TATHAGATA_GHOSH_mnist_cnn.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "TATHAGATA GHOSH - mnist_cnn.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.6.8"
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/tathatj/Image-Classifier/blob/master/TATHAGATA_GHOSH_mnist_cnn.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "MvErumT0n0_-"
      },
      "source": [
        "# Classification of handwritten digits using MLN"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "zh0uH8bDn1AE"
      },
      "source": [
        "# MNIST Dataset\n",
        "\n",
        "<img src=\"https://upload.wikimedia.org/wikipedia/commons/2/27/MnistExamples.png\" title=\"MNIST dataset\" align=\"center\"/>\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "1K6Im4RYn1AK"
      },
      "source": [
        "# Problem\n",
        "Classify handwritten digits from 0 - 9. <br>\n",
        "Each image is 28x28 pixels\n",
        "\n",
        "<img src=\"https://corochann.com/wp-content/uploads/2017/02/mnist_plot.png\" title=\"\" align=\"center\"/>\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "EVex4puqn1AO"
      },
      "source": [
        "# Understanding the data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "deletable": false,
        "id": "alqmfbOIn1AR",
        "nbgrader": {
          "checksum": "f1c9b1f58906b8bf2546a86766ba19a8",
          "grade": false,
          "grade_id": "cell-81641c87749d35fd",
          "locked": false,
          "schema_version": 1,
          "solution": true
        },
        "outputId": "85dbea4f-610c-43cf-e707-0ac709d0f3f5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 73
        }
      },
      "source": [
        "# MNIST data is present in the keras library. You may load it from there\n",
        "from keras.datasets import mnist\n",
        "def load_data():\n",
        "    \"\"\"\n",
        "    Inputs:\n",
        "        None\n",
        "    Outputs:\n",
        "        train_samples, train_labels, test_samples, test_labels: numpy arrays\n",
        "    \n",
        "    Load the train/test of mnist data into these variables\n",
        "    \"\"\"\n",
        "    # YOUR CODE HERE\n",
        "    (train_samples, train_labels), (test_samples, test_labels) = mnist.load_data()\n",
        "\n",
        "    return train_samples, train_labels, test_samples, test_labels \n",
        "\n",
        "train_samples, train_labels, test_samples, test_labels = load_data()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Using TensorFlow backend.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Downloading data from https://s3.amazonaws.com/img-datasets/mnist.npz\n",
            "11493376/11490434 [==============================] - 0s 0us/step\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yK_GoHzu5RUW",
        "colab_type": "code",
        "outputId": "f921a37c-0b87-4e0b-ea2f-f4e9680e05f1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 36
        }
      },
      "source": [
        "\"\"\"Test cases\"\"\"\n",
        "assert train_samples.shape == (60000, 28, 28)\n",
        "assert test_labels.shape == (10000,)\n",
        "print('Test passed', '\\U0001F44D')\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Test passed 👍\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "3CNRCPbrn1Ae"
      },
      "source": [
        "### Shape of Data\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "_pAn6o3Vn1Ag",
        "outputId": "b372372a-695c-43db-ff27-69be0d32500a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 92
        }
      },
      "source": [
        "print(train_samples.shape)\n",
        "print(train_labels.shape)\n",
        "print(test_samples.shape)\n",
        "print(test_labels.shape)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(60000, 28, 28)\n",
            "(60000,)\n",
            "(10000, 28, 28)\n",
            "(10000,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "ubf7aJPIn1Ax"
      },
      "source": [
        "### Range of Values"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "-jiF1J8Jn1A0",
        "outputId": "c11f9e20-1a30-40e5-aebe-208452d7bbfd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 36
        }
      },
      "source": [
        "import numpy as np\n",
        "np.amax(train_samples) # Max value"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "255"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "Mu18byDmn1A-",
        "outputId": "9b66bce3-435e-4b1e-df4b-51cc81a60fb5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 36
        }
      },
      "source": [
