mnist_dataset_Sequential_model.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "authorship_tag": "ABX9TyMESBwZLdsP7t2wSJGlyoo1",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/aaolcay/Convolutional-Neural-Network-Examples/blob/main/mnist_dataset_Sequential_model.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## **CNN for `mnist` Dataset**\n",
        "In this notebook, we applied one simple CNN architecture by using Sequential model. "
      ],
      "metadata": {
        "id": "fs4ij0aKHtcQ"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 70,
      "metadata": {
        "id": "FEKNIwwCuKEt"
      },
      "outputs": [],
      "source": [
        "import tensorflow.keras as keras\n",
        "from keras.datasets import mnist\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### **Load Dataset**\n",
        "Data chosen in this implementation is quite famous, and most of the tutorials on CNNs include this dataset as it is very easy to load and to use with CNN model considering model's high accuracy score."
      ],
      "metadata": {
        "id": "tpG06b5EumyA"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "(x_train, y_train), (x_test,y_test) = mnist.load_data()"
      ],
      "metadata": {
        "id": "fZCme5B0ulvU"
      },
      "execution_count": 71,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### **Data Shapes**\n",
        "We need to know the data shape and which type of data we deal with. This should also be noticed that 2 dimensional convolutional layer (`Conv2D`) needs a shape of (**#samples**, **height**, **width**, **#channels**). For example, imagine that we have 200 training data consist of images whose heights (i.e., #rows) are 28 and widths are 28 (i.e., #columns) as well. For colorful image the number of channels is 3 (RGB: Red, Green, Blue) but gray images have only one channel that is what we consider in this implementation for making prediction.  "
      ],
      "metadata": {
        "id": "opHme_NRvbfr"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print(f'The shape of training data is : {x_train.shape}.\\nThe shape of test data is {x_test.shape}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "w0NWf-fUvbAe",
        "outputId": "a97f81db-e3cf-4612-f551-9a03b91638e0"
      },
      "execution_count": 72,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The shape of training data is : (60000, 28, 28).\n",
            "The shape of test data is (10000, 28, 28)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "The data shapes are not appropriate to be used within the 2D convolutional layers. Therefore, we need to add channel number to the datasets.\n",
        "\n",
        "                                  (60000, 28, 28) -> (60000, 28, 28, 1)\n",
        "\n",
        "                                  (10000, 28, 28) -> (10000, 28, 28, 1)\n"
      ],
      "metadata": {
        "id": "j9KRFQ55xO8D"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "x_train = x_train[..., np.newaxis]\n",
        "x_test = x_test[..., np.newaxis]\n",
        "print(f'The shape of training data is : {x_train.shape}.\\nThe shape of test data is {x_test.shape}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7fJliZANxNOl",
        "outputId": "16ef6b6e-7575-4641-8a27-c0fc83674857"
      },
      "execution_count": 73,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The shape of training data is : (60000, 28, 28, 1).\n",
            "The shape of test data is (10000, 28, 28, 1)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Let's now look at the labels and convert them to categorical form. Watch out CNN is a deep learning model and needs labeled data that makes it a supervised method. "
      ],
      "metadata": {
        "id": "YFCul16mycQn"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print(f'Labels of train dataset: {y_train[:20]}\\nLabels of test dataset: {y_test[:20]}')\n",
        "print(f'The shape of training data is : {y_train.shape}.\\nThe shape of test data is {y_test.shape}')\n",
        "ytrain = y_train.copy()\n",
        "ytest = y_test.copy()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xfTyR3dayObo",
        "outputId": "d982a1a6-91a2-4b16-de28-1c2614f01e9a"
      },
      "execution_count": 74,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Labels of train dataset: [5 0 4 1 9 2 1 3 1 4 3 5 3 6 1 7 2 8 6 9]\n",
            "Labels of test dataset: [7 2 1 0 4 1 4 9 5 9 0 6 9 0 1 5 9 7 3 4]\n",
            "The shape of training data is : (60000,).\n",
            "The shape of test data is (10000,)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Visualize each label with the data (images of numbers)."
