Exercise_2_Handwriting_Recognition_Question.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Exercise 2 Handwriting Recognition Question.ipynb",
      "provenance": [],
      "authorship_tag": "ABX9TyOWbcdeJrqFtOuwNTTeB23x",
      "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/JooVLC/Coursera-TensorFlow-in-practice-specialization/blob/main/Exercise_2_Handwriting_Recognition_Question.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "B4_B2pVHY99v",
        "outputId": "2eedab86-5695-4bfa-b4a2-b73bca0dc045"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "2.7.0\n"
          ]
        }
      ],
      "source": [
        "import tensorflow as tf\n",
        "import numpy as np\n",
        "from tensorflow import keras\n",
        "\n",
        "print(tf.__version__)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "class tf_stop_training_after_hit_accuracy_callback(keras.callbacks.Callback):\n",
        "  def __init__(self, desired_accuracy):\n",
        "    self.desired_accuracy = desired_accuracy\n",
        "    super().__init__()\n",
        "\n",
        "  def on_epoch_end(self, epoch, logs={}):\n",
        "    current_epoch_accuracy = logs.get('accuracy')\n",
        "    if current_epoch_accuracy is not None and current_epoch_accuracy >= self.desired_accuracy:\n",
        "      self.model.stop_training = True\n"
      ],
      "metadata": {
        "id": "6-ZWD6R_m4LB"
      },
      "execution_count": 68,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "selected_dataset = keras.datasets.mnist"
      ],
      "metadata": {
        "id": "Y1ORLKHwreG4"
      },
      "execution_count": 55,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "(train_images, train_labels), (test_images, test_labels) = selected_dataset.load_data()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5OJ3l7olaBVj",
        "outputId": "4258bb33-51ef-437d-b013-f53a7f2d806c"
      },
      "execution_count": 56,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/mnist.npz\n",
            "11493376/11490434 [==============================] - 0s 0us/step\n",
            "11501568/11490434 [==============================] - 0s 0us/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "plt.imshow(train_images[30])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "caKhdyK8c5OX",
        "outputId": "0ce7db85-fd52-47f6-f23e-e34d833e4d38"
      },
      "execution_count": 58,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7fa01f3b5ed0>"
            ]
          },
          "metadata": {},
          "execution_count": 58
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAPsAAAD4CAYAAAAq5pAIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAANzElEQVR4nO3dXawc9XnH8d/Prl+KjYMdjHEdNxDeAoXGaU8cCiglQqUOTWWoGgoXCU3dnFxAFaSoDUoubOXKggRUkTTCARQnolCqQOHCbUOtUIKSEA4vMTYOmFIMdg824LbYNH7l6cUZRwc4+9/j3dmdxc/3I612d56dnUdj/87Mzuzs3xEhAEe/KU03AKA/CDuQBGEHkiDsQBKEHUji1/q5sOmeETM1q5+LBFLZqze0P/Z5olpXYbe9TNLfSpoq6daIWF16/UzN0kd9UTeLBFDwSKxvWet4N972VEnflPQJSWdJutL2WZ2+H4De6uYz+1JJz0XE8xGxX9JdkpbX0xaAunUT9kWSXhr3fFs17S1sD9sesT1yQPu6WByAbvT8aHxErImIoYgYmqYZvV4cgBa6Cft2SYvHPX9fNQ3AAOom7I9KOs32ybanS7pC0v31tAWgbh2feouIg7avkfSvGjv1dntEbKqtMwC16uo8e0Ssk7Supl4A9