        "np.amin(train_samples) # Min Value"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "_OkrzUFQn1BI",
        "outputId": "b94284b5-04ff-41de-fddf-1e1961fdd437",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 288
        }
      },
      "source": [
        "%matplotlib inline\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "pixels = train_samples[0] # Shape (28, 28)\n",
        "plt.imshow(pixels, cmap='gray')\n",
        "plt.show()\n",
        "print('Label of image is', train_labels[0])"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Label of image is 5\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "N6JiZgztn1BP"
      },
      "source": [
        "# Data Preparation"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "7hUHie4_n1BQ"
      },
      "source": [
        "### Normalize inputs to (0, 1)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "deletable": false,
        "id": "iGKZASBBn1BS",
        "nbgrader": {
          "checksum": "517fb19d6f3ebbc315f78688e3cf580a",
          "grade": false,
          "grade_id": "cell-6d7ec54fdf30ec6a",
          "locked": false,
          "schema_version": 1,
          "solution": true
        },
        "colab": {}
      },
      "source": [
        "def convert_dtype(x):\n",
        "    \"\"\"\n",
        "    Inputs:\n",
        "        x: numpy array\n",
        "    Outputs:\n",
        "        x_float: numpy array, dtype of elements changed to 'float32'\n",
        "    \"\"\"\n",
        "    # YOUR CODE HERE\n",
        "    x_float = x.astype('float32')\n",
        "    \n",
        "    return x_float\n",
        "\n",
        "train_samples = convert_dtype(train_samples)\n",
        "test_samples = convert_dtype(test_samples)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IB2huSQl5RVW",
        "colab_type": "code",
        "outputId": "add802bb-7a26-4159-f6d2-720cdb0cf004",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 36
        }
      },
      "source": [
        "\"\"\"Test cases\"\"\"\n",
        "assert str(train_samples.dtype) == 'float32'\n",
        "print('Test passed', '\\U0001F44D')\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Test passed 👍\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "deletable": false,
        "id": "gbaIejYPn1BX",
        "nbgrader": {
          "checksum": "460dfe1a17dd7cbfe4b2783f3f7d66dc",
          "grade": false,
          "grade_id": "cell-f942b7b27fdbdc51",
          "locked": false,
          "schema_version": 1,
          "solution": true
        },
        "colab": {}
      },
      "source": [
        "def normalize(x):\n",
        "    \"\"\"\n",
        "    Inputs:\n",
        "        x: numpy array\n",
        "    Outputs:\n",
        "        x_n: numpy array, elements normalized to be between (0, 1)\n",
        "    \"\"\"    \n",
        "    # YOUR CODE HERE\n",
        "    x_n = x/255.\n",
        "    return x_n\n",
        "    \n",
        "train_samples = normalize(train_samples)\n",
        "test_samples = normalize(test_samples) \n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cNZ-cCOuIMbK",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PUPxc8Qb5RVj",
        "colab_type": "code",
        "outputId": "5077d811-ab75-4bb4-8957-ad5e4b523ae2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 36
        }
      },
      "source": [
        "\"\"\"Test cases\"\"\"\n",
        "assert np.isclose(np.amax(train_samples), 1)\n",
        "print('Test passed', '\\U0001F44D')\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Test passed 👍\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TVHu6HY5IZ_m",
        "colab_type": "code",
        "outputId": "62fc9a58-b914-43b0-f0d0-89ac0c322f91",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 36
        }
      },
      "source": [
        "train_samples.shape"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(60000, 28, 28)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "deletable": false,
        "nbgrader": {
          "checksum": "95f321beea0331f5d325a1dd113645da",
          "grade": false,
          "grade_id": "cell-8a7f6b723861a02a",
          "locked": false,
          "schema_version": 1,
          "solution": true
        },
        "id": "aCR6DCgA5RVp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def reshape(x):\n",
        "    \"\"\"\n",
        "    We need to reshape our train_data to be of shape (samples, height, width, channels) pass to Conv2D layer of keras\n",
        "    Inputs:\n",
        "        x: numpy array of shape(samples, height, width)\n",
