mid_project.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "05. mnist dataset hand written digits classification.ipynb",
      "provenance": [],
      "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/shakib04/practice_cvpr/blob/main/mid_project.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "14OLb6rhr8Eh"
      },
      "source": [
        "Program to classify the hand written numbers"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "F31_kEQTsY3y"
      },
      "source": [
        "import tensorflow as tf\n",
        "#from tensorflow import keras\n",
        "#from tensorflow.keras.datasets import mnist\n",
        "\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt"
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-9ksYE_ps0K-",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6d44e17a-bc07-4845-8f21-bbb69ceb6dbb"
      },
      "source": [
        "(train_imgs, train_labels), (test_imgs, test_labels) = tf.keras.datasets.mnist.load_data()"
      ],
      "execution_count": 2,
      "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",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ST1VrT2DtUI3",
        "outputId": "2b693b30-828d-4a62-ec38-acba2c5547d5"
      },
      "source": [
        "print(f'Train img shape {train_imgs.shape}')\n",
        "print(f'Train Label shape {train_labels.shape}')"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Train img shape (60000, 28, 28)\n",
            "Train Label shape (60000,)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "TqdR0BH0tfFp",
        "outputId": "5a78f6fb-71be-478e-9eb4-1d2dbe8b0e4a"
      },
      "source": [
        "i = 98\n",
        "plt.figure()\n",
        "plt.imshow(train_imgs[i], cmap=plt.cm.gray_r)\n",
        "plt.show()\n",
        "print(f'Image Label: {train_labels[i]}')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAPsAAAD4CAYAAAAq5pAIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAANi0lEQVR4nO3db6xU9Z3H8c9ntX0gkPDn3hC0RGrjE7JmKY6kSUll0ywBjcE+ISXGsAmBaiC2sWqNG1OiDzSbhVrNpglVAtWuFQNGEqGLizXIk8bRoCBklSVAQYSLJhbiA5R+98E9NFe8c+YyZ/7h9/1KJjNzvnPmfHO4H87M+c3MzxEhAF9//9DrBgB0B2EHkiDsQBKEHUiCsANJXNnNjQ0MDMSMGTO6uUkglcOHD+v06dMerVYp7LYXSPq1pCskPR0Rj5c9fsaMGarX61U2CaBErVZrWGv5ZbztKyT9p6SFkmZKWmJ7ZqvPB6CzqrxnnyPpYEQciohzkv4gaVF72gLQblXCfo2kv4y4f6xY9iW2V9iu264PDQ1V2ByAKjp+Nj4i1kVELSJqg4ODnd4cgAaqhP24pOkj7n+rWAagD1UJ+5uSrrf9bdvflPRjSVvb0xaAdmt56C0ivrC9StJ/a3jobX1EvNe2zgC0VaVx9ojYJmlbm3oB0EF8XBZIgrADSRB2IAnCDiRB2IEkCDuQBGEHkiDsQBKEHUiCsANJEHYgCcIOJEHYgSQIO5AEYQeSIOxAEoQdSIKwA0kQdiAJwg4kQdiBJAg7kARhB5Ig7EAShB1IgrADSRB2IAnCDiRB2IEkKs3iiu749NNPS+vPPvtsw9qjjz5auu7Q0FBpPSJK67feemtpfdGiRQ1ry5cvL10X7VUp7LYPSzoj6bykLyKi1o6mALRfO47s/xwRp9vwPAA6iPfsQBJVwx6Sdth+y/aK0R5ge4Xtuu16s/eHADqnatjnRsRsSQslrbT9g4sfEBHrIqIWEbXBwcGKmwPQqkphj4jjxfUpSS9JmtOOpgC0X8thtz3O9oQLtyXNl7SvXY0BaK8qZ+OnSnrJ9oXn+a+I+GNbukrm4MGDpfWFCxeW1g8dOtTytot/v5Zt3769tP7OO+80rN14442l686ePbulnjC6lsMeEYck/VMbewHQQQy9AUkQdiAJwg4kQdiBJAg7kARfce2CF198sbT+wAMPlNaPHj1aWh8YGGhYu+OOO0rXbTas99lnn5XWV61aVVr/8MMPG9aefPLJ0nU3bNhQWsel4cgOJEHYgSQIO5AEYQeSIOxAEoQdSIKwA0kwzt4FR44cqVSfMmVKaf2FF15oWJs3b17pulVt2bKltP7cc891dPsYO47sQBKEHUiCsANJEHYgCcIOJEHYgSQIO5AE4+xdcPfdd5fWz549W1q/5557SuuTJ0++5J7G6tSpU6X1zZs3t/zctRqT/nYTR3YgCcIOJEHYgSQIO5AEYQeSIOxAEoQdSIJx9i4YN25caX316tXdaaQFr7zySmm92e/Kz58/v2Gt2W/Oo72aHtltr7d9yva+Ecsm237V9gfF9aTOtgmgqrG8jN8gacFFyx6UtDMirpe0s7gPoI81DXtE7JL0yUWLF0naWNzeKOn2NvcFoM1aPUE3NSJOFLc/kjS10QNtr7Bdt10fGhpqcXMAqqp8Nj4iQlKU1NdFRC0iaoODg1U3B6BFrYb9pO1pklRcl381CkDPtRr2rZKWFreXSnq5Pe0A6JSm4+y2n5c0T9KA7WOSfinpcUmbbC+TdETS4k42ida9//77pfU33nijtL58+fLS+lVXXVVaX7yYP41+0TTsEbGkQemHbe4FQAfxcVkgCcIOJEHYgSQIO5AEYQeS4Cuul4EzZ86U1u+7776GtbLpnMfy3M1ceWX5n9D58+db3vaECRNa6gmj48gOJEHYgSQIO5AEYQeSIOxAEoQdSIKwA0kwzn4Z2LRpU2n96aef7lInX9VsrPyuu+5qWFu7dm3purt37y6tT5kypbSOL+PIDiRB2IEkCDuQBGEHkiDsQBKEHUiCsANJMM5+GVi4cGFp/dprr21YGz9+fKXnvvfee0vr27dvL60/8sgjDWvNfub6qaeeKq3381TX/YgjO5AEYQeSIOxAEoQdSIKwA0kQdiAJwg4k4Yjo2sZqtVrU6/WubQ+9t2PHjoa1BQsWVHru1157rbQ+b968Ss9/OarVaqrX6x6t1vTIbnu97VO2941Yttr2cdt7isst7WwYQPuN5WX8Bkmj/Rf8q4iYVVy2tbctAO3WNOwRsUvSJ13oBUAHVTlBt8r2u8XL/EmNHmR7he267frQ0FCFzQGootWw/0bSdyTNknRC0ppGD4yIdRFRi4ja4OBgi5sDUFVLYY+IkxFxPiL+Jum3kua0ty0A7dZS2G1PG3H3R5L2NXosgP7Q9Pvstp+XNE/SgO1jkn4paZ7tWZJC0mFJP+lgj7iM3XDDDQ1rZd/Dl6SjR4+W1pt9ZiPjOHuZpmGPiCWjLH6mA70A6CA+LgskQdiBJAg7kARhB5Ig7EAS/JQ0OmratGkNaxMnTixdt9nQGy4NR3YgCcIOJEHYgSQIO5AEYQeSIOxAEoQdSIKwA0kQdiAJwg4kQdiBJAg7kARhB5Ig7EAShB1Igu+zo6Nef/31hrUDBw50rxFwZAeyIOxAEoQdSIKwA0kQdiAJwg4kQdiBJBhnRyWff/55aX3btm0Na+fOnau07ZtvvrnS+tk0PbLbnm77T7b3237P9k+L5ZNtv2r7g+J6UufbBdCqsbyM/0LSzyNipqTvSVppe6akByXtjIjrJe0s7gPoU03DHhEnIuLt4vYZSQckXSNpkaSNxcM2Srq9U00CqO6STtDZniHpu5L+LGlqRJwoSh9JmtpgnRW267brQ0NDFVoFUMWYw257vKTNkn4WEX8dWYuIkBSjrRcR6yKiFhG1wcHBSs0CaN2Ywm77GxoO+u8jYkux+KTtaUV9mqRTnWkRQDs0HXqzbUnPSDoQEWtHlLZKWirp8eL65Y50mECz4av9+/eX1idMmNCwdt1117XU0wXNenv44YdL62vWrGlYG/7Tamzp0qWl9Ztuuqm0ji8byzj79yXdKWmv7T3Fsoc0HPJNtpdJOiJpcWdaBNAOTcMeEbslNfov+IftbQdAp/BxWSAJwg4kQdiBJAg7kARhB5LgK6594OOPPy6tz549u7Q+ZcqUhrUnnniidN2BgYHS+mOPPVZa37VrV2m9zNVXX11aX7t2bWkdl4YjO5AEYQeSIOxAEoQdSIKwA0kQdiAJwg4kwTj710DZOP2dd95Z6bmHf4SosbIxfklauXJlw9qyZctK1504cWJpHZeGIzuQBGEHkiDsQBKEHUiCsANJEHYgCcIOJME4ex8YN25caX3mzJml9Wa/K1/F/fffX1q/7bbbSutz585tZzuogCM7kARhB5Ig7EAShB1IgrADSRB2IAnCDiQxlvnZp0v6naSpkkLSuoj4te3VkpZLGioe+lBEbOtUo19nZfOrS9LevXu71Am+zsbyoZovJP08It62PUHSW7ZfLWq/ioj/6Fx7ANplLPOzn5B0orh9xvYBSdd0ujEA7XVJ79ltz5D0XUl/Lhatsv2u7fW2JzVYZ4Xtuu360NDQaA8B0AVjDrvt8ZI2S/pZRPxV0m8kfUfSLA0f+deMtl5ErIuIWkTUBgcH29AygFaMKey2v6HhoP8+IrZIUkScjIjzEfE3Sb+VNKdzbQKoqmnYbVvSM5IORMTaEcunjXjYjyTta397ANplLGfjvy/pTkl7be8plj0kaYntWRoejjss6Scd6RBAW4zlbPxuSR6lxJg6cBnhE3RAEoQdSIKwA0kQdiAJwg4kQdiBJAg7kARhB5Ig7EAShB1IgrADSRB2IAnCDiRB2IEkHBHd25g9JOnIiEUDkk53rYFL06+99WtfEr21qp29XRsRo/7+W1fD/pWN2/WIqPWsgRL92lu/9iXRW6u61Rsv44EkCDuQRK/Dvq7H2y/Tr731a18SvbWqK7319D07gO7p9ZEdQJcQdiCJnoTd9gLb/2v7oO0He9FDI7YP295re4/teo97WW/7lO19I5ZNtv2q7Q+K61Hn2OtRb6ttHy/23R7bt/Sot+m2/2R7v+33bP+0WN7TfVfSV1f2W9ffs9u+QtL7kv5F0jFJb0paEhH7u9pIA7YPS6pFRM8/gGH7B5LOSvpdRPxjsezfJX0SEY8X/1FOiohf9ElvqyWd7fU03sVsRdNGTjMu6XZJ/6oe7ruSvharC/utF0f2OZIORsShiDgn6Q+SFvWgj74XEbskfXLR4kWSNha3N2r4j6XrGvTWFyLiRES8Xdw+I+nCNOM93XclfXVFL8J+jaS/jLh/TP0133tI2mH7Ldsret3MKKZGxIni9keSpvaymVE0nca7my6aZrxv9l0r059XxQm6r5obEbMlLZS0sni52pdi+D1YP42djmka724ZZZrxv+vlvmt1+vOqehH245Kmj7j/rWJZX4iI48X1KUkvqf+moj55YQbd4vpUj/v5u36axnu0acbVB/uul9Of9yLsb0q63va3bX9T0o8lbe1BH19he1xx4kS2x0mar/6binqrpKXF7aWSXu5hL1/SL9N4N5pmXD3edz2f/jwiun6RdIuGz8j/n6R/60UPDfq6TtI7xeW9Xvcm6XkNv6z7XMPnNpZJmiJpp6QPJP2PpMl91NuzkvZKelfDwZrWo97mavgl+ruS9hSXW3q970r66sp+4+OyQBKcoAOSIOxAEoQdSIKwA0kQdiAJwg4kQdiBJP4fonYezIy9nJ0AAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Image Label: 3\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 228
