Building_Complex_Deep_Learning_Models_Using_Keras_Functional_API.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "authorship_tag": "ABX9TyMfyzhS20eJSMmiUsByCRRj",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU",
    "gpuClass": "standard"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/youssefHosni/Machine-Learning-Practical-Guide/blob/main/Building_Complex_Deep_Learning_Models_Using_Keras_Functional_API.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Building Complex Deep Learning Models Using Keras Functional API"
      ],
      "metadata": {
        "id": "La7xjp0XO8dO"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "The Functional API in Keras is a powerful extension to the Sequential model. It allows us to create complex models by composing simple building blocks that each takes one or more inputs, process them and return one output. This article will give you an overview of the Keras Functional API architecture as well as concrete examples of how to implement it for your own use cases."
      ],
      "metadata": {
        "id": "9DioVF6lPHEI"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Table of Contents:\n",
        "\n",
        "1. Loading the Data & Building a Simple Model\n",
        "2. Building a Wide & Deep Neural Network\n",
        "3. Building Wide & Deep Neural Network With Multiple Inputs\n",
        "4. Building Wide & Deep Neural Network With Multiple Inputs & Outputs\n"
      ],
      "metadata": {
        "id": "I4yTrCJlPLsZ"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 1. Loading the Data & Building a Simple Model\n",
        "First, we will load the data we will use through this article which will be the [MNSIT Fashion](https://www.tensorflow.org/datasets/catalog/fashion_mnist) dataset. The data consists of different types of fashion and contains ten classes."
      ],
      "metadata": {
        "id": "b2hQ2lFGPbvy"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Let's first load the required libraries:"
      ],
      "metadata": {
        "id": "SsyzXL05PmSe"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "ep_lS8gKm5y-"
      },
      "outputs": [],
      "source": [
        "from tensorflow import keras\n",
        "from tensorflow.keras.utils import plot_model\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Next, let's load the dataset from the **Keras.datasets** library and then split the data into training and validation and finally normalize the data:"
      ],
      "metadata": {
        "id": "HhNB_YOnPpbf"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "fashion_mnist = keras.datasets.fashion_mnist\n",
        "(X_train_full, y_train_full), (X_test, y_test) = fashion_mnist.load_data()\n",
        "\n",
        "X_valid, X_train = X_train_full[:5000] / 255., X_train_full[5000:] / 255.\n",
        "y_valid, y_train = y_train_full[:5000], y_train_full[5000:]\n",
        "X_test = X_test / 255."
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QLs11PnIcsWv",
        "outputId": "cb262b0a-e82f-4780-bc78-54f289a971a6"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/train-labels-idx1-ubyte.gz\n",
            "29515/29515 [==============================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/train-images-idx3-ubyte.gz\n",
            "26421880/26421880 [==============================] - 1s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/t10k-labels-idx1-ubyte.gz\n",
            "5148/5148 [==============================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/t10k-images-idx3-ubyte.gz\n",
            "4422102/4422102 [==============================] - 0s 0us/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now let's print the size of the datasets and the labels:"
      ],
      "metadata": {
        "id": "D4u53LlvPvF6"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print('The size of the trainig data is {} and the size of the train labels is {}' .format(X_train.shape , y_train.shape))\n",
        "print('The size of the validation data is {} and the size of the validation labels is {}' .format(X_valid.shape , y_valid.shape))\n",
        "print('The size of the test data is {} and the size of the test labels is {}' .format(X_test.shape , y_test.shape))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0NvlaqkLCcXV",
        "outputId": "198f64bf-9a3e-4130-8b6b-c51ff18de62d"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The size of the trainig data is (55000, 28, 28) and the size of the train labels is (55000,)\n",
            "The size of the validation data is (5000, 28, 28) and the size of the validation labels is (5000,)\n",
            "The size of the test data is (10000, 28, 28) and the size of the test labels is (10000,)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "As we can see the size of the training data is 55000 and 5000 for validation data and test data is 10000. The size of each image is 28*28. \n"
      ],
      "metadata": {
        "id": "E7rNNlQYP3mu"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "class_names = [\"T-shirt/top\", \"Trouser\", \"Pullover\", \"Dress\", \"Coat\",\n",
        "               \"Sandal\", \"Shirt\", \"Sneaker\", \"Bag\", \"Ankle boot\"]"
      ],
      "metadata": {
        "id": "4y-Rd1ilC-LN"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Let's plot a subset of the data to have a better idea of what we working with. The data is shown in figure 1."