        "    Outputs:\n",
        "        x_r: numpy array of shape(samples, height, width, 1)\n",
        "    \"\"\"\n",
        "    # YOUR CODE HERE\n",
        "    x_r = x.reshape(x.shape[0], 28,28,1)\n",
        "    return x_r\n",
        "\n",
        "train_samples = reshape(train_samples)\n",
        "test_samples = reshape(test_samples)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fUrOyiSU5RVv",
        "colab_type": "code",
        "outputId": "20880fee-b23c-4394-8dc0-b5aa6a89dfd1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 36
        }
      },
      "source": [
        "train_labels.shape"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(60000,)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "jzvfaeh9n1Bj"
      },
      "source": [
        "### Convert outputs to 1-hot vectors\n",
        "\\begin{equation*}\n",
        "Eg: 5 \\rightarrow [0, 0, 0, 0, 0, 1, 0, 0, 0, 0]\n",
        "\\end{equation*}"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "deletable": false,
        "id": "kCxrh0-nn1Bl",
        "nbgrader": {
          "checksum": "64b5aebaeb79b72c5f8d6b5fc4158bd1",
          "grade": false,
          "grade_id": "cell-c11f23149e00b861",
          "locked": false,
          "schema_version": 1,
          "solution": true
        },
        "colab": {}
      },
      "source": [
        "def oneHot(y, Ny):\n",
        "    \"\"\"\n",
        "    Inputs:\n",
        "        y: numpy array if shape (samples, ) with class labels\n",
        "        Ny: number of classes\n",
        "    Outputs:\n",
        "        y_oh: numpy array of shape (samples, Ny) of one hot vectors\n",
        "    \"\"\"\n",
        "    # YOUR CODE HERE\n",
        "    from tensorflow.keras.utils import to_categorical\n",
        "\n",
        "    y_oh = to_categorical(y,Ny)\n",
        "    return y_oh\n",
        "\n",
        "# example\n",
        "train_labels = oneHot(train_labels, 10)\n",
        "test_labels = oneHot(test_labels, 10)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Uot6K52Z5RWB",
        "colab_type": "code",
        "outputId": "200d6c5b-84e1-43ae-dc69-7f34128fd9c6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 36
        }
      },
      "source": [
        "\"\"\"Test cases\"\"\"\n",
        "assert train_labels.shape[1] == 10\n",
        "print('Test passed', '\\U0001F44D')\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Test passed 👍\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ntJekQJ85RWI",
        "colab_type": "text"
      },
      "source": [
        "## Create a convolutional neural network model\n",
        "You may design whatever cnn model you like. But following are hints to get started.<br>\n",
        "Make the following layers:\n",
        "1. cnn layer with kernel_size = (5, 5) and 32 kernels\n",
        "2. cnn layer with kernel_size = (3, 3) and 20 kernels\n",
        "3. Maxpooling layer of size (2, 2)\n",
        "4. Flatten layer\n",
        "5. Dense layer of appropriate size\n",
        "6. Output layer of appropriate size"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "deletable": false,
        "id": "amqRJpZun1Bt",
        "nbgrader": {
          "checksum": "2c2396a31ef46bae9d5b0fa50a2db297",
          "grade": false,
          "grade_id": "cell-f2918a15da09ae15",
          "locked": false,
          "schema_version": 1,
          "solution": true
        },
        "outputId": "140bf427-af7c-485e-caf6-461f1fe7a4f5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 281
        }
      },
      "source": [
        "def create_model():\n",
        "    \"\"\"\n",
        "    Inputs:\n",
        "        None\n",
        "    Outputs:\n",
        "        model: compiled keras model\n",
        "    \"\"\"\n",
        "    # YOUR CODE HERE\n",
        "    from keras.models import Sequential\n",
        "    from keras.layers import Dense, Dropout, Flatten\n",
        "    from keras.layers import Conv2D, MaxPooling2D\n",
        "    from tensorflow.keras import optimizers\n",
        "\n",
        "    model = Sequential()\n",
        "    model.add(Conv2D(32, kernel_size=(3, 3),\n",
        "                 activation='relu',\n",
        "                 input_shape=(28,28,1)))\n",
        "    \n",
        "    model.add(Conv2D(64, kernel_size=(3, 3), activation='relu'))\n",
        "    model.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "    model.add(Flatten())\n",
        "\n",
        "    model.add(Dense(128, activation='relu'))\n",
        "    model.add(Dense(10, activation='softmax'))\n",
