This function prints and plots the confusion matrix.\n",
        "    Normalization can be applied by setting `normalize=True`.\n",
        "    \"\"\"\n",
        "    plt.imshow(cm, interpolation='nearest', cmap=cmap)\n",
        "    plt.title(title)\n",
        "    plt.colorbar()\n",
        "    tick_marks = np.arange(len(classes))\n",
        "    plt.xticks(tick_marks, classes, rotation=45)\n",
        "    plt.yticks(tick_marks, classes)\n",
        "\n",
        "    if normalize:\n",
        "        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
        "\n",
        "    thresh = cm.max() / 2.\n",
        "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
        "        plt.text(j, i, cm[i, j],\n",
        "                 horizontalalignment=\"center\",\n",
        "                 color=\"white\" if cm[i, j] > thresh else \"black\")\n",
        "\n",
        "    plt.tight_layout()\n",
        "    plt.ylabel('True label')\n",
        "    plt.xlabel('Predicted label')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "wI_NzhPVOXlw",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 415
        },
        "outputId": "d6a25ac8-dd91-4edf-887c-3599e585c67c"
      },
      "cell_type": "code",
      "source": [
        "# Predict the values from the validation dataset\n",
        "Y_pred = model.predict(X_val)\n",
        "# Convert predictions classes to one hot vectors \n",
        "Y_pred_classes = np.argmax(Y_pred,axis = 1) \n",
        "# Convert validation observations to one hot vectors\n",
        "Y_true = np.argmax(Y_val,axis = 1) \n",
        "# compute the confusion matrix\n",
        "confusion_mtx = confusion_matrix(Y_true, Y_pred_classes) \n",
        "# plot the confusion matrix\n",
        "plot_confusion_matrix(confusion_mtx, classes = range(10)) "
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAb4AAAGOCAYAAADo92ZEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nhU8IIYRNkcInhBDCpkjhE0IIYVP+D4AQ/xIqEn80AAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7fea238e2390>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "GhLzeLM8OlGX",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 347
        },
        "outputId": "a1eab442-67f7-4280-eff8-24e302d60d5e"
      },
      "cell_type": "code",
      "source": [
        "\n",
        "\n",
        "\n",
        "# Plot the loss and accuracy curves for training and validation \n",
        "fig, ax = plt.subplots(2,1)\n",
        "ax[0].plot(history.history['loss'], color='b', label=\"Training loss\")\n",
        "ax[0].plot(history.history['val_loss'], color='r', label=\"validation loss\",axes =ax[0])\n",
        "legend = ax[0].legend(loc='best', shadow=True)\n",
        "\n",
        "ax[1].plot(history.history['acc'], color='b', label=\"Training accuracy\")\n",
        "ax[1].plot(history.history['val_acc'], color='r',label=\"Validation accuracy\")\n",
        "legend = ax[1].legend(loc='best', shadow=True)\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAe8AAAFKCAYAAAA0WNeQAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAElFTkSuQmCC\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7feaa8078588>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "psDRlNhfOsJy",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "\n",
        "# Display some error results \n",
        "\n",
        "# Errors are difference between predicted labels and true labels\n",
        "errors = (Y_pred_classes - Y_true != 0)\n",
        "\n",
        "Y_pred_classes_errors = Y_pred_classes[errors]\n",
        "Y_pred_errors = Y_pred[errors]\n",
        "Y_true_errors = Y_true[errors]\n",
        "X_val_errors = X_val[errors]\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "85YL_uctOyuO",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "\n",
        "def display_errors(errors_index,img_errors,pred_errors, obs_errors):\n",
        "    \"\"\" This function shows 6 images with their predicted and real labels\"\"\"\n",
        "    n = 0\n",
        "    nrows = 2\n",
        "    ncols = 3\n",
        "    fig, ax = plt.subplots(nrows,ncols,sharex=True,sharey=True)\n",
        "    for row in range(nrows):\n",
        "        for col in range(ncols):\n",
        "            error = errors_index[n]\n",
        "            ax[row,col].imshow((img_errors[error]).reshape((28,28)))\n",