      ],
      "metadata": {
        "id": "TY-MymH_zVMz"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "fig = plt.figure(figsize=(5,5))\n",
        "for i in range(20):\n",
        "  x_test[i,:,:,:]\n",
        "  fig.add_subplot(4,5,i+1)\n",
        "  plt.imshow(x_train[i,:,:,:], cmap='gray')\n",
        "  plt.title(f'{y_train[i]}')\n",
        "  fig.tight_layout(pad=0.2)\n",
        "\n",
        "fig = plt.figure(figsize=(5,5))\n",
        "for k in range(20):\n",
        "  fig.add_subplot(4,5,k+1)\n",
        "  plt.imshow(x_test[k,:,:,:], cmap='gray')\n",
        "  plt.title(f'{y_test[k]}')\n",
        "  fig.tight_layout(pad=0.2)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 729
        },
        "id": "g1-Xo9wxy-v_",
        "outputId": "cb37d659-2898-480c-efd4-7dfce7b8d40b"
      },
      "execution_count": 75,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 360x360 with 20 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 360x360 with 20 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now, let's convert each label into categorical from:\n",
        "\n",
        "0 -> [1 0 0 0 0 0 0 0 0 0]\n",
        "\n",
        "1 -> [0 1 0 0 0 0 0 0 0 0]\n",
        "\n",
        "2 -> [0 0 1 0 0 0 0 0 0 0]\n",
        "\n",
        "3 -> [0 0 0 1 0 0 0 0 0 0]\n",
        "\n",
        "4 -> [0 0 0 0 1 0 0 0 0 0]\n",
        "\n",
        "5 -> [0 0 0 0 0 1 0 0 0 0]\n",
        "\n",
        "6 -> [0 0 0 0 0 0 1 0 0 0]\n",
        "\n",
        "7 -> [0 0 0 0 0 0 0 1 0 0]\n",
        "\n",
        "8 -> [0 0 0 0 0 0 0 0 1 0]\n",
        "\n",
        "9 -> [0 0 0 0 0 0 0 0 0 1]"
      ],
      "metadata": {
        "id": "uj48ICTP-q9r"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "num_class = 10\n",
        "y_train = keras.utils.to_categorical(y_train, num_class)\n",
        "y_test = keras.utils.to_categorical(y_test, num_class)"
      ],
      "metadata": {
        "id": "PRWIjhq60fEl"
      },
      "execution_count": 76,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "for i in range(20):\n",
        "  print(f'Labels of train dataset:\\n{ytrain[i]}->{y_train[i]}\\nLabels of test dataset:\\n{ytest[i]}->{y_test[i]}')\n",
        "\n",
        "print(f'The shape of training data is : {y_train.shape}.\\nThe shape of test data is {y_test.shape}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Q8PtOqHPAHpA",
        "outputId": "58c57f27-4b40-4b2c-cf92-70aa7e326141"
      },
      "execution_count": 77,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Labels of train dataset:\n",
            "5->[0. 0. 0. 0. 0. 1. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "7->[0. 0. 0. 0. 0. 0. 0. 1. 0. 0.]\n",
            "Labels of train dataset:\n",
            "0->[1. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "2->[0. 0. 1. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "4->[0. 0. 0. 0. 1. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "1->[0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "1->[0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "0->[1. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "9->[0. 0. 0. 0. 0. 0. 0. 0. 0. 1.]\n",
            "Labels of test dataset:\n",
            "4->[0. 0. 0. 0. 1. 0. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "2->[0. 0. 1. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "1->[0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "1->[0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "4->[0. 0. 0. 0. 1. 0. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "3->[0. 0. 0. 1. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "9->[0. 0. 0. 0. 0. 0. 0. 0. 0. 1.]\n",
            "Labels of train dataset:\n",
            "1->[0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "5->[0. 0. 0. 0. 0. 1. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "4->[0. 0. 0. 0. 1. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "9->[0. 0. 0. 0. 0. 0. 0. 0. 0. 1.]\n",
            "Labels of train dataset:\n",
            "3->[0. 0. 0. 1. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "0->[1. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "5->[0. 0. 0. 0. 0. 1. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "6->[0. 0. 0. 0. 0. 0. 1. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "3->[0. 0. 0. 1. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "9->[0. 0. 0. 0. 0. 0. 0. 0. 0. 1.]\n",