      ],
      "metadata": {
        "id": "WwsC_XXVPzZ0"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "n_rows = 4\n",
        "n_cols = 10\n",
        "plt.figure(figsize=(n_cols * 1.2, n_rows * 1.2))\n",
        "for row in range(n_rows):\n",
        "    for col in range(n_cols):\n",
        "        index = n_cols * row + col\n",
        "        plt.subplot(n_rows, n_cols, index + 1)\n",
        "        plt.imshow(X_train[index], cmap=\"binary\", interpolation=\"nearest\")\n",
        "        plt.axis('off')\n",
        "        plt.title(class_names[y_train[index]], fontsize=12)\n",
        "plt.subplots_adjust(wspace=0.2, hspace=0.5)\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 308
        },
        "id": "n681r1a8C7I_",
        "outputId": "97108c3a-c0a2-4d9d-b01e-75a3418ccf2f"
      },
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 864x345.6 with 40 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now for simplicity reasons, we will flatten the data to be easily used within the neural networks:"
      ],
      "metadata": {
        "id": "3HxKameVP69S"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "X_train = X_train.reshape(X_train.shape[0], 28*28)\n",
        "X_valid = X_valid.reshape(X_valid.shape[0], 28*28)\n",
        "X_test = X_test.reshape(X_test.shape[0], 28*28)"
      ],
      "metadata": {
        "id": "oNyUBI7REZ9n"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X_train.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nUZpNYfKgpbE",
        "outputId": "c7f08cd0-d762-4c71-f54b-607579085b3e"
      },
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(55000, 784)"
            ]
          },
          "metadata": {},
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Let's build and train a simple MLP neural network as shown in the figure below:"
      ],
      "metadata": {
        "id": "Oa5g5L8FQAJp"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "![1_bSrwOYQgJnZLkzhtLL2mYw.png](data:image/png;base64,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)"
      ],
      "metadata": {
        "id": "5DjQ-eNhQFcX"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Let's now build the model shown in figure2. The code is self-explanatory we first define the input layer with the expected input size. Next, we define the first hidden layer with 30 neurons and the Relu activation function and then the second hidden layer with the same parameters.\n",
        " \n",
        "Finally, the output layer with the number of classes we have which in our case is 10. Finally, we will plot the model using the plot_model function from the Keras utils library."
      ],
      "metadata": {
        "id": "c9XcQdz-QTcN"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "input_ = keras.layers.Input(shape=[28*28])\n",
        "hidden1 = keras.layers.Dense(30, activation=\"relu\")(input_)\n",
        "hidden2 = keras.layers.Dense(30, activation=\"relu\")(hidden1)\n",
        "output = output = keras.layers.Dense(10, activation=\"softmax\", name=\"output\")(hidden2)\n",
        "model = keras.models.Model(inputs=[input_], outputs=[output])\n",
        "\n",
        "keras.utils.plot_model(model, \"Neural_Network.png\", show_shapes=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 422
        },
        "id": "AOBlelZLDPI0",
        "outputId": "e2e9d83f-60ed-4a97-b971-dcd4b45ac154"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now let's compile the model and fit it to the training data and plot the learning curve using the code below:"
      ],
      "metadata": {
        "id": "LCkXpEadQV4i"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model.compile(loss=\"sparse_categorical_crossentropy\",\n",
        "              optimizer=\"sgd\",\n",
        "              metrics=[\"accuracy\"])\n",
        "\n",
        "history = model.fit(X_train, y_train, epochs=30, validation_data=(X_valid, y_valid))\n",
        "\n",
        "pd.DataFrame(history.history).plot(figsize=(8, 5))\n",
        "plt.grid(True)\n",
        "plt.gca().set_ylim(0, 1)\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "gmimFLpwDuDd",
        "outputId": "aa8addb7-af6b-4c38-a274-428d9f05855c"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.9042 - accuracy: 0.6882 - val_loss: 0.5814 - val_accuracy: 0.8034\n",
            "Epoch 2/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.5402 - accuracy: 0.8147 - val_loss: 0.4917 - val_accuracy: 0.8348\n",
            "Epoch 3/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4828 - accuracy: 0.8323 - val_loss: 0.4449 - val_accuracy: 0.8504\n",
            "Epoch 4/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4544 - accuracy: 0.8397 - val_loss: 0.4370 - val_accuracy: 0.8520\n",
            "Epoch 5/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4348 - accuracy: 0.8470 - val_loss: 0.4433 - val_accuracy: 0.8394\n",
            "Epoch 6/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4196 - accuracy: 0.8526 - val_loss: 0.4033 - val_accuracy: 0.8664\n",
            "Epoch 7/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4079 - accuracy: 0.8562 - val_loss: 0.4014 - val_accuracy: 0.8598\n",
            "Epoch 8/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3969 - accuracy: 0.8589 - val_loss: 0.4167 - val_accuracy: 0.8526\n",
            "Epoch 9/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3890 - accuracy: 0.8621 - val_loss: 0.4067 - val_accuracy: 0.8588\n",