        },
        "id": "GuHIHBonuUFC",
        "outputId": "99652702-a400-468d-e269-0c601369fab8"
      },
      "source": [
        "R = 5\n",
        "C = 20\n",
        "i = 0\n",
        "\n",
        "plt.figure()\n",
        "for r in range(R):\n",
        "  for c in range(C):\n",
        "    plt.subplot(R, C, i + 1)\n",
        "    plt.imshow(train_imgs[i])\n",
        "    #remove extra info\n",
        "    plt.xticks([])\n",
        "    plt.yticks([])\n",
        "    i += 1"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 100 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "v_hZyfEfv6FW"
      },
      "source": [
        "#convert numpy 2d array to 1d array # astype # normalization by diving 255\n",
        "train_imgs_reshaped = train_imgs.reshape(-1, 784).astype('float32') / 255\n",
        "test_imgs_reshaped = test_imgs.reshape(-1, 784).astype('float32') / 255\n",
        "#we are losing some information by converting 2d to 1d"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "b4wmcjL2zMpj"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3qKXwuqLw-FN",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6856cfd8-59e1-4168-dbd5-dc4b65eef425"
      },
      "source": [
        "#step: 2 #Sequential layers create hobe (one layer is connected with previous layer)\n",
        "model = tf.keras.models.Sequential([\n",
        "       tf.keras.Input(shape= 28 * 28),\n",
        "       tf.keras.layers.Dense(units=256, activation='relu'),\n",
        "       tf.keras.layers.Dense(units=128, activation='relu'),\n",
        "       tf.keras.layers.Dense(units=10, activation='softmax') \n",
        "])\n",
        "\n",
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_3\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense_7 (Dense)              (None, 256)               200960    \n",
            "_________________________________________________________________\n",
            "dense_8 (Dense)              (None, 128)               32896     \n",
            "_________________________________________________________________\n",
            "dense_9 (Dense)              (None, 10)                1290      \n",
            "=================================================================\n",
            "Total params: 235,146\n",
            "Trainable params: 235,146\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WPc_glTmcqPh"
      },
      "source": [
        "ADAM Optimizer"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1tcxruan15JH"
      },
      "source": [
        "\n",
        "model.compile(\n",
        "    optimizer='adam', \n",
        "    loss='sparse_categorical_crossentropy', \n",
        "    metrics=['accuracy']\n",
        ")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4099bme91-yo",
        "outputId": "8ec05511-5a79-454b-b35b-cca49edfd648"
      },
      "source": [
        "tf.config.experimental_run_functions_eagerly(True)\n",