            "Labels of train dataset:\n",
            "6->[0. 0. 0. 0. 0. 0. 1. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "0->[1. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "1->[0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "1->[0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "7->[0. 0. 0. 0. 0. 0. 0. 1. 0. 0.]\n",
            "Labels of test dataset:\n",
            "5->[0. 0. 0. 0. 0. 1. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "2->[0. 0. 1. 0. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "9->[0. 0. 0. 0. 0. 0. 0. 0. 0. 1.]\n",
            "Labels of train dataset:\n",
            "8->[0. 0. 0. 0. 0. 0. 0. 0. 1. 0.]\n",
            "Labels of test dataset:\n",
            "7->[0. 0. 0. 0. 0. 0. 0. 1. 0. 0.]\n",
            "Labels of train dataset:\n",
            "6->[0. 0. 0. 0. 0. 0. 1. 0. 0. 0.]\n",
            "Labels of test dataset:\n",
            "3->[0. 0. 0. 1. 0. 0. 0. 0. 0. 0.]\n",
            "Labels of train dataset:\n",
            "9->[0. 0. 0. 0. 0. 0. 0. 0. 0. 1.]\n",
            "Labels of test dataset:\n",
            "4->[0. 0. 0. 0. 1. 0. 0. 0. 0. 0.]\n",
            "The shape of training data is : (60000, 10).\n",
            "The shape of test data is (10000, 10)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### **Build CNN architecture**"
      ],
      "metadata": {
        "id": "LNdOBquzBijh"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model = keras.Sequential()\n",
        "\n",
        "#######################################################\n",
        "#-------------- 1st convolutional layer --------------#\n",
        "#------------------ with 64 filters ------------------#\n",
        "#######################################################\n",
        "model.add(keras.layers.Conv2D(64, \n",
        "                              kernel_size=(3,3), # filter/grid size\n",
        "                              activation='relu', # Rectified Linear Unit used as an activation function\n",
        "                              input_shape=(x_train.shape[1], x_train.shape[2], x_train.shape[3]), # As it is the first layer, needs to enter input shape \n",
        "                              padding='same' # zero padding to apply filters into whole image pixels \n",
        "                              )\n",
        "          )\n",
        "model.add(keras.layers.MaxPool2D((3,3), # grid size\n",
        "                                     strides=(2,2), # strides\n",
        "                                     padding='same'# applies max pooling to whole input\n",
        "                                     ))\n",
        "model.add(keras.layers.BatchNormalization()) # standardizes/normalizes the contributions\n",
        "                                                 # to a layer for every mini-batch\n",
        "\n",
        "#######################################################\n",
        "#-------------- 2nd convolutional layer --------------#\n",
        "#------------------ with 32 filters ------------------#\n",
        "#######################################################\n",
        "model.add(keras.layers.Conv2D(32,\n",
        "                              kernel_size=(3,3),\n",
        "                              activation='relu',\n",
        "                              padding='same'\n",
        "                              )\n",
        "          )\n",
        "model.add(keras.layers.MaxPool2D((3,3),\n",
        "                                     strides=(2,2),\n",
        "                                     padding='same'\n",
        "                                     ))\n",
        "model.add(keras.layers.BatchNormalization())\n",
        "\n",
        "#######################################################\n",
        "#---------------------- Flatten ----------------------#\n",
        "#######################################################\n",
        "model.add(keras.layers.Flatten())\n",
        "\n",
        "#######################################################\n",