BBflwWSIOxAEoQdSIKwA0kQdiAJwg4kQdiBJAg7kARhB5Ig7EAShB1IgrADSRB2IAnCDiRB2IEkCDuQBGEHkiDsQBKEHUiCsANJEHYgCcIOJEHYgSQIO5AEYQeSIOxAEoQdSIKwA0kQdiAJwg4k0dWQzbZfkLRb0iFJByNiqI6mANSvq7BXPh4Rr9bwPgB6iN14IIluwx6SfmD7MdvDE73A9rDtEdsjB7Svy8UB6FS3u/EXRMR22ydIesD2LyLiofEviIg1ktZI0hzPiy6XB6BDXW3ZI2J7db9T0r2SltbRFID6dRx227NsH3v4saSLJW2sqzEA9epmN36BpHttH36fv4+If6mlKwC16zjsEfG8pA/V2AuAHuLUG5AEYQeSIOxAEoQdSIKwA0nUcSHMUWHqce8p1j1rVsvatk+dVJz39XP2d9LSUeGDN+1pWXtz4y/62AnYsgNJEHYgCcIOJEHYgSQIO5AEYQeSIOxAEkfNefadV59XrL9+7i+L9RUf+nGx/tfvffqIe4L03fMWtazds+wjxXkPbn2p7nZSY8sOJEHYgSQIO5AEYQeSIOxAEoQdSIKwA0kcNefZn/jK3xXrB+JQV/V735h/xD0d9tWNnyzW33j1mGJ99rPTOl52t/acWb4W/9lltxTrn5mzvWXthr/4k+K871/JefY6sWUHkiDsQBKEHUiCsANJEHYgCcIOJEHYgSSOmvPst/3vicX6rkOtf/ddku66+eJi/fhbfnLEPR22SJs6nrfXpp5xarH+4nkzerbs2S9Gz94b79R2y277dts7bW8cN22e7Qdsb6nu5/a2TQDdmsxu/HckLXvbtOskrY+I0yStr54DGGBtwx4RD0na9bbJyyWtrR6vlXRpzX0BqFmnn9kXRMRo9fhlSQtavdD2sKRhSZqp8nfAAfRO10fjIyIktTzSEhFrImIoIoamqXcHewCUdRr2HbYXSlJ1v7O+lgD0Qqdhv1/SVdXjqyTdV087AHql7Wd223dKulDS8ba3SVopabWku22vkLRV0uW9bHIy7j6zfJ69nePV+Xn0QTbl7A8W60u+V/49/PtOeKKr5V+08U9b1k74p2eK85Z/YQBHqm3YI+LKFqWLau4FQA/xdVkgCcIOJEHYgSQIO5AEYQeSOGoucT2aTZ0zp1h/9bLfalm7aeU3i/MundHby0xnrmrd+6HX/rOny8ZbsWUHkiDsQBKEHUiCsANJEHYgCcIOJEHYgSQ4z/4usPn68mWqz/7xN/rUyZFbdNPzLWuj/7eoq/fe8vPFxfoZt/53y9qhTeXLa49GbNmBJAg7kARhB5Ig7EAShB1IgrADSRB2IAnOs78LLD75laZb6NiaxQ/27s3PKJfPP/2KlrV5y8v/9ePgwU46Gmhs2YEkCDuQBGEHkiDsQBKEHUiCsANJEHYgCUf09nfDx5vjefFRM/jrETv3t4vlvfNn9qmRd3pjwdRi/dg/+6/W897xG8V5d/+mi/Uffe6GYv09U1qvl2u2X1Ccd+v55QGj48D+Yr0pj8R6vR67Jlxxbbfstm+3vdP2xnHTVtnebvvJ6nZJnQ0DqN9kduO/I2nZBNNviogl1W1dvW0BqFvbsEfEQ5J29aEXAD3UzQG6a2xvqHbz57Z6ke1h2yO2Rw5oXxeLA9CNTsP+LUmnSFoiaVTS11u9MCLWRMRQRAxN04wOFwegWx2FPSJ2RMShiHhT0rclLa23LQB16yjstheOe3qZpI2tXgtgMLS9nt32nZIulHS87W2SVkq60PYSSSHpBUmf72GP+OmGYrmXZ9n3/dFHivVz/rL8d37HZ09sWZu++SfFeVseCKpc8aO/KtaX3jjSsvaNRQ8X5/3k7362vPA2/yaDqG3YI+LKCSbf1oNeAPQQX5cFkiDsQBKEHUiCsANJEHYgCS5xTW7/svKptdO/Wj619szKs4v1Gf/86BH3VJcXV53XsrbhczcX531o7/Ri/fpTzumop17r6hJXAEcHwg4kQdiBJAg7kARhB5Ig7EAShB1IgiGbk9t/7WvF+vzpe4r1bU+8WKw3OfDxgp8daFnbs6L8E2kfa3Pd8PWdNNQwtuxAEoQdSIKwA0kQdiAJwg4kQdiBJAg7kATn2ZM7/bhXivWV858s1j/1D22GXb6x9fghv37fz4rztvPq8O8V6wcv+Z+WtdlT8o1OxJYdSIKwA0kQdiAJwg4kQdiBJAg7kARhB5LgPHtyDz59RvkFix8slv/x1HXF+ujNv2xZe/SG8jn6dv7wmJ8W6zM8reP3HvpaeTjoE/Xjjt+7KW237LYX2/6h7adtb7L9hWr6PNsP2N5S3bcbThtAgyazG39Q0hcj4ixJ50q62vZZkq6TtD4iTpO0vnoOYEC1DXtEjEbE49Xj3ZI2S1okabmktdXL1kq6tFdNAujeEX1mt32SpA9LekTSgogYrUovS1rQYp5hScOSNFPHdNongC5N+mi87dmSvi/p2oh4fXwtxkaHnHCEyIhYExFDETE0TfkuPgAGxaTCbnuaxoJ+R0TcU03eYXthVV8oaWdvWgRQh7ZDNtu2xj6T74qIa8dNv0HSaxGx2vZ1kuZFxN+U3oshmwfPlGPKH6223Hp6sb7592+rs52+OfPfVxTrp36mPFR1HGzyR7JbKw3ZPJnP7OdL+rSkp2wfvrj5y5JWS7rb9gpJWyVdXkezAHqjbdgj4mFJE/6lkMRmGniX4OuyQBKEHUiCsANJEHYgCcIOJNH2PHudOM/+7jNlZnns4ilzjyvWn7325Ja1g/N6e6567mOtTzbNv6V8eaz6mIs6lc6zs2UHkiDsQBKEHUiCsANJEHYgCcIOJEHYgST4KWkUvbl3b7k++nKx/oEvlevoH7bsQBKEHUiCsANJEHYgCcIOJEHYgSQIO5AEYQeSIOxAEoQdSIKwA0kQdiAJwg4kQdiBJAg7kETbsNtebPuHtp+2vcn2F6rpq2xvt/1kdbuk9+0C6NRkfrzioKQvRsTjto+V9JjtB6raTRHxtd61B6AukxmffVTSaPV4t+3Nkhb1ujEA9Tqiz+y2T5L0YUmPVJOusb3B9u2257aYZ9j2iO2RA9rXVbMAOjfpsNueLen7kq6NiNclfUvSKZKWaGzL//WJ5ouINRExFBFD0zSjhpYBdGJSYbc9TWNBvyMi7pGkiNgREYci4k1J35a0tHdtAujWZI7GW9JtkjZHxI3jpi8c97LLJG2svz0AdZnM0fjzJX1a0lO2n6ymfVnSlbaXSApJL0j6fE86BFCLyRyNf1jSROM9r6u/HQC9wjfogCQIO5AEYQeSIOxAEoQdSIKwA0kQdiAJwg4kQdiBJAg7kARhB5Ig7EAShB1IgrADSTgi+rcw+xVJW8dNOl7Sq31r4MgMam+D2pdEb52qs7f3R8T8iQp9Dfs7Fm6PRMRQYw0UDGpvg9qXRG+d6ldv7MYDSRB2IImmw76m4eWXDGpvg9qXRG+d6ktvjX5mB9A/TW/ZAfQJYQeSaCTstpfZfsb2c7ava6KHVmy/YPupahjqkYZ7ud32Ttsbx02bZ/sB21uq+wnH2Guot4EYxrswzHij667p4c/7/pnd9lRJz0r6A0nbJD0q6cqIeLqvjbRg+wVJQxHR+BcwbH9M0h5J342Is6tp10vaFRGrqz+UcyPiSwPS2ypJe5oexrsarWjh+GHGJV0q6c/V4Lor9HW5+rDemtiyL5X0XEQ8HxH7Jd0laXkDfQy8iHhI0q63TV4uaW31eK3G/rP0XYveBkJEjEbE49Xj3ZIODzPe6Lor9NUXTYR9kaSXxj3fpsEa7z0k/cD2Y7aHm25mAgsiYrR6/LKkBU02M4G2w3j309uGGR+YddfJ8Ofd4gDdO10QEb8j6ROSrq52VwdSjH0GG6Rzp5MaxrtfJhhm/FeaXHedDn/erSbCvl3S4nHP31dNGwgRsb263ynpXg3eUNQ7Do+gW93vbLifXxmkYbwnGmZcA7Dumhz+vImwPyrpNNsn254u6QpJ9zfQxzvYnlUdOJHtWZIu1uANRX2/pKuqx1dJuq/BXt5iUIbxbjXMuBped40Pfx4Rfb9JukRjR+T/Q9JXmuihRV8fkPTz6rap6d4k3amx3boDGju2sULSeyWtl7RF0r9JmjdAvX1P0lOSNmgsWAsb6u0Cje2ib5D0ZHW7pOl1V+irL+uNr8sCSXCADkiCsANJEHYgCcIOJEHYgSQIO5AEYQeS+H8C1yeXqBsAMwAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train_images_normalized = train_images / 255.0\n",
        "test_images_normalized = test_images / 255.0"
      ],
      "metadata": {
        "id": "I3esx7ZVc86m"
      },
      "execution_count": 59,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model = keras.Sequential()\n",
        "model.add(keras.layers.Flatten())\n",
        "model.add(keras.layers.Dense(units=512, activation=tf.nn.relu))\n",
        "model.add(keras.layers.Dense(units=10, activation=tf.nn.softmax))"
      ],
      "metadata": {
        "id": "UQT-zVpcezd2"
      },
      "execution_count": 63,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "callback_to_stop_training = tf_stop_training_after_hit_accuracy_callback(0.99)"
      ],
      "metadata": {
        "id": "5FWJtnaMoeyN"
      },
      "execution_count": 69,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model.compile(optimizer=tf.optimizers.Adam(), loss='sparse_categorical_crossentropy', metrics=['accuracy'])\n",
        "model.fit(train_images_normalized, train_labels, epochs=10000, callbacks=[callback_to_stop_training])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "o9SUCPzgf7zh",
        "outputId": "e61cc507-ae10-480b-cbf8-41ff41c4a683"
      },
      "execution_count": 70,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10000\n",
            "1875/1875 [==============================] - 13s 7ms/step - loss: 0.0203 - accuracy: 0.9937\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7fa01f4ab090>"
            ]
          },
          "metadata": {},
          "execution_count": 70
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.evaluate(test_images_normalized, test_labels)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nDEskADfg8_j",
        "outputId": "b7ef9eb0-338c-42b9-8a3c-5cdb46969e73"
      },
      "execution_count": 71,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 1s 4ms/step - loss: 0.0631 - accuracy: 0.9818\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.06310193240642548, 0.9818000197410583]"
            ]
          },
          "metadata": {},
          "execution_count": 71
        }
      ]
    }
  ]
}