            "Epoch 10/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3817 - accuracy: 0.8650 - val_loss: 0.3709 - val_accuracy: 0.8740\n",
            "Epoch 11/30\n",
            "1719/1719 [==============================] - 4s 3ms/step - loss: 0.3751 - accuracy: 0.8682 - val_loss: 0.3921 - val_accuracy: 0.8630\n",
            "Epoch 12/30\n",
            "1719/1719 [==============================] - 4s 3ms/step - loss: 0.3687 - accuracy: 0.8699 - val_loss: 0.3697 - val_accuracy: 0.8710\n",
            "Epoch 13/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3620 - accuracy: 0.8719 - val_loss: 0.3672 - val_accuracy: 0.8728\n",
            "Epoch 14/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3568 - accuracy: 0.8727 - val_loss: 0.3577 - val_accuracy: 0.8764\n",
            "Epoch 15/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3511 - accuracy: 0.8746 - val_loss: 0.3727 - val_accuracy: 0.8692\n",
            "Epoch 16/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3472 - accuracy: 0.8757 - val_loss: 0.3586 - val_accuracy: 0.8770\n",
            "Epoch 17/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3417 - accuracy: 0.8769 - val_loss: 0.3662 - val_accuracy: 0.8732\n",
            "Epoch 18/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3381 - accuracy: 0.8786 - val_loss: 0.3921 - val_accuracy: 0.8574\n",
            "Epoch 19/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3339 - accuracy: 0.8800 - val_loss: 0.3504 - val_accuracy: 0.8772\n",
            "Epoch 20/30\n",
            "1719/1719 [==============================] - 4s 3ms/step - loss: 0.3300 - accuracy: 0.8810 - val_loss: 0.3589 - val_accuracy: 0.8724\n",
            "Epoch 21/30\n",
            "1719/1719 [==============================] - 4s 3ms/step - loss: 0.3263 - accuracy: 0.8818 - val_loss: 0.3445 - val_accuracy: 0.8748\n",
            "Epoch 22/30\n",
            "1719/1719 [==============================] - 4s 3ms/step - loss: 0.3234 - accuracy: 0.8833 - val_loss: 0.3412 - val_accuracy: 0.8790\n",
            "Epoch 23/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3195 - accuracy: 0.8846 - val_loss: 0.3542 - val_accuracy: 0.8726\n",
            "Epoch 24/30\n",
            "1719/1719 [==============================] - 4s 3ms/step - loss: 0.3170 - accuracy: 0.8855 - val_loss: 0.3795 - val_accuracy: 0.8634\n",
            "Epoch 25/30\n",
            "1719/1719 [==============================] - 4s 3ms/step - loss: 0.3137 - accuracy: 0.8866 - val_loss: 0.3523 - val_accuracy: 0.8762\n",
            "Epoch 26/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3116 - accuracy: 0.8878 - val_loss: 0.3399 - val_accuracy: 0.8800\n",
            "Epoch 27/30\n",
            "1719/1719 [==============================] - 4s 3ms/step - loss: 0.3085 - accuracy: 0.8886 - val_loss: 0.3336 - val_accuracy: 0.8806\n",
            "Epoch 28/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3060 - accuracy: 0.8892 - val_loss: 0.3414 - val_accuracy: 0.8774\n",
            "Epoch 29/30\n",
            "1719/1719 [==============================] - 4s 3ms/step - loss: 0.3031 - accuracy: 0.8905 - val_loss: 0.3372 - val_accuracy: 0.8792\n",
            "Epoch 30/30\n",
            "1719/1719 [==============================] - 4s 3ms/step - loss: 0.3016 - accuracy: 0.8903 - val_loss: 0.3346 - val_accuracy: 0.8832\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 576x360 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "The final step is to evaluate the model on the test data and print the accuracy and the loss. The accuracy of the model is 86.7% which is good compared to the simplicity of the model."
      ],
      "metadata": {
        "id": "dMhr-oUQQc83"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model.evaluate(X_test, y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "I2FtMVxJDu38",
        "outputId": "c479da07-17ca-4aa6-d281-f49535a00a1c"
      },
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 1s 2ms/step - loss: 0.3656 - accuracy: 0.8674\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.3655974566936493, 0.8673999905586243]"
            ]
          },
          "metadata": {},
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 2. Building a Wide & Deep Neural Network"
      ],
      "metadata": {
        "id": "7M-g011HGUr5"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "The first example of complex models is the wide & Deep network which is a nonsequential neural network that was first introduced in the [Wide & Deep Learning for Recommender Systems paper](https://arxiv.org/abs/1606.07792) by Heng-Tze Cheng et al. \n",
        "\n",
        "In this model part of the inputs or all of it is connected directly to the output layer as shown in figure 5. This architecture allows the model to learn both deep information using the deep path and simple information directly from the input through the wide path."
      ],
      "metadata": {
        "id": "CphLLlSZQhT4"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "![1_sbK9g6tHn7hKgA9HT6dZlA.png](data:image/png;base64,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)"
      ],
      "metadata": {
        "id": "ozJw5iNEQ--R"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Here is the code to build this architecture using Keras API and to make sure that the model built is the same as the required one we will plot the model architecture using the **Keras.utils.plot_model** method. The resulting model is shown in figure 6."