        "h = model.fit(x=train_imgs_reshaped, y=train_labels, epochs=5, batch_size=64, validation_split=0.3)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "WARNING:tensorflow:From <ipython-input-27-fbf08a043493>:1: experimental_run_functions_eagerly (from tensorflow.python.eager.def_function) is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "Use `tf.config.run_functions_eagerly` instead of the experimental version.\n",
            "Epoch 1/5\n",
            "  4/657 [..............................] - ETA: 11s - loss: 2.1676 - accuracy: 0.2461"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/tensorflow/python/data/ops/dataset_ops.py:4212: UserWarning: Even though the `tf.config.experimental_run_functions_eagerly` option is set, this option does not apply to tf.data functions. To force eager execution of tf.data functions, please use `tf.data.experimental.enable_debug_mode()`.\n",
            "  \"Even though the `tf.config.experimental_run_functions_eagerly` \"\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "657/657 [==============================] - 14s 21ms/step - loss: 0.2712 - accuracy: 0.9213 - val_loss: 0.1443 - val_accuracy: 0.9566\n",
            "Epoch 2/5\n",
            "657/657 [==============================] - 13s 20ms/step - loss: 0.1048 - accuracy: 0.9692 - val_loss: 0.1107 - val_accuracy: 0.9666\n",
            "Epoch 3/5\n",
            "657/657 [==============================] - 13s 20ms/step - loss: 0.0652 - accuracy: 0.9804 - val_loss: 0.1032 - val_accuracy: 0.9701\n",
            "Epoch 4/5\n",
            "657/657 [==============================] - 13s 19ms/step - loss: 0.0483 - accuracy: 0.9850 - val_loss: 0.1025 - val_accuracy: 0.9693\n",
            "Epoch 5/5\n",
            "657/657 [==============================] - 13s 20ms/step - loss: 0.0345 - accuracy: 0.9885 - val_loss: 0.1043 - val_accuracy: 0.9711\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 334
        },
        "id": "UtPSwuYN2gMb",
        "outputId": "ec47518d-03f6-4399-9117-802d581bc8f7"
      },
      "source": [
        "plt.figure(figsize=(15,5))\n",
        "plt.subplot(1,2,1)\n",
        "plt.plot(h.history['accuracy'], 'o-', label='train accuracy')\n",
        "plt.plot(h.history['val_accuracy'], 'o-', label = 'validation accuracy')\n",
        "plt.xlabel('Epoch')\n",
        "plt.ylabel('Accuracy')\n",
        "plt.grid(True)\n",
        "plt.legend(loc='lower right')\n",
        "\n",
        "plt.subplot(1,2,2)\n",
        "plt.plot(h.history['loss'], 'o-', label='train loss')\n",
        "plt.plot(h.history['val_loss'], 'o-', label='validation loss')\n",
        "plt.xlabel('Epoch')\n",
        "plt.ylabel('Loss')\n",
        "plt.grid(True)\n",
        "plt.legend(loc='upper right')\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3SMYwjRJ4v6H"
      },
      "source": [
        "Evaluate the model on the test dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LwMiELhy4nvH",
        "outputId": "d58663b6-d359-4600-8ea4-145cce13e71b"
      },
      "source": [
        "test_loss, test_acc = model.evaluate(test_imgs_reshaped, test_labels)\n",