        "    model.compile(optimizer='sgd', loss='categorical_crossentropy', metrics=['accuracy'])\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "    return model\n",
        "\n",
        "model = create_model()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING: Logging before flag parsing goes to stderr.\n",
            "W0623 18:02:17.159910 140594691581824 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:74: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n",
            "\n",
            "W0623 18:02:17.196750 140594691581824 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n",
            "\n",
            "W0623 18:02:17.205101 140594691581824 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n",
            "\n",
            "W0623 18:02:17.244240 140594691581824 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:3976: The name tf.nn.max_pool is deprecated. Please use tf.nn.max_pool2d instead.\n",
            "\n",
            "W0623 18:02:17.287856 140594691581824 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
            "\n",
            "W0623 18:02:17.309474 140594691581824 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:3295: The name tf.log is deprecated. Please use tf.math.log instead.\n",
            "\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kl2bX0hh5RWQ",
        "colab_type": "code",
        "outputId": "02991009-6937-4c96-b3b7-583e5ebb7068",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 373
        }
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d_1 (Conv2D)            (None, 26, 26, 32)        320       \n",
            "_________________________________________________________________\n",
            "conv2d_2 (Conv2D)            (None, 24, 24, 64)        18496     \n",
            "_________________________________________________________________\n",
            "max_pooling2d_1 (MaxPooling2 (None, 12, 12, 64)        0         \n",
            "_________________________________________________________________\n",
            "flatten_1 (Flatten)          (None, 9216)              0         \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 128)               1179776   \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 10)                1290      \n",
            "=================================================================\n",
            "Total params: 1,199,882\n",
            "Trainable params: 1,199,882\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "u-2Sfsxtn1B7",
        "outputId": "5b55a608-23f9-4b50-c8e2-888b4539cb5e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 225
        }
      },
      "source": [
        "history = model.fit(train_samples, train_labels, validation_split = 0.1, epochs=2, batch_size=200)\n",
        "# Use 10% of samples for validation, validation_split is the relevant parameter\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "W0623 18:02:17.437232 140594691581824 deprecation.py:323] From /usr/local/lib/python3.6/dist-packages/tensorflow/python/ops/math_grad.py:1250: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "Use tf.where in 2.0, which has the same broadcast rule as np.where\n",
            "W0623 18:02:17.494397 140594691581824 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:986: The name tf.assign_add is deprecated. Please use tf.compat.v1.assign_add instead.\n",
            "\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Train on 54000 samples, validate on 6000 samples\n",
            "Epoch 1/2\n",
            "54000/54000 [==============================] - 10s 181us/step - loss: 1.0884 - acc: 0.6917 - val_loss: 0.3125 - val_acc: 0.9117\n",
            "Epoch 2/2\n",
            "54000/54000 [==============================] - 3s 56us/step - loss: 0.3339 - acc: 0.9019 - val_loss: 0.2289 - val_acc: 0.9342\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "deletable": false,
        "nbgrader": {
          "checksum": "152e035c796daf8de8495ef4dd4b509b",
          "grade": false,
          "grade_id": "cell-2e082c9902d9c054",
          "locked": false,
          "schema_version": 1,
          "solution": true
        },
        "id": "otxaaM845RWi",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def predict(x):\n",
        "    \"\"\"\n",
        "    Inputs:\n",
        "        x: input samples\n",
        "        model: keras model\n",
        "    Outputs:\n",
        "        y: predicted labels\n",
        "    \"\"\"\n",
        "    # YOUR CODE HERE\n",
        "    y = model.predict(x)\n",