      ],
      "metadata": {
        "id": "bYX76VEWRHGK"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "input_ = keras.layers.Input(shape=[28*28])\n",
        "hidden1 = keras.layers.Dense(30, activation=\"relu\")(input_)\n",
        "hidden2 = keras.layers.Dense(30, activation=\"relu\")(hidden1)\n",
        "concat = keras.layers.concatenate([input_, hidden2])\n",
        "output = output = keras.layers.Dense(10, activation=\"softmax\", name=\"output\")(concat)\n",
        "model = keras.models.Model(inputs=[input_], outputs=[output])\n",
        "\n",
        "keras.utils.plot_model(model, \"Deep & wide.png\", show_shapes=True)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 533
        },
        "id": "1KJtdP6uwPBp",
        "outputId": "10bc34d2-4f25-469f-912d-db0ec46929c0"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Let's explain the code in more detail:\n",
        "\n",
        "* **input_**: In the first line we created an Input object. This defines the input of the model which includes its type and shape. The model can have multiple objects as we will see in the coming models. \n",
        "* **hidden1:** The next step is to create the first hidden layer with 30 neurons and the ReLU activation function and we pass to it the input we created in the line before. You can see that it is similar to calling a function that's why it is called Functional API.\n",
        "* **hidden2:** Next we created another hidden layer with the same properties as the previously hidden layer and we pass to it the output of the previously hidden layer. \n",
        "* **concat:** This is a Concatenate layer, and is used to concatenate the input and the output of the second hidden layer. \n",
        "output: This is the output layer which consists of one single neuron and no activation function for the sake of simplicity but you can add what you need for your model. \n",
        "* **model:** Finally we create the Keras model and specify which input and outputs to use.\n",
        "\n",
        "The next step is to compile the model and fit it to the training data and plot the learning curve. This can be done using the code below:"
      ],
      "metadata": {
        "id": "39iMaKBVRMJk"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model.compile(loss=\"sparse_categorical_crossentropy\",\n",
        "              optimizer=\"sgd\",\n",
        "              metrics=[\"accuracy\"])\n",
        "\n",
        "history = model.fit(X_train, y_train, epochs=30, validation_data=(X_valid, y_valid))\n",
        "\n",
        "pd.DataFrame(history.history).plot(figsize=(8, 5))\n",
        "plt.grid(True)\n",
        "plt.gca().set_ylim(0, 1)\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "sUnG500Jw994",
        "outputId": "1d97fc9b-880e-4054-e1a9-9a321ad95ada"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/30\n",
            "1719/1719 [==============================] - 6s 3ms/step - loss: 0.8344 - accuracy: 0.7308 - val_loss: 0.6276 - val_accuracy: 0.8010\n",
            "Epoch 2/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.5869 - accuracy: 0.8081 - val_loss: 0.5452 - val_accuracy: 0.8252\n",
            "Epoch 3/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.5299 - accuracy: 0.8236 - val_loss: 0.5007 - val_accuracy: 0.8376\n",
            "Epoch 4/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4974 - accuracy: 0.8329 - val_loss: 0.4751 - val_accuracy: 0.8446\n",
            "Epoch 5/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4764 - accuracy: 0.8376 - val_loss: 0.4617 - val_accuracy: 0.8430\n",
            "Epoch 6/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4607 - accuracy: 0.8415 - val_loss: 0.4554 - val_accuracy: 0.8462\n",
            "Epoch 7/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4479 - accuracy: 0.8451 - val_loss: 0.4412 - val_accuracy: 0.8468\n",
            "Epoch 8/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4374 - accuracy: 0.8498 - val_loss: 0.4368 - val_accuracy: 0.8546\n",
            "Epoch 9/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4280 - accuracy: 0.8516 - val_loss: 0.4164 - val_accuracy: 0.8572\n",
            "Epoch 10/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4193 - accuracy: 0.8550 - val_loss: 0.4136 - val_accuracy: 0.8584\n",
            "Epoch 11/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4119 - accuracy: 0.8577 - val_loss: 0.4081 - val_accuracy: 0.8638\n",
            "Epoch 12/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4040 - accuracy: 0.8599 - val_loss: 0.3998 - val_accuracy: 0.8626\n",
            "Epoch 13/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3967 - accuracy: 0.8620 - val_loss: 0.3959 - val_accuracy: 0.8630\n",
            "Epoch 14/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3909 - accuracy: 0.8642 - val_loss: 0.4001 - val_accuracy: 0.8650\n",
            "Epoch 15/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3846 - accuracy: 0.8663 - val_loss: 0.3884 - val_accuracy: 0.8670\n",
            "Epoch 16/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3796 - accuracy: 0.8681 - val_loss: 0.3796 - val_accuracy: 0.8708\n",
            "Epoch 17/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3744 - accuracy: 0.8704 - val_loss: 0.3810 - val_accuracy: 0.8708\n",
            "Epoch 18/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3690 - accuracy: 0.8719 - val_loss: 0.3800 - val_accuracy: 0.8720\n",
            "Epoch 19/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3639 - accuracy: 0.8729 - val_loss: 0.3733 - val_accuracy: 0.8758\n",
            "Epoch 20/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3598 - accuracy: 0.8748 - val_loss: 0.3719 - val_accuracy: 0.8752\n",
            "Epoch 21/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3551 - accuracy: 0.8764 - val_loss: 0.3860 - val_accuracy: 0.8702\n",
            "Epoch 22/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3519 - accuracy: 0.8776 - val_loss: 0.3772 - val_accuracy: 0.8664\n",
            "Epoch 23/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3471 - accuracy: 0.8792 - val_loss: 0.3638 - val_accuracy: 0.8780\n",
            "Epoch 24/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3431 - accuracy: 0.8807 - val_loss: 0.3644 - val_accuracy: 0.8776\n",
            "Epoch 25/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3395 - accuracy: 0.8823 - val_loss: 0.3621 - val_accuracy: 0.8760\n",
            "Epoch 26/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3373 - accuracy: 0.8819 - val_loss: 0.3548 - val_accuracy: 0.8786\n",
            "Epoch 27/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3340 - accuracy: 0.8827 - val_loss: 0.3677 - val_accuracy: 0.8758\n",
            "Epoch 28/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3302 - accuracy: 0.8840 - val_loss: 0.3640 - val_accuracy: 0.8782\n",
            "Epoch 29/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3268 - accuracy: 0.8859 - val_loss: 0.3668 - val_accuracy: 0.8748\n",
            "Epoch 30/30\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3243 - accuracy: 0.8858 - val_loss: 0.3591 - val_accuracy: 0.8738\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 576x360 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "The final step is to evaluate the model on the test data and print the accuracy and the loss. The accuracy of the model is 85.9% which is good compared to the simplicity of the model although it is smaller than the simple MLP model introduced in the previous section."
