3_CNN_(Convolutional_Neural_Network)_in_TensorFlow_2_0.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "3 - CNN (Convolutional Neural Network) in TensorFlow 2.0",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/pratjz/Tensorflow/blob/master/3_CNN_(Convolutional_Neural_Network)_in_TensorFlow_2_0.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zJMS8p3wfPXO"
      },
      "source": [
        "## Step 1: Installing dependencies and notebook gpu setup"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PXNUanZvcwwZ"
      },
      "source": [
        "#!pip install tensorflow-gpu==2.0.0.alpha0"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lRHFdNYAfWKJ"
      },
      "source": [
        "## Step 2: Importing the libraries"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tS5xFeQwe9Xu",
        "outputId": "004821b0-f2f4-4756-a0f8-8576c009e790",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "import tensorflow as tf\n",
        "import matplotlib.pyplot as plt\n",
        "from tensorflow.keras.datasets import cifar10\n",
        "\n",
        "%matplotlib inline\n",
        "tf.__version__"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'2.3.0'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 2
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "e8zYubaPfy-S"
      },
      "source": [
        "## Step 3: Data Preprocessing"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AOAGjiPogF0w"
      },
      "source": [
        "### Loading the Cifar10 dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "v1N6LbS-e9a1"
      },
      "source": [
        "# Setting class names in the dataset\n",
        "class_names = ['airplane', 'automobile', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'truck']"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "f13-8-m_fqKP",
        "outputId": "4ba92e29-9754-48bc-9251-235e0b6c7a23",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "# Loading the dataset\n",
        "(X_train, y_train), (X_test, y_test) = cifar10.load_data()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading data from https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz\n",
            "170500096/170498071 [==============================] - 2s 0us/step\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XqDIpsJWgUkz"
      },
      "source": [
        "### Image normalization"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-QFncPlpgNmU"
      },
      "source": [
        "X_train = X_train / 255.0"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "L1rFHX67gc6G",
        "outputId": "4c095033-4e6b-496e-b7a2-19be497c0d0b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "X_train.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(50000, 32, 32, 3)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ve2WbRGWgc3I"
      },
      "source": [
        "X_test = X_test / 255.0"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7jwvNZ82gh9d",
        "outputId": "933c91fa-ab0f-4040-9850-a577e86b1558",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 283
        }
      },
      "source": [
        "plt.imshow(X_test[10])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7f4ff2751b38>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HXiuZulIguyf"
      },
      "source": [
        "## Step 4: Building a Convolutional Neural Network"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZFTetk8ngy0f"
      },
      "source": [
        "### Defining the model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1TR0JGP5gq2i"
      },
      "source": [
        "model = tf.keras.models.Sequential()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "j92d4FE0hTZV"
      },
      "source": [
        "### Adding the first convolutional layer\n",
        "\n",
        "CNN layer hyper-parameters:\n",
        "- filters: 32\n",
        "- kernel_size: 3\n",
        "- padding: same\n",
        "- activation: relu\n",
        "- input_shape: (32, 32, 3)\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LSkL1iOvg_dE"
      },
      "source": [
        "model.add(tf.keras.layers.Conv2D(filters=32, kernel_size=3, padding=\"same\", activation=\"relu\", input_shape=[32, 32, 3]))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gfRAaRmWiSlZ"
      },
      "source": [
        "### Adding the second convolutional layer and the max-pooling layer\n",
        "\n",
        "CNN layer hyper-parameters:\n",
        "- filters: 32\n",
        "- kernel_size:3\n",
        "- padding: same\n",
        "- activation: relu\n",
        "\n",
        "MaxPool layer hyper-parameters:\n",
        "- pool_size: 2\n",
        "- strides: 2\n",
        "- padding: valid"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sSl7Es5yidMp"
      },
      "source": [
        "model.add(tf.keras.layers.Conv2D(filters=32, kernel_size=3, padding=\"same\", activation=\"relu\"))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wmP9h5wliAR6"
      },
      "source": [
        "model.add(tf.keras.layers.MaxPool2D(pool_size=2, strides=2, padding='valid'))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wd8ERDyvin-0"
      },
      "source": [
        "### Adding the third convolutional layer\n",
        "\n",
        "CNN layer hyper-parameters:\n",
        "\n",
        "    filters: 64\n",
        "    kernel_size:3\n",
        "    padding: same\n",
        "    activation: relu\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "i9HWy6aFixEw"
      },
      "source": [
        "model.add(tf.keras.layers.Conv2D(filters=64, kernel_size=3, padding=\"same\", activation=\"relu\"))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "O55kyOQGi44V"
      },
      "source": [
        "###  Adding the fourth convolutional layer and max-pooling layer\n",
        "\n",
        "CNN layer hyper-parameters:\n",
        "\n",
        "    filters: 64\n",
        "    kernel_size:3\n",
        "    padding: same\n",
        "    activation: relu\n",
        "\n",
        "MaxPool layer hyper-parameters:\n",
        "\n",