        "    return y"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "deletable": false,
        "nbgrader": {
          "checksum": "51bf0f6f01341d1e097cfcce995f7842",
          "grade": false,
          "grade_id": "cell-b55a1773f8fcdf85",
          "locked": false,
          "schema_version": 1,
          "solution": true
        },
        "id": "kvd9eiKM5RWs",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def oneHot_tolabel(y):\n",
        "    \"\"\"\n",
        "    Inputs:\n",
        "        y: numpy array of shape (samples, Ny)\n",
        "    Outputs:\n",
        "        y_b: numpy array of shape (samples,) where one hot encoding is converted back to class labels\n",
        "    \"\"\"\n",
        "    # YOUR CODE HERE\n",
        "     \n",
        "    y_b=np.argmax(y, axis=1)\n",
        "    return y_b\n",
        "    "
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "deletable": false,
        "id": "B12vN44Ln1CF",
        "nbgrader": {
          "checksum": "177572664de00d24d4f6271664cbc95f",
          "grade": false,
          "grade_id": "cell-7b7fec626903c453",
          "locked": false,
          "schema_version": 1,
          "solution": true
        },
        "colab": {}
      },
      "source": [
        "def create_confusion_matrix(true_labels, predicted_labels):\n",
        "    \"\"\"\n",
        "    Inputs:\n",
        "        true_labels: numpy array of shape (samples, ) with true_labels\n",
        "        test_labels: numpy array of shape(samples, ) with test_labels\n",
        "    Outputs:\n",
        "        cm: numpy array of shape (Ny, Ny), confusion matrix. Ny -> number of unique classes in y\n",
        "    \"\"\"\n",
        "    # YOUR CODE HERE\n",
        "    from sklearn.metrics import confusion_matrix\n",
        "    \n",
        "    cm = confusion_matrix(true_labels, predicted_labels)\n",
        "\n",
        "    return cm"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ntzOmfSO5RW1",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "predicted_labels_train = predict(train_samples)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nluDi-n8THb9",
        "colab_type": "code",
        "outputId": "da693c62-5945-484a-c719-bf1ba96ae2ad",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 36
        }
      },
      "source": [
        "print(predicted_labels_train.shape)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(60000, 10)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KNrcI9nZ5RW-",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "cm = create_confusion_matrix(oneHot_tolabel(train_labels), oneHot_tolabel(predict(train_samples)))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "t--cvIRI5RXD",
        "colab_type": "code",
        "outputId": "bdd1d8c8-8374-4271-8312-177ff5b73a31",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 205
        }
      },
      "source": [
        "print(cm)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[5672    1    8   18   10   51   59   17   79    8]\n",
            " [   0 6582   40   14    6   29   12   20   29   10]\n",
            " [  53   72 5141   87  110   15  157  119  174   30]\n",
            " [  33   50   83 5342    5  197   41  107  215   58]\n",
            " [   5   29   28    0 5426    2   83   30   32  207]\n",
            " [  60   52   19  171   49 4707  129   33  151   50]\n",
            " [  26   19   12    2   46   56 5721    4   31    1]\n",
            " [  26   52   56    7   67    8    1 5904   13  131]\n",
            " [  23  113   34  124   19   82   49   29 5301   77]\n",
            " [  31   40   16   73  183   23    1  249   63 5270]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "28mectC3Kn2Q",
        "colab_type": "code",
        "outputId": "c94b400c-d6e7-4d49-904f-54529d333a43",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 73
        }
      },
      "source": [
        "precision = np.diag(cm) / np.sum(cm, axis = 0)\n",
        "print(precision )\n",
        "min=np.amin(precision)\n",
        "print(min)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[0.95665374 0.93894437 0.94555821 0.9150394  0.91639926 0.91044487\n",
            " 0.91492084 0.90663391 0.8707293  0.90208833]\n",
            "0.8707293035479632\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "sdXS58IwWOXP",
        "outputId": "bfc8e956-eae7-4cd8-9ffc-7465e41dabec",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 269
        }
      },
      "source": [
        "history.history.keys()\n",
        "import matplotlib.pyplot as plt\n",