      ],
      "metadata": {
        "id": "wU714xwpRiGo"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model.evaluate(X_test, y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sxpiF6IpgSAM",
        "outputId": "ce437117-1ec4-4a98-ad97-ad91d408cd88"
      },
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 1s 2ms/step - loss: 0.3919 - accuracy: 0.8595\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.3919452726840973, 0.859499990940094]"
            ]
          },
          "metadata": {},
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 3. Building Wide & Deep Neural Network With Multiple Inputs"
      ],
      "metadata": {
        "id": "8JtbFOANGcrH"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Another important variation you can to the previous model is to have two paths for the input, this will be very useful if you would like for example to send part of the input features through the wide path and part through the deep path. This model is shown in figure 8."
      ],
      "metadata": {
        "id": "PjzffnlBRl5J"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "![1_jn6uV19wHadNn5FW_5Lm8g.png](data:image/png;base64,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)"
      ],
      "metadata": {
        "id": "td26SD2XRo_Q"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "So let's build this and we will split half of the features to go through the deep path and the other half to go through the wide path and then we will plot the created model as shown in figure 9."
      ],
      "metadata": {
        "id": "O-UpYjrfR5UC"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "input_A = keras.layers.Input(shape=[392], name=\"wide_input\")\n",
        "input_B = keras.layers.Input(shape=[392], name=\"deep_input\")\n",
        "hidden1 = keras.layers.Dense(30, activation=\"relu\")(input_B)\n",
        "hidden2 = keras.layers.Dense(30, activation=\"relu\")(hidden1)\n",
        "concat = keras.layers.concatenate([input_A, hidden2])\n",
        "output = keras.layers.Dense(10, activation=\"softmax\", name=\"output\")(concat)\n",
        "model = keras.models.Model(inputs=[input_A, input_B], outputs=[output])\n",
        "\n",
        "keras.utils.plot_model(model, \"multiple_inputs_model.png\", show_shapes=True)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 533
        },
        "id": "ESWNNf0LvZUB",
        "outputId": "c80f7b3c-6f44-4d27-beae-e8698939fbfc"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "The code is very similar to the previous one except that we have to define two inputs (input_A and input_B) when creating the model each with the input shape of 392 which is half the number of features we have. \n",
        "\n",
        "Next, let's compile and fit the model to the data using the code below:"
      ],
      "metadata": {
        "id": "zoIx9GtIR193"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model.compile(loss=\"sparse_categorical_crossentropy\",\n",
        "              optimizer=\"sgd\",\n",
        "              metrics=[\"accuracy\"])\n",
        "\n",
        "X_train_A, X_train_B = X_train[:, :392], X_train[:, 392:]\n",
        "X_valid_A, X_valid_B = X_valid[:, :392], X_valid[:, 392:]\n",
        "X_test_A, X_test_B = X_test[:, :392], X_test[:, 392:]\n",
        "\n",
        "history = model.fit((X_train_A, X_train_B), y_train, epochs=20,\n",
        "                    validation_data=((X_valid_A, X_valid_B), y_valid))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9PTI620cwAPz",
        "outputId": "734223aa-4117-4aba-8f82-c5a70065667f"
      },
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "1719/1719 [==============================] - 7s 3ms/step - loss: 0.9552 - accuracy: 0.7015 - val_loss: 0.7040 - val_accuracy: 0.7656\n",
            "Epoch 2/20\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.6274 - accuracy: 0.7931 - val_loss: 0.5759 - val_accuracy: 0.8108\n",
            "Epoch 3/20\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.5524 - accuracy: 0.8132 - val_loss: 0.5202 - val_accuracy: 0.8298\n",
            "Epoch 4/20\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.5130 - accuracy: 0.8245 - val_loss: 0.4958 - val_accuracy: 0.8328\n",
            "Epoch 5/20\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4893 - accuracy: 0.8302 - val_loss: 0.4727 - val_accuracy: 0.8450\n",
            "Epoch 6/20\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4722 - accuracy: 0.8361 - val_loss: 0.4583 - val_accuracy: 0.8510\n",
            "Epoch 7/20\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4586 - accuracy: 0.8411 - val_loss: 0.4591 - val_accuracy: 0.8500\n",
            "Epoch 8/20\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4470 - accuracy: 0.8459 - val_loss: 0.4400 - val_accuracy: 0.8556\n",
            "Epoch 9/20\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4373 - accuracy: 0.8488 - val_loss: 0.4288 - val_accuracy: 0.8584\n",
            "Epoch 10/20\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4287 - accuracy: 0.8513 - val_loss: 0.4333 - val_accuracy: 0.8554\n",
            "Epoch 11/20\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4215 - accuracy: 0.8543 - val_loss: 0.4312 - val_accuracy: 0.8560\n",
            "Epoch 12/20\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4149 - accuracy: 0.8564 - val_loss: 0.4094 - val_accuracy: 0.8620\n",