        "print('\\nTest accuracy:', test_acc)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            " 16/313 [>.............................] - ETA: 2s - loss: 0.0685 - accuracy: 0.9805"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/tensorflow/python/data/ops/dataset_ops.py:4212: UserWarning: Even though the `tf.config.experimental_run_functions_eagerly` option is set, this option does not apply to tf.data functions. To force eager execution of tf.data functions, please use `tf.data.experimental.enable_debug_mode()`.\n",
            "  \"Even though the `tf.config.experimental_run_functions_eagerly` \"\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 2s 7ms/step - loss: 0.0830 - accuracy: 0.9761\n",
            "\n",
            "Test accuracy: 0.9761000275611877\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uLqwcwFk5jfk"
      },
      "source": [
        "Model with differnt parameters(**sigmoid activation** on output layer)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LneRmxlF5DmV",
        "outputId": "98cf617b-aa82-4e80-9e3f-5cad3a0710b4"
      },
      "source": [
        "\n",
        "model = tf.keras.models.Sequential([\n",
        "       tf.keras.Input(shape= 28 * 28),\n",
        "       tf.keras.layers.Dense(units=256, activation='relu'),\n",
        "       tf.keras.layers.Dense(units=128, activation='relu'),\n",
        "       tf.keras.layers.Dense(units=10, activation='sigmoid') \n",
        "])\n",
        "\n",
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_5\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense_14 (Dense)             (None, 256)               200960    \n",
            "_________________________________________________________________\n",
            "dense_15 (Dense)             (None, 128)               32896     \n",
            "_________________________________________________________________\n",
            "dense_16 (Dense)             (None, 10)                1290      \n",
            "=================================================================\n",
            "Total params: 235,146\n",
            "Trainable params: 235,146\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "n5N0t1QWdp-0"
      },
      "source": [
        "SGD optimizer"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZcHBvqLH51gG"
      },
      "source": [
        "model.compile(\n",
        "    optimizer='SGD', \n",
        "    loss='sparse_categorical_crossentropy', \n",
        "    metrics=['accuracy']\n",
        ")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-eKe3T6ed_fu"
      },
      "source": [
        "99.76% accuracy"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fNz6_SkK6f97",
        "outputId": "fcc26f27-0053-48e8-bf02-9a04f55f7215"
      },
      "source": [
        "tf.config.experimental_run_functions_eagerly(True)\n",
        "h = model.fit(x=train_imgs_reshaped, y=train_labels, epochs=5, batch_size=64, validation_split=0.3)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/5\n",
            "  9/657 [..............................] - ETA: 9s - loss: 0.0304 - accuracy: 0.9844"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/tensorflow/python/data/ops/dataset_ops.py:4212: UserWarning: Even though the `tf.config.experimental_run_functions_eagerly` option is set, this option does not apply to tf.data functions. To force eager execution of tf.data functions, please use `tf.data.experimental.enable_debug_mode()`.\n",