        "            ax[row,col].set_title(\"Predicted label :{}\\nTrue label :{}\".format(pred_errors[error],obs_errors[error]))\n",
        "            n += 1"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "B7gh838wO5dd",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 376
        },
        "outputId": "3aaec465-ec74-4a95-cdcb-330e097144a5"
      },
      "cell_type": "code",
      "source": [
        "\n",
        "# Probabilities of the wrong predicted numbers\n",
        "Y_pred_errors_prob = np.max(Y_pred_errors,axis = 1)\n",
        "\n",
        "# Predicted probabilities of the true values in the error set\n",
        "true_prob_errors = np.diagonal(np.take(Y_pred_errors, Y_true_errors, axis=1))\n",
        "\n",
        "# Difference between the probability of the predicted label and the true label\n",
        "delta_pred_true_errors = Y_pred_errors_prob - true_prob_errors\n",
        "\n",
        "# Sorted list of the delta prob errors\n",
        "sorted_dela_errors = np.argsort(delta_pred_true_errors)\n",
        "\n",
        "# Top 6 errors \n",
        "most_important_errors = sorted_dela_errors[-6:]\n",
        "\n",
        "# Show the top 6 errors\n",
        "display_errors(most_important_errors, X_val_errors, Y_pred_classes_errors, Y_true_errors)\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAd8AAAFnCAYAAADnmUNAAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7feaa8078eb8>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "dFR7PTn-O-Xr",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "\n",
        "# predict results\n",
        "results = model.predict(X_test)\n",
        "\n",
        "# select the indix with the maximum probability\n",
        "results = np.argmax(results,axis = 1)\n",
        "\n",
        "results = pd.Series(results,name=\"Label\")\n",
        "\n",
        "\n",
        "submission = pd.concat([pd.Series(range(1,28001),name = \"ImageId\"),results],axis = 1)\n",
        "\n",
        "submission.to_csv(\"cnn_mnist_datagen.csv\",index=False)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "hgyF5YtFPGLp",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        },
        "outputId": "4c2c9c91-60b8-42ac-89f1-903b6479a2a0"
      },
      "cell_type": "code",
      "source": [
        "!ls"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "cnn_mnist_datagen.csv  model-cnn-minst.h5  sample_submission.csv  train.csv\n",
            "kaggle.json\t       sample_data\t   test.csv\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "yB0ZJHnTPL1h",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 774
        },
        "outputId": "541abd5e-275a-4046-aa68-1defda4aa97a"
      },
      "cell_type": "code",
      "source": [
        "# Save final model to convert and use later\n",
        "model.save('new.h5')\n",
        "\n",
        "# Now let's convert the model to JSON using tensorflowjs and use in the angular web app\n",
        "!pip install tensorflowjs"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting tensorflowjs\n",
            "  Downloading https://files.pythonhosted.org/packages/7d/96/20120de578fe12c9e25e2b08dcf472a47bb0704bd1c894893f1a1f7e532d/tensorflowjs-0.5.7-py3-none-any.whl\n",
            "Requirement already satisfied: keras==2.1.6 in /usr/local/lib/python3.6/dist-packages (from tensorflowjs) (2.1.6)\n",
            "Requirement already satisfied: h5py==2.8.0 in /usr/local/lib/python3.6/dist-packages (from tensorflowjs) (2.8.0)\n",
            "Requirement already satisfied: six==1.11.0 in /usr/local/lib/python3.6/dist-packages (from tensorflowjs) (1.11.0)\n",
            "Collecting tensorflow-hub==0.1.0 (from tensorflowjs)\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/aa/8e/ab0d5cf0f2dc2ed1c645bf1cb4acc7beeb0be5a51ebb0c91cbb23eb1fe10/tensorflow_hub-0.1.0-py2.py3-none-any.whl (53kB)\n",
            "\u001b[K    100% |████████████████████████████████| 61kB 8.2MB/s \n",
            "\u001b[?25hCollecting tensorflow==1.9.0 (from tensorflowjs)\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/86/95/274190b39950e1e9eef4b071acefea832ac3e2c19bb4b442fa54f3214d2e/tensorflow-1.9.0-cp36-cp36m-manylinux1_x86_64.whl (51.1MB)\n",
            "\u001b[K    100% |████████████████████████████████| 51.1MB 782kB/s \n",
            "\u001b[?25hCollecting numpy==1.14.1 (from tensorflowjs)\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/de/7d/348c5d8d44443656e76285aa97b828b6dbd9c10e5b9c0f7f98eff0ff70e4/numpy-1.14.1-cp36-cp36m-manylinux1_x86_64.whl (12.2MB)\n",
            "\u001b[K    100% |████████████████████████████████| 12.2MB 3.6MB/s \n",
            "\u001b[?25hRequirement already satisfied: pyyaml in /usr/local/lib/python3.6/dist-packages (from keras==2.1.6->tensorflowjs) (3.13)\n",
            "Requirement already satisfied: scipy>=0.14 in /usr/local/lib/python3.6/dist-packages (from keras==2.1.6->tensorflowjs) (0.19.1)\n",
            "Requirement already satisfied: protobuf>=3.4.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow-hub==0.1.0->tensorflowjs) (3.6.1)\n",
            "Requirement already satisfied: termcolor>=1.1.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow==1.9.0->tensorflowjs) (1.1.0)\n",
            "Requirement already satisfied: gast>=0.2.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow==1.9.0->tensorflowjs) (0.2.0)\n",
            "Requirement already satisfied: grpcio>=1.8.6 in /usr/local/lib/python3.6/dist-packages (from tensorflow==1.9.0->tensorflowjs) (1.14.2)\n",
            "Requirement already satisfied: absl-py>=0.1.6 in /usr/local/lib/python3.6/dist-packages (from tensorflow==1.9.0->tensorflowjs) (0.4.1)\n",
            "Collecting tensorboard<1.10.0,>=1.9.0 (from tensorflow==1.9.0->tensorflowjs)\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/9e/1f/3da43860db614e294a034e42d4be5c8f7f0d2c75dc1c428c541116d8cdab/tensorboard-1.9.0-py3-none-any.whl (3.3MB)\n",
            "\u001b[K    100% |████████████████████████████████| 3.3MB 7.2MB/s \n",
            "\u001b[?25hRequirement already satisfied: setuptools<=39.1.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow==1.9.0->tensorflowjs) (39.1.0)\n",
            "Requirement already satisfied: astor>=0.6.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow==1.9.0->tensorflowjs) (0.7.1)\n",
            "Requirement already satisfied: wheel>=0.26 in /usr/local/lib/python3.6/dist-packages (from tensorflow==1.9.0->tensorflowjs) (0.31.1)\n",
            "Requirement already satisfied: markdown>=2.6.8 in /usr/local/lib/python3.6/dist-packages (from tensorboard<1.10.0,>=1.9.0->tensorflow==1.9.0->tensorflowjs) (2.6.11)\n",
            "Requirement already satisfied: werkzeug>=0.11.10 in /usr/local/lib/python3.6/dist-packages (from tensorboard<1.10.0,>=1.9.0->tensorflow==1.9.0->tensorflowjs) (0.14.1)\n",
            "Installing collected packages: numpy, tensorflow-hub, tensorboard, tensorflow, tensorflowjs\n",
            "  Found existing installation: numpy 1.14.5\n",
            "    Uninstalling numpy-1.14.5:\n",
            "      Successfully uninstalled numpy-1.14.5\n",
            "  Found existing installation: tensorflow-hub 0.1.1\n",
            "    Uninstalling tensorflow-hub-0.1.1:\n",
            "      Successfully uninstalled tensorflow-hub-0.1.1\n",
            "  Found existing installation: tensorboard 1.10.0\n",
            "    Uninstalling tensorboard-1.10.0:\n",
            "      Successfully uninstalled tensorboard-1.10.0\n",
            "  Found existing installation: tensorflow 1.10.1\n",
            "    Uninstalling tensorflow-1.10.1:\n",
            "      Successfully uninstalled tensorflow-1.10.1\n",
            "Successfully installed numpy-1.14.1 tensorboard-1.9.0 tensorflow-1.9.0 tensorflow-hub-0.1.0 tensorflowjs-0.5.7\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "2BvQoFYkQicw",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "5d6b679a-cbbe-4ba3-97a9-b37b2b839f39"
      },
      "cell_type": "code",
      "source": [
        "# Use tensorflowjs converter to convert the saved model to JSON format\n",
        "\n",
        "!tensorflowjs_converter \\\n",
        "    --input_format=keras \\\n",
        "    --output_format=tensorflowjs \\\n",
        "    new.h5 \\\n",
        "    model"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Using TensorFlow backend.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "UYovvA-UtVCs",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# download saved h5 keras model \n",
        "files.download('new.h5')\n",
        "\n",
        "# download converted JSON model and weights under the converted 'model' folder\n",
        "files.download('/model/*.*')"
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}