            "Epoch 13/20\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4092 - accuracy: 0.8581 - val_loss: 0.4105 - val_accuracy: 0.8614\n",
            "Epoch 14/20\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.4032 - accuracy: 0.8600 - val_loss: 0.4032 - val_accuracy: 0.8654\n",
            "Epoch 15/20\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3990 - accuracy: 0.8613 - val_loss: 0.3971 - val_accuracy: 0.8666\n",
            "Epoch 16/20\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3940 - accuracy: 0.8624 - val_loss: 0.3974 - val_accuracy: 0.8672\n",
            "Epoch 17/20\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3897 - accuracy: 0.8652 - val_loss: 0.3955 - val_accuracy: 0.8670\n",
            "Epoch 18/20\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3861 - accuracy: 0.8666 - val_loss: 0.3949 - val_accuracy: 0.8682\n",
            "Epoch 19/20\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3829 - accuracy: 0.8668 - val_loss: 0.3882 - val_accuracy: 0.8682\n",
            "Epoch 20/20\n",
            "1719/1719 [==============================] - 5s 3ms/step - loss: 0.3791 - accuracy: 0.8693 - val_loss: 0.3858 - val_accuracy: 0.8714\n",
            "313/313 [==============================] - 1s 2ms/step - loss: 0.4151 - accuracy: 0.8543\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "As you can see the compile and fit methods are similar to the previous code the only difference is in the fit method in which we have to give two inputs for training data, validation, and test data. Finally, we will evaluate this model using the test data using the code below:"
      ],
      "metadata": {
        "id": "dPEYfcdTR8jg"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "test = model.evaluate((X_test_A, X_test_B), y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "AyWS4E8_klr-",
        "outputId": "94cfd096-eed3-4a3c-ee01-03e287fc4f8c"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 1s 2ms/step - loss: 0.4151 - accuracy: 0.8543\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 4. Building Wide & Deep Neural Network With Multiple Inputs & Outputs"
      ],
      "metadata": {
        "id": "7W2NGMw_g-uM"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Finally, we will build a Wide and deep neural network with both multiple inputs and multiple outputs as shown in figure 10. This can be used in multiple cases such as:\n",
        "* You have multiple outputs such as in object detection and classification you \n",
        "would like to return the coordinates of the object and its class. \n",
        "* You might be working on multiple tasks and everyone will have different output layers. For example, you might be working on facial detection tasks and emotion recognition and identify identification. So you can train the same network and have separate output layers for each of them. \n",
        "* Another possible use is for regularization purposes which we will be using multiple outputs for in this example. Regularization methods are used to reduce overfitting and improve models' ability to generalize to new data. \n",
        "\n",
        "In the example below we will use an auxiliary output in a neural network to ensure that the underlying part of the network will learn useful information on its own without relying on the rest of the network."
      ],
      "metadata": {
        "id": "P148M5fXSA8_"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "![1_4c-9-6GjMKH_ZDWHswsiVA.png](data:image/png;base64,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)"
      ],
      "metadata": {
        "id": "hhBxHFJSSKwj"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "input_A = keras.layers.Input(shape=[392], name=\"wide_input\")\n",
        "input_B = keras.layers.Input(shape=[392], name=\"deep_input\")\n",
        "\n",
        "hidden1 = keras.layers.Dense(30, activation=\"relu\")(input_B)\n",
        "hidden2 = keras.layers.Dense(30, activation=\"relu\")(hidden1)\n",
        "\n",
        "concat = keras.layers.concatenate([input_A, hidden2])\n",
        "\n",
        "output = keras.layers.Dense(10, activation=\"softmax\",name=\"main_output\")(concat)\n",
        "aux_output = keras.layers.Dense(10, activation=\"softmax\", name=\"aux_output\")(hidden2)\n",
        "\n",
        "model = keras.models.Model(inputs=[input_A, input_B],\n",
        "                           outputs=[output, aux_output])\n",
        "\n",
        "keras.utils.plot_model(model, \"model.png\", show_shapes=True)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 533
        },
        "id": "Q6Hn6bG4vXHB",
        "outputId": "7620eedb-d363-4613-9176-647c386bddc9"
      },
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now let's compile and fit the model to the training data. Since we are using two outputs therefore we will have to define two loss functions (if we pass one loss function then it will assume that it will be used for both outputs) also the weight of each loss should be declared else it will assume both outputs will contribute to the model loss with the same amount. In this case, we would like to give more weight to the main output than the auxiliary output. So the weight of the losses will be 0.7 to the main output and 0.3 to the auxiliary output."