        "#------------------ Fully Connected ------------------#\n",
        "#######################################################\n",
        "model.add(keras.layers.Dense(32, activation='relu')) # number of parameters in output:\n",
        "                                                     # (1568:weights + 1:bias)*32:number of units\n",
        "                                                     # = 50208 \n",
        "# to avoid overfitting we add dropout\n",
        "model.add(keras.layers.Dropout(0.3))\n",
        "\n",
        "#######################################################\n",
        "#-------------------- Output Layer -------------------#\n",
        "#######################################################\n",
        "model.add(keras.layers.Dense(10, activation='softmax'))"
      ],
      "metadata": {
        "id": "n7DpTIyUAISd"
      },
      "execution_count": 78,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VsOuLu9SDw4k",
        "outputId": "c76e32f6-d445-4df7-d688-d1936a67a4b8"
      },
      "execution_count": 79,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_3\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " conv2d_5 (Conv2D)           (None, 28, 28, 64)        640       \n",
            "                                                                 \n",
            " max_pooling2d_4 (MaxPooling  (None, 14, 14, 64)       0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " batch_normalization_4 (Batc  (None, 14, 14, 64)       256       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " conv2d_6 (Conv2D)           (None, 14, 14, 32)        18464     \n",
            "                                                                 \n",
            " max_pooling2d_5 (MaxPooling  (None, 7, 7, 32)         0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " batch_normalization_5 (Batc  (None, 7, 7, 32)         128       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " flatten_2 (Flatten)         (None, 1568)              0         \n",
            "                                                                 \n",
            " dense_4 (Dense)             (None, 32)                50208     \n",
            "                                                                 \n",
            " dropout_2 (Dropout)         (None, 32)                0         \n",
            "                                                                 \n",
            " dense_5 (Dense)             (None, 10)                330       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 70,026\n",
            "Trainable params: 69,834\n",
            "Non-trainable params: 192\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Compile the model\n",
        "model.compile(optimizer=keras.optimizers.Adam(learning_rate=0.001), loss=keras.losses.BinaryCrossentropy(), metrics=['accuracy'])"
      ],
      "metadata": {
        "id": "BMQsO43aEVYr"
      },
      "execution_count": 80,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Train the model\n",
        "history = model.fit(x_train, y_train, batch_size=32, epochs=20, validation_data=(x_test, y_test), shuffle=True, verbose =1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "W96JqybjEmxT",
        "outputId": "ddd8977c-aa2a-4416-e803-766bcbc92e23"
      },
      "execution_count": 81,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "1875/1875 [==============================] - 93s 49ms/step - loss: 0.0693 - accuracy: 0.8986 - val_loss: 0.0102 - val_accuracy: 0.9855\n",
            "Epoch 2/20\n",
            "1875/1875 [==============================] - 85s 45ms/step - loss: 0.0223 - accuracy: 0.9720 - val_loss: 0.0078 - val_accuracy: 0.9896\n",
            "Epoch 3/20\n",
            "1875/1875 [==============================] - 81s 43ms/step - loss: 0.0164 - accuracy: 0.9788 - val_loss: 0.0073 - val_accuracy: 0.9890\n",
            "Epoch 4/20\n",
            "1875/1875 [==============================] - 80s 43ms/step - loss: 0.0138 - accuracy: 0.9827 - val_loss: 0.0071 - val_accuracy: 0.9895\n",
            "Epoch 5/20\n",
            "1875/1875 [==============================] - 81s 43ms/step - loss: 0.0121 - accuracy: 0.9845 - val_loss: 0.0064 - val_accuracy: 0.9905\n",