            "  \"Even though the `tf.config.experimental_run_functions_eagerly` \"\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "657/657 [==============================] - 12s 19ms/step - loss: 0.0177 - accuracy: 0.9949 - val_loss: 0.0865 - val_accuracy: 0.9772\n",
            "Epoch 2/5\n",
            "657/657 [==============================] - 13s 19ms/step - loss: 0.0133 - accuracy: 0.9963 - val_loss: 0.0852 - val_accuracy: 0.9782\n",
            "Epoch 3/5\n",
            "657/657 [==============================] - 12s 18ms/step - loss: 0.0119 - accuracy: 0.9970 - val_loss: 0.0842 - val_accuracy: 0.9782\n",
            "Epoch 4/5\n",
            "657/657 [==============================] - 13s 19ms/step - loss: 0.0110 - accuracy: 0.9974 - val_loss: 0.0845 - val_accuracy: 0.9781\n",
            "Epoch 5/5\n",
            "657/657 [==============================] - 12s 19ms/step - loss: 0.0103 - accuracy: 0.9976 - val_loss: 0.0842 - val_accuracy: 0.9787\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "W64ILIU16lms",
        "outputId": "02e14be0-001e-4079-fcb9-7536766d8053"
      },
      "source": [
        "test_loss, test_acc = model.evaluate(test_imgs_reshaped, test_labels)\n",
        "print('\\nTest accuracy:', test_acc)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            " 15/313 [>.............................] - ETA: 2s - loss: 0.0327 - accuracy: 0.9875"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/tensorflow/python/data/ops/dataset_ops.py:4212: UserWarning: Even though the `tf.config.experimental_run_functions_eagerly` option is set, this option does not apply to tf.data functions. To force eager execution of tf.data functions, please use `tf.data.experimental.enable_debug_mode()`.\n",
            "  \"Even though the `tf.config.experimental_run_functions_eagerly` \"\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 2s 8ms/step - loss: 0.0675 - accuracy: 0.9817\n",
            "\n",
            "Test accuracy: 0.9817000031471252\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2u1KLlrXeE5e"
      },
      "source": [
        ""
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ao1BlDmz73A9",
        "outputId": "5780e02b-b693-436d-be9c-e4c7a782885b"
      },
      "source": [
        "model = tf.keras.models.Sequential([\n",
        "       tf.keras.Input(shape= 28 * 28),\n",
        "       tf.keras.layers.Dense(units=256, activation='relu'),\n",
        "       tf.keras.layers.Dense(units=128, activation='relu'),\n",
        "       tf.keras.layers.Dense(units=10, activation='softmax') \n",
        "])\n",
        "\n",
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_6\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense_17 (Dense)             (None, 256)               200960    \n",
            "_________________________________________________________________\n",
            "dense_18 (Dense)             (None, 128)               32896     \n",
            "_________________________________________________________________\n",
            "dense_19 (Dense)             (None, 10)                1290      \n",
            "=================================================================\n",
            "Total params: 235,146\n",
            "Trainable params: 235,146\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pRAlTBteeK4s"