        "plt.plot(range(len(history.history['val_acc'])), history.history['val_acc'])\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAD8CAYAAAB3u9PLAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nCCIiEqRAEBERAP4/n8ANVbofyG8AAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "deletable": false,
        "nbgrader": {
          "checksum": "a9dfa82486edd7a314c8ac7fd47e389e",
          "grade": false,
          "grade_id": "cell-aba6e55dfdf66734",
          "locked": false,
          "schema_version": 1,
          "solution": true
        },
        "id": "Khcn6sy55RXP",
        "colab_type": "code",
        "outputId": "e50271c9-4a7f-4ef1-d9ad-da275e8b0eb9",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 36
        }
      },
      "source": [
        "def accuracy(x_test, y_test, model):\n",
        "    \"\"\"\n",
        "    Inputs:\n",
        "        x_test: test samples\n",
        "        y_test : test labels\n",
        "        model: keras model\n",
        "    Ouputs:\n",
        "        acc: float, accuracy of test data on model\n",
        "    \"\"\"\n",
        "    # YOUR CODE HERE\n",
        "    loss, acc = model.evaluate(x_test, y_test,verbose=0)\n",
        "    return acc\n",
        "\n",
        "acc = accuracy(test_samples, test_labels, model)\n",
        "print('Test accuracy is, ', acc*100, '%')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Test accuracy is,  92.05 %\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "DySmpDX7n1CY",
        "colab": {}
      },
      "source": [
        "### Advanced\n",
        "1. Tune the hyperparameters to better the performance\n",
        "2. Find the classes which are getting most confused from confusion matrix. Take out those samples. Plot them and see why they are getting confused."
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GP4UQ7Zz5RXW",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def create_model():\n",
        "    \"\"\"\n",
        "    Inputs:\n",
        "        None\n",
        "    Outputs:\n",
        "        model: compiled keras model\n",
        "    \"\"\"\n",
        "    # YOUR CODE HERE\n",
        "    from keras.models import Sequential\n",
        "    from keras.layers import Dense, Dropout, Flatten\n",
        "    from keras.layers import Conv2D, MaxPooling2D\n",
        "    from tensorflow.keras import optimizers\n",
        "\n",
        "    model = Sequential()\n",
        "    model.add(Conv2D(64, kernel_size=(3, 3),\n",
        "                 activation='relu',\n",
        "                 input_shape=(28,28,1)))\n",
        "    \n",
        "    model.add(Conv2D(128, kernel_size=(3, 3), activation='tanh'))\n",
        "    model.add(MaxPooling2D(pool_size=(2, 2)))\n",
        "    model.add(Flatten())\n",
        "\n",
        "    model.add(Dense(128, activation='relu'))\n",
        "    model.add(Dense(10, activation='softmax'))\n",
        "    model.compile(optimizer='sgd', loss='categorical_crossentropy', metrics=['accuracy'])\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "    return model\n",
        "\n",
        "model = create_model()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JxIlZACPFlaP",
        "colab_type": "code",
        "outputId": "3f83ef93-aabd-4157-d8bc-6feab1faed1d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 373
        }
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d_11 (Conv2D)           (None, 26, 26, 64)        640       \n",
            "_________________________________________________________________\n",
            "conv2d_12 (Conv2D)           (None, 24, 24, 128)       73856     \n",
            "_________________________________________________________________\n",
            "max_pooling2d_6 (MaxPooling2 (None, 12, 12, 128)       0         \n",
            "_________________________________________________________________\n",
            "flatten_6 (Flatten)          (None, 18432)             0         \n",
            "_________________________________________________________________\n",
            "dense_11 (Dense)             (None, 128)               2359424   \n",
            "_________________________________________________________________\n",
            "dense_12 (Dense)             (None, 10)                1290      \n",
            "=================================================================\n",
            "Total params: 2,435,210\n",
            "Trainable params: 2,435,210\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8GGLxfjRFoW4",
        "colab_type": "code",