        "    pool_size: 2\n",
        "    strides: 2\n",
        "    padding: valid\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5b7vAuhjjCF2"
      },
      "source": [
        "model.add(tf.keras.layers.Conv2D(filters=64, kernel_size=3, padding=\"same\", activation=\"relu\"))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oc493G2BjFhg"
      },
      "source": [
        "model.add(tf.keras.layers.MaxPool2D(pool_size=2, strides=2, padding='valid'))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-hGnR3aXjKbZ"
      },
      "source": [
        "### Adding the flattening layer"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QLzu2cCVjI5Z"
      },
      "source": [
        "model.add(tf.keras.layers.Flatten())"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GpeRUvVWjR1W"
      },
      "source": [
        "### Adding the first fully-connected layer\n",
        "\n",
        "Dense layer hyper-parameters:\n",
        "- units/neurons: 128\n",
        "- activation: relu"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FWzYY8kKjhnZ"
      },
      "source": [
        "model.add(tf.keras.layers.Dense(units=128, activation='relu'))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UaakKTqRjrkF"
      },
      "source": [
        "### Adding the output layer\n",
        "\n",
        "Dense layer hyper-parameters:\n",
        "\n",
        " - units/neurons: 10 (number of classes)\n",
        " - activation: softmax\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4t-JmzRvjnBj"
      },
      "source": [
        "model.add(tf.keras.layers.Dense(units=10, activation='softmax'))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aRr3bCU-ti06",
        "outputId": "b9a3cfdf-e885-415a-d135-5d6b11f6a054",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 459
        }
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d (Conv2D)              (None, 32, 32, 32)        896       \n",
            "_________________________________________________________________\n",
            "conv2d_1 (Conv2D)            (None, 32, 32, 32)        9248      \n",
            "_________________________________________________________________\n",
            "max_pooling2d (MaxPooling2D) (None, 16, 16, 32)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_2 (Conv2D)            (None, 16, 16, 64)        18496     \n",
            "_________________________________________________________________\n",
            "conv2d_3 (Conv2D)            (None, 16, 16, 64)        36928     \n",
            "_________________________________________________________________\n",
            "max_pooling2d_1 (MaxPooling2 (None, 8, 8, 64)          0         \n",
            "_________________________________________________________________\n",
            "flatten (Flatten)            (None, 4096)              0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 128)               524416    \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 10)                1290      \n",
            "=================================================================\n",
            "Total params: 591,274\n",
            "Trainable params: 591,274\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DYgvbNihtprw"
      },
      "source": [
        "### Compiling the model\n",
        "\n",
        "#### sparse_categorical_accuracy\n",
        "sparse_categorical_accuracy checks to see if the maximal true value is equal to the index of the maximal predicted value.\n",
        "\n",
        "https://stackoverflow.com/questions/44477489/keras-difference-between-categorical-accuracy-and-sparse-categorical-accuracy "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oYHELxz4tsa-"
      },
      "source": [
        "model.compile(loss=\"sparse_categorical_crossentropy\", optimizer=\"Adam\", metrics=[\"sparse_categorical_accuracy\"])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Gox3SmwUtwgX"
      },
      "source": [
        "### Training the model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "D3MHvRYKe9fN",
        "outputId": "3c8d3bd3-8ec8-4510-8834-f4801e04d9ae",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        }
      },
      "source": [
        "model.fit(X_train, y_train, epochs=5)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/5\n",
            "1563/1563 [==============================] - 6s 4ms/step - loss: 1.3512 - sparse_categorical_accuracy: 0.5097\n",
            "Epoch 2/5\n",
            "1563/1563 [==============================] - 6s 4ms/step - loss: 0.8949 - sparse_categorical_accuracy: 0.6883\n",
            "Epoch 3/5\n",
            "1563/1563 [==============================] - 6s 4ms/step - loss: 0.7158 - sparse_categorical_accuracy: 0.7521\n",
            "Epoch 4/5\n",
            "1563/1563 [==============================] - 6s 4ms/step - loss: 0.5893 - sparse_categorical_accuracy: 0.7939\n",
            "Epoch 5/5\n",
            "1563/1563 [==============================] - 6s 4ms/step - loss: 0.4777 - sparse_categorical_accuracy: 0.8331\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tensorflow.python.keras.callbacks.History at 0x7f4fe008b7f0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "D8C7Pm0NuOrJ"
      },
      "source": [
        "### Evaluating the model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Z9r8TtNet3D0",
        "outputId": "8bf4d5ba-dee6-449a-d43c-88273e26e7ab",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "test_loss, test_accuracy = model.evaluate(X_test, y_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "313/313 [==============================] - 1s 2ms/step - loss: 0.7733 - sparse_categorical_accuracy: 0.7519\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0rpAPpfzuV0p",
        "outputId": "4dbe24b3-6010-41ec-cc4a-4dc90b4958f2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "print(\"Test accuracy: {}\".format(test_accuracy))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Test accuracy: 0.7519000172615051\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}