            "Epoch 6/20\n",
            "1875/1875 [==============================] - 80s 43ms/step - loss: 0.0115 - accuracy: 0.9850 - val_loss: 0.0080 - val_accuracy: 0.9898\n",
            "Epoch 7/20\n",
            "1875/1875 [==============================] - 81s 43ms/step - loss: 0.0104 - accuracy: 0.9869 - val_loss: 0.0051 - val_accuracy: 0.9933\n",
            "Epoch 8/20\n",
            "1875/1875 [==============================] - 81s 43ms/step - loss: 0.0092 - accuracy: 0.9884 - val_loss: 0.0055 - val_accuracy: 0.9922\n",
            "Epoch 9/20\n",
            "1875/1875 [==============================] - 80s 43ms/step - loss: 0.0085 - accuracy: 0.9894 - val_loss: 0.0060 - val_accuracy: 0.9926\n",
            "Epoch 10/20\n",
            "1875/1875 [==============================] - 82s 44ms/step - loss: 0.0078 - accuracy: 0.9895 - val_loss: 0.0065 - val_accuracy: 0.9914\n",
            "Epoch 11/20\n",
            "1875/1875 [==============================] - 81s 43ms/step - loss: 0.0077 - accuracy: 0.9902 - val_loss: 0.0053 - val_accuracy: 0.9938\n",
            "Epoch 12/20\n",
            "1875/1875 [==============================] - 80s 43ms/step - loss: 0.0074 - accuracy: 0.9909 - val_loss: 0.0056 - val_accuracy: 0.9928\n",
            "Epoch 13/20\n",
            "1875/1875 [==============================] - 81s 43ms/step - loss: 0.0064 - accuracy: 0.9917 - val_loss: 0.0060 - val_accuracy: 0.9916\n",
            "Epoch 14/20\n",
            "1875/1875 [==============================] - 83s 44ms/step - loss: 0.0063 - accuracy: 0.9919 - val_loss: 0.0058 - val_accuracy: 0.9925\n",
            "Epoch 15/20\n",
            "1875/1875 [==============================] - 81s 43ms/step - loss: 0.0060 - accuracy: 0.9927 - val_loss: 0.0060 - val_accuracy: 0.9935\n",
            "Epoch 16/20\n",
            "1875/1875 [==============================] - 81s 43ms/step - loss: 0.0061 - accuracy: 0.9925 - val_loss: 0.0066 - val_accuracy: 0.9932\n",
            "Epoch 17/20\n",
            "1875/1875 [==============================] - 82s 44ms/step - loss: 0.0059 - accuracy: 0.9932 - val_loss: 0.0072 - val_accuracy: 0.9918\n",
            "Epoch 18/20\n",
            "1875/1875 [==============================] - 80s 43ms/step - loss: 0.0056 - accuracy: 0.9933 - val_loss: 0.0065 - val_accuracy: 0.9930\n",
            "Epoch 19/20\n",
            "1875/1875 [==============================] - 81s 43ms/step - loss: 0.0053 - accuracy: 0.9935 - val_loss: 0.0082 - val_accuracy: 0.9915\n",
            "Epoch 20/20\n",
            "1875/1875 [==============================] - 80s 43ms/step - loss: 0.0049 - accuracy: 0.9940 - val_loss: 0.0061 - val_accuracy: 0.9928\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.figure(figsize=(5,5))\n",
        "plt.plot(history.history['accuracy'], label='Train')\n",
        "plt.plot(history.history['val_accuracy'], label='Test')\n",
        "plt.xlabel('# epochs')\n",
        "plt.ylabel('Accuracy in Model')\n",
        "plt.legend()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 351
        },
        "id": "Btnl3YouFs9s",
        "outputId": "f444ec33-87be-4f6f-cbbc-4fe8bf983856"
      },
      "execution_count": 85,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f795407c850>"
            ]
          },
          "metadata": {},
          "execution_count": 85
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 360x360 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.figure(figsize=(5,5))\n",
        "plt.plot(history.history['loss'], label='Train')\n",
        "plt.plot(history.history['val_loss'], label='Test')\n",
        "plt.xlabel('# epochs')\n",
        "plt.ylabel('Loss in Model')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 351
        },
        "id": "5pdzu8GaHaCj",
        "outputId": "d21c931f-1e6a-42fb-c2ae-a83eaf28d7be"
      },
      "execution_count": 86,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0, 0.5, 'Loss in Model')"
            ]
          },
          "metadata": {},
          "execution_count": 86
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 360x360 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Prediction\n",