      },
      "source": [
        "RMSprop Optimizer"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JLIDw5H_8BLD"
      },
      "source": [
        "model.compile(\n",
        "    optimizer='RMSprop', \n",
        "    loss='sparse_categorical_crossentropy', \n",
        "    metrics=['accuracy']\n",
        ")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vP7ItC7B8VdL",
        "outputId": "4dad1412-0c88-402d-e593-3eb01d5ba655"
      },
      "source": [
        "h = model.fit(x=train_imgs_reshaped, y=train_labels, epochs=5, batch_size=64, validation_split=0.3)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/5\n",
            "  5/657 [..............................] - ETA: 17s - loss: 1.8246 - accuracy: 0.4219"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/tensorflow/python/data/ops/dataset_ops.py:4212: UserWarning: Even though the `tf.config.experimental_run_functions_eagerly` option is set, this option does not apply to tf.data functions. To force eager execution of tf.data functions, please use `tf.data.experimental.enable_debug_mode()`.\n",
            "  \"Even though the `tf.config.experimental_run_functions_eagerly` \"\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "657/657 [==============================] - 19s 28ms/step - loss: 0.2608 - accuracy: 0.9213 - val_loss: 0.1835 - val_accuracy: 0.9414\n",
            "Epoch 2/5\n",
            "657/657 [==============================] - 18s 27ms/step - loss: 0.1056 - accuracy: 0.9684 - val_loss: 0.1378 - val_accuracy: 0.9586\n",
            "Epoch 3/5\n",
            "657/657 [==============================] - 18s 28ms/step - loss: 0.0705 - accuracy: 0.9790 - val_loss: 0.1034 - val_accuracy: 0.9709\n",
            "Epoch 4/5\n",
            "657/657 [==============================] - 18s 28ms/step - loss: 0.0524 - accuracy: 0.9835 - val_loss: 0.1229 - val_accuracy: 0.9684\n",
            "Epoch 5/5\n",
            "657/657 [==============================] - 18s 28ms/step - loss: 0.0402 - accuracy: 0.9878 - val_loss: 0.1039 - val_accuracy: 0.9725\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sOb7QIE-8d1b",
        "outputId": "15ae94d0-636b-43df-9dc8-1f9a44d5a596"
      },
      "source": [
        "test_loss, test_acc = model.evaluate(test_imgs_reshaped, test_labels)\n",
        "print('\\nTest accuracy:', test_acc)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            " 14/313 [>.............................] - ETA: 2s - loss: 0.0337 - accuracy: 0.9866"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/tensorflow/python/data/ops/dataset_ops.py:4212: UserWarning: Even though the `tf.config.experimental_run_functions_eagerly` option is set, this option does not apply to tf.data functions. To force eager execution of tf.data functions, please use `tf.data.experimental.enable_debug_mode()`.\n",
            "  \"Even though the `tf.config.experimental_run_functions_eagerly` \"\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 3s 8ms/step - loss: 0.0675 - accuracy: 0.9817\n",
            "\n",
            "Test accuracy: 0.9817000031471252\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GBT2wMfm-jU_"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}