      ],
      "metadata": {
        "id": "gfCu7nyySPDw"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model.compile(loss=[\"sparse_categorical_crossentropy\",\"sparse_categorical_crossentropy\" ],\n",
        "              optimizer=\"sgd\",\n",
        "              loss_weights=[0.7, 0.3],\n",
        "              metrics=[\"accuracy\"])"
      ],
      "metadata": {
        "id": "WyTul1m_wmqB"
      },
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Let's now fit the model on the training data:"
      ],
      "metadata": {
        "id": "uawmdSI0SSaX"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "X_train_A, X_train_B = X_train[:, :392], X_train[:, 392:]\n",
        "X_valid_A, X_valid_B = X_valid[:, :392], X_valid[:, 392:]\n",
        "X_test_A, X_test_B = X_test[:, :392], X_test[:, 392:]\n",
        "\n",
        "history = model.fit((X_train_A, X_train_B), y_train, epochs=20,\n",
        "                    validation_data=((X_valid_A, X_valid_B), y_valid))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6KfVcNCnMGz3",
        "outputId": "0b800156-16b8-48d4-d761-130fd2c7eeef"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "1719/1719 [==============================] - 13s 6ms/step - loss: 1.1672 - main_output_loss: 1.0265 - aux_output_loss: 1.4954 - main_output_accuracy: 0.6779 - aux_output_accuracy: 0.5028 - val_loss: 0.8206 - val_main_output_loss: 0.7142 - val_aux_output_loss: 1.0687 - val_main_output_accuracy: 0.7702 - val_aux_output_accuracy: 0.6474\n",
            "Epoch 2/20\n",
            "1719/1719 [==============================] - 7s 4ms/step - loss: 0.7457 - main_output_loss: 0.6476 - aux_output_loss: 0.9748 - main_output_accuracy: 0.7847 - aux_output_accuracy: 0.6561 - val_loss: 0.6693 - val_main_output_loss: 0.5821 - val_aux_output_loss: 0.8728 - val_main_output_accuracy: 0.8082 - val_aux_output_accuracy: 0.6830\n",
            "Epoch 3/20\n",
            "1719/1719 [==============================] - 6s 4ms/step - loss: 0.6537 - main_output_loss: 0.5644 - aux_output_loss: 0.8620 - main_output_accuracy: 0.8078 - aux_output_accuracy: 0.6865 - val_loss: 0.6146 - val_main_output_loss: 0.5326 - val_aux_output_loss: 0.8057 - val_main_output_accuracy: 0.8194 - val_aux_output_accuracy: 0.7052\n",
            "Epoch 4/20\n",
            "1719/1719 [==============================] - 6s 4ms/step - loss: 0.6110 - main_output_loss: 0.5257 - aux_output_loss: 0.8102 - main_output_accuracy: 0.8186 - aux_output_accuracy: 0.7073 - val_loss: 0.5842 - val_main_output_loss: 0.5041 - val_aux_output_loss: 0.7711 - val_main_output_accuracy: 0.8280 - val_aux_output_accuracy: 0.7290\n",
            "Epoch 5/20\n",
            "1719/1719 [==============================] - 7s 4ms/step - loss: 0.5838 - main_output_loss: 0.5016 - aux_output_loss: 0.7755 - main_output_accuracy: 0.8267 - aux_output_accuracy: 0.7224 - val_loss: 0.5601 - val_main_output_loss: 0.4833 - val_aux_output_loss: 0.7392 - val_main_output_accuracy: 0.8354 - val_aux_output_accuracy: 0.7288\n",
            "Epoch 6/20\n",
            "1719/1719 [==============================] - 7s 4ms/step - loss: 0.5634 - main_output_loss: 0.4834 - aux_output_loss: 0.7502 - main_output_accuracy: 0.8330 - aux_output_accuracy: 0.7328 - val_loss: 0.5383 - val_main_output_loss: 0.4626 - val_aux_output_loss: 0.7148 - val_main_output_accuracy: 0.8454 - val_aux_output_accuracy: 0.7462\n",
            "Epoch 7/20\n",
            "1719/1719 [==============================] - 6s 3ms/step - loss: 0.5482 - main_output_loss: 0.4698 - aux_output_loss: 0.7312 - main_output_accuracy: 0.8379 - aux_output_accuracy: 0.7398 - val_loss: 0.5446 - val_main_output_loss: 0.4719 - val_aux_output_loss: 0.7143 - val_main_output_accuracy: 0.8354 - val_aux_output_accuracy: 0.7426\n",
            "Epoch 8/20\n",
            "1719/1719 [==============================] - 6s 4ms/step - loss: 0.5353 - main_output_loss: 0.4585 - aux_output_loss: 0.7145 - main_output_accuracy: 0.8413 - aux_output_accuracy: 0.7460 - val_loss: 0.5166 - val_main_output_loss: 0.4437 - val_aux_output_loss: 0.6866 - val_main_output_accuracy: 0.8500 - val_aux_output_accuracy: 0.7570\n",
            "Epoch 9/20\n",
            "1719/1719 [==============================] - 6s 3ms/step - loss: 0.5240 - main_output_loss: 0.4484 - aux_output_loss: 0.7004 - main_output_accuracy: 0.8464 - aux_output_accuracy: 0.7510 - val_loss: 0.5109 - val_main_output_loss: 0.4390 - val_aux_output_loss: 0.6788 - val_main_output_accuracy: 0.8522 - val_aux_output_accuracy: 0.7576\n",
            "Epoch 10/20\n",
            "1719/1719 [==============================] - 6s 4ms/step - loss: 0.5140 - main_output_loss: 0.4397 - aux_output_loss: 0.6872 - main_output_accuracy: 0.8489 - aux_output_accuracy: 0.7563 - val_loss: 0.5015 - val_main_output_loss: 0.4297 - val_aux_output_loss: 0.6690 - val_main_output_accuracy: 0.8588 - val_aux_output_accuracy: 0.7610\n",
            "Epoch 11/20\n",