        "predictions = model.predict(x_test[:10])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZNhCwE6hHksJ",
        "outputId": "dd03a0a5-b817-4623-cead-97e924e542a5"
      },
      "execution_count": 96,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 0s 23ms/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "predictions"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wQCqMl-NdklT",
        "outputId": "311b02ac-c7df-4b81-d936-e4a4b6552b17"
      },
      "execution_count": 97,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0.00000000e+00, 0.00000000e+00, 3.34200276e-36, 0.00000000e+00,\n",
              "        0.00000000e+00, 0.00000000e+00, 0.00000000e+00, 9.99999940e-01,\n",
              "        0.00000000e+00, 1.11053654e-37],\n",
              "       [4.37298088e-22, 3.37619195e-20, 9.99999940e-01, 1.89883519e-27,\n",
              "        6.43535902e-32, 7.31100533e-28, 1.77164181e-23, 1.82274106e-23,\n",
              "        7.22285144e-21, 3.78643289e-38],\n",
              "       [2.75312220e-33, 9.99999940e-01, 4.79544147e-21, 6.40837936e-27,\n",
              "        1.52076201e-24, 8.70851656e-25, 8.72346290e-21, 3.08348404e-23,\n",
              "        2.23567914e-25, 4.19238567e-29],\n",
              "       [9.99999940e-01, 1.21985417e-28, 1.28849265e-19, 2.83560781e-28,\n",
              "        5.62754752e-20, 9.35022395e-24, 4.23555437e-14, 3.67920566e-21,\n",
              "        2.46672393e-19, 9.81182720e-22],\n",
              "       [6.63591320e-33, 1.25364375e-24, 1.62336492e-22, 9.85451984e-36,\n",
              "        9.99999940e-01, 6.41962628e-37, 6.84530538e-26, 9.76407792e-21,\n",
              "        1.37560570e-21, 1.98968569e-15],\n",
              "       [6.31053782e-36, 9.99999940e-01, 7.07056156e-23, 6.27622060e-32,\n",
              "        4.18261793e-25, 9.07847626e-29, 1.40928975e-23, 8.13108320e-25,\n",
              "        1.67785264e-27, 3.19843596e-31],\n",
              "       [6.88337974e-33, 4.33152724e-21, 5.24877409e-20, 4.17594086e-31,\n",
              "        9.99999940e-01, 1.79871795e-30, 1.23626228e-24, 5.50938849e-18,\n",
              "        1.12216776e-20, 3.30442609e-15],\n",
              "       [4.02734283e-14, 3.03935690e-20, 8.55064209e-17, 8.01034933e-11,\n",
              "        1.34705058e-09, 2.68247447e-12, 9.20288414e-22, 7.02144787e-10,\n",
              "        2.44840155e-11, 9.99999940e-01],\n",
              "       [3.45246970e-14, 6.12473024e-20, 4.10850688e-13, 1.12551957e-09,\n",
              "        1.72503681e-13, 9.99997497e-01, 1.69550319e-07, 2.67704064e-15,\n",
              "        2.34357844e-06, 4.95859298e-10],\n",
              "       [8.34131333e-28, 1.40414340e-31, 5.01688011e-30, 9.42333425e-26,\n",
              "        4.22020105e-14, 1.12153834e-28, 9.19137222e-38, 4.77025341e-14,\n",
              "        4.83576410e-21, 1.00000000e+00]], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 97
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "predictions = np.argmax(predictions,axis=1)"
      ],
      "metadata": {
        "id": "JuRk_FsrdNgj"
      },
      "execution_count": 98,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "predictions"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HZekhzoDdP2x",
        "outputId": "1b566177-9390-4068-a0a6-494a84c709d3"
      },
      "execution_count": 99,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([7, 2, 1, 0, 4, 1, 4, 9, 5, 9])"
            ]
          },
          "metadata": {},
          "execution_count": 99
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "fig = plt.figure(figsize=(5,5))\n",
        "for k in range(len(predictions)):\n",
        "  fig.add_subplot(5,2,k+1)\n",
        "  plt.imshow(x_test[k], cmap='gray')\n",
        "  plt.title(f'prediction:{predictions[k]}')\n",
        "  fig.tight_layout(pad=0.2)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 387
        },
        "id": "xNpVIAifdgzf",
        "outputId": "ffacdfb7-0ff6-4cfd-99db-28c7e8b047c0"
      },
      "execution_count": 103,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 360x360 with 10 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}