        "outputId": "ae4d3c1e-ff5b-4594-9ded-ee490f5ec928",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 186
        }
      },
      "source": [
        "history = model.fit(train_samples, train_labels, validation_split = 0.2, epochs=4, batch_size=200)\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train on 48000 samples, validate on 12000 samples\n",
            "Epoch 1/4\n",
            "48000/48000 [==============================] - 6s 123us/step - loss: 1.1301 - acc: 0.7276 - val_loss: 0.3855 - val_acc: 0.8949\n",
            "Epoch 2/4\n",
            "48000/48000 [==============================] - 5s 114us/step - loss: 0.3589 - acc: 0.8967 - val_loss: 0.2936 - val_acc: 0.9198\n",
            "Epoch 3/4\n",
            "48000/48000 [==============================] - 5s 114us/step - loss: 0.2952 - acc: 0.9141 - val_loss: 0.2572 - val_acc: 0.9263\n",
            "Epoch 4/4\n",
            "48000/48000 [==============================] - 5s 114us/step - loss: 0.2604 - acc: 0.9235 - val_loss: 0.2368 - val_acc: 0.9308\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rND2nk68SxW4",
        "colab_type": "code",
        "outputId": "74c911af-bf03-46c9-e7f2-5c4dc412a27a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 205
        }
      },
      "source": [
        "print(cm)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[5672    1    8   18   10   51   59   17   79    8]\n",
            " [   0 6582   40   14    6   29   12   20   29   10]\n",
            " [  53   72 5141   87  110   15  157  119  174   30]\n",
            " [  33   50   83 5342    5  197   41  107  215   58]\n",
            " [   5   29   28    0 5426    2   83   30   32  207]\n",
            " [  60   52   19  171   49 4707  129   33  151   50]\n",
            " [  26   19   12    2   46   56 5721    4   31    1]\n",
            " [  26   52   56    7   67    8    1 5904   13  131]\n",
            " [  23  113   34  124   19   82   49   29 5301   77]\n",
            " [  31   40   16   73  183   23    1  249   63 5270]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZIsyrdaaV9Lk",
        "colab_type": "code",
        "outputId": "07f756ad-fc05-45ae-8c4e-dc1f26c4de64",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 73
        }
      },
      "source": [
        "precision = np.diag(cm) / np.sum(cm, axis = 0)\n",
        "print(precision )\n",
        "min=np.amin(precision)\n",
        "print(min)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[0.95665374 0.93894437 0.94555821 0.9150394  0.91639926 0.91044487\n",
            " 0.91492084 0.90663391 0.8707293  0.90208833]\n",
            "0.8707293035479632\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jXCjzk-mWhAH",
        "colab_type": "code",
        "outputId": "b0a468de-a926-446c-8efe-88d3e13921da",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 36
        }
      },
      "source": [
        "train_samples.shape"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(60000, 28, 28, 1)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 109
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wnGT6iHKW2GI",
        "colab_type": "code",
        "outputId": "4316d41f-1109-4dcf-d13f-9e5c65d27365",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 36
        }
      },
      "source": [
        "tx=train_samples.reshape(train_samples.shape[0], 28,28)\n",
        "tx.shape"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(60000, 28, 28)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 32
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "C6v_28y3XNbN",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7vy_ScuwWXuH",
        "colab_type": "code",
        "outputId": "79b8065f-3125-432a-f693-eb1cbae8fc94",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1371
        }
      },
      "source": [
        "%matplotlib inline\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "for i in range(5):\n",
        "  if oneHot_tolabel(predict(train_samples))[i]==8:\n",
        "    pixels = tx[i] # Shape (28, 28)\n",
        "  plt.imshow(pixels, cmap='gray')\n",
        "  plt.show()\n",
        "  print('Label of image is', oneHot_tolabel(predict(train_samples))[i])\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Label of image is 5\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Label of image is 0\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Label of image is 4\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Label of image is 1\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Label of image is 9\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}