            "1719/1719 [==============================] - 6s 3ms/step - loss: 0.5057 - main_output_loss: 0.4327 - aux_output_loss: 0.6761 - main_output_accuracy: 0.8505 - aux_output_accuracy: 0.7600 - val_loss: 0.5070 - val_main_output_loss: 0.4398 - val_aux_output_loss: 0.6639 - val_main_output_accuracy: 0.8530 - val_aux_output_accuracy: 0.7658\n",
            "Epoch 12/20\n",
            "1719/1719 [==============================] - 6s 4ms/step - loss: 0.4976 - main_output_loss: 0.4256 - aux_output_loss: 0.6656 - main_output_accuracy: 0.8534 - aux_output_accuracy: 0.7629 - val_loss: 0.4885 - val_main_output_loss: 0.4203 - val_aux_output_loss: 0.6476 - val_main_output_accuracy: 0.8582 - val_aux_output_accuracy: 0.7688\n",
            "Epoch 13/20\n",
            "1719/1719 [==============================] - 7s 4ms/step - loss: 0.4901 - main_output_loss: 0.4193 - aux_output_loss: 0.6554 - main_output_accuracy: 0.8563 - aux_output_accuracy: 0.7676 - val_loss: 0.4894 - val_main_output_loss: 0.4211 - val_aux_output_loss: 0.6490 - val_main_output_accuracy: 0.8584 - val_aux_output_accuracy: 0.7706\n",
            "Epoch 14/20\n",
            "1719/1719 [==============================] - 6s 3ms/step - loss: 0.4837 - main_output_loss: 0.4139 - aux_output_loss: 0.6466 - main_output_accuracy: 0.8580 - aux_output_accuracy: 0.7693 - val_loss: 0.4854 - val_main_output_loss: 0.4183 - val_aux_output_loss: 0.6422 - val_main_output_accuracy: 0.8616 - val_aux_output_accuracy: 0.7762\n",
            "Epoch 15/20\n",
            "1719/1719 [==============================] - 6s 4ms/step - loss: 0.4772 - main_output_loss: 0.4082 - aux_output_loss: 0.6384 - main_output_accuracy: 0.8600 - aux_output_accuracy: 0.7722 - val_loss: 0.4729 - val_main_output_loss: 0.4080 - val_aux_output_loss: 0.6244 - val_main_output_accuracy: 0.8632 - val_aux_output_accuracy: 0.7806\n",
            "Epoch 16/20\n",
            "1719/1719 [==============================] - 6s 3ms/step - loss: 0.4722 - main_output_loss: 0.4039 - aux_output_loss: 0.6317 - main_output_accuracy: 0.8610 - aux_output_accuracy: 0.7749 - val_loss: 0.4694 - val_main_output_loss: 0.4050 - val_aux_output_loss: 0.6196 - val_main_output_accuracy: 0.8624 - val_aux_output_accuracy: 0.7808\n",
            "Epoch 17/20\n",
            "1719/1719 [==============================] - 6s 3ms/step - loss: 0.4665 - main_output_loss: 0.3989 - aux_output_loss: 0.6242 - main_output_accuracy: 0.8633 - aux_output_accuracy: 0.7783 - val_loss: 0.4868 - val_main_output_loss: 0.4222 - val_aux_output_loss: 0.6375 - val_main_output_accuracy: 0.8634 - val_aux_output_accuracy: 0.7698\n",
            "Epoch 18/20\n",
            "1719/1719 [==============================] - 7s 4ms/step - loss: 0.4625 - main_output_loss: 0.3955 - aux_output_loss: 0.6189 - main_output_accuracy: 0.8638 - aux_output_accuracy: 0.7797 - val_loss: 0.4649 - val_main_output_loss: 0.3994 - val_aux_output_loss: 0.6178 - val_main_output_accuracy: 0.8628 - val_aux_output_accuracy: 0.7824\n",
            "Epoch 19/20\n",
            "1719/1719 [==============================] - 6s 3ms/step - loss: 0.4580 - main_output_loss: 0.3919 - aux_output_loss: 0.6123 - main_output_accuracy: 0.8651 - aux_output_accuracy: 0.7825 - val_loss: 0.4614 - val_main_output_loss: 0.3982 - val_aux_output_loss: 0.6087 - val_main_output_accuracy: 0.8674 - val_aux_output_accuracy: 0.7864\n",
            "Epoch 20/20\n",
            "1719/1719 [==============================] - 6s 3ms/step - loss: 0.4530 - main_output_loss: 0.3875 - aux_output_loss: 0.6057 - main_output_accuracy: 0.8668 - aux_output_accuracy: 0.7842 - val_loss: 0.4622 - val_main_output_loss: 0.3991 - val_aux_output_loss: 0.6095 - val_main_output_accuracy: 0.8644 - val_aux_output_accuracy: 0.7808\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Finally, let's evaluate the model on the test data:"
      ],
      "metadata": {
        "id": "dsRQ1MHQSU-V"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model.evaluate([X_test_A, X_test_B], [y_test, y_test])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "aSjIrMDoMKGR",
        "outputId": "b9c13e42-7879-4634-b62c-9ea7e2eac010"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 1s 3ms/step - loss: 0.4938 - main_output_loss: 0.4289 - aux_output_loss: 0.6451 - main_output_accuracy: 0.8497 - aux_output_accuracy: 0.7645\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.49377304315567017,\n",
              " 0.42891421914100647,\n",
              " 0.6451098918914795,\n",
              " 0.8496999740600586,\n",
              " 0.7645000219345093]"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "As you can see this will return five numbers as the following: \n",
        "* Total loss: 0.4938\n",
        "* Main output loss: 0.4289\n",
        "* Auxiliary output loss: 0.6451\n",
        "* Main output accuracy:  0.8497\n",
        "* Auxiliary output accuracy:  0.7645"
      ],
      "metadata": {
        "id": "GzGtQw1aSXuh"
      }
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "1CSVGR5TScZA"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}