chapter04_getting-started-with-neural-networks.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "id": "hv0wmCcBHa3t"
      },
      "outputs": [],
      "source": [
        "from tensorflow.keras.datasets import imdb\n",
        "(train_data, train_labels), (test_data, test_labels) = imdb.load_data(\n",
        "    num_words=10000)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "aMCAonX5Ha3u"
      },
      "outputs": [],
      "source": [
        "train_data[0]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "id": "Tffr8DoAHa3u",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a556b49b-ba34-4dfe-d796-4fa9fce78996"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "1"
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ],
      "source": [
        "train_labels[0]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "id": "27blWmeFHa3u",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "970e7c5f-d8cd-434e-ca73-fdcf5213f3fc"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "9999"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ],
      "source": [
        "max([max(sequence) for sequence in train_data])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "id": "sAED3pfMHa3u",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "1dcdc6c6-2cba-4ee5-e658-35175b31006b"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/imdb_word_index.json\n",
            "1641221/1641221 [==============================] - 0s 0us/step\n"
          ]
        }
      ],
      "source": [
        "word_index = imdb.get_word_index()\n",
        "reverse_word_index = dict(\n",
        "    [(value, key) for (key, value) in word_index.items()])\n",
        "decoded_review = \" \".join(\n",
        "    [reverse_word_index.get(i - 3, \"?\") for i in train_data[0]])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "h47rp_0IHa3u"
      },
      "source": [
        "### Preparing the data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "id": "Ln63JFkaHa3u"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "def vectorize_sequences(sequences, dimension=10000):\n",
        "    results = np.zeros((len(sequences), dimension))\n",
        "    for i, sequence in enumerate(sequences):\n",
        "        for j in sequence:\n",
        "            results[i, j] = 1.\n",
        "    return results\n",
        "x_train = vectorize_sequences(train_data)\n",
        "x_test = vectorize_sequences(test_data)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "id": "tiYXPJWZHa3u",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5b7a02d8-9319-414b-c904-50d4f339266e"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0., 1., 1., ..., 0., 0., 0.])"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ],
      "source": [
        "x_train[0]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {
        "id": "YAYUnwLUHa3u"
      },
      "outputs": [],
      "source": [
        "y_train = np.asarray(train_labels).astype(\"float32\")\n",
        "y_test = np.asarray(test_labels).astype(\"float32\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "T0Bp7AbgHa3u"
      },
      "source": [
        "### Building your model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "id": "iG7g9dgvHa3v"
      },
      "outputs": [],
      "source": [
        "from tensorflow import keras\n",
        "from tensorflow.keras import layers\n",
        "\n",
        "model = keras.Sequential([\n",
        "    layers.Dense(16, activation=\"relu\"),\n",
        "    layers.Dense(16, activation=\"relu\"),\n",
        "    layers.Dense(1, activation=\"sigmoid\")\n",
        "])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "id": "ow2iJi3eHa3v"
      },
      "outputs": [],
      "source": [
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"binary_crossentropy\",\n",
        "              metrics=[\"accuracy\"])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "a0_DFJhlHa3v"
      },
      "source": [
        "### Validating your approach"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "id": "7nmlMS_DHa3v"
      },
      "outputs": [],
      "source": [
        "x_val = x_train[:10000]\n",
        "partial_x_train = x_train[10000:]\n",
        "y_val = y_train[:10000]\n",
        "partial_y_train = y_train[10000:]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 13,
      "metadata": {
        "id": "FFu89fhtHa3v",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a7f56638-3f0d-4140-efbd-694292336ab1"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "30/30 [==============================] - 3s 92ms/step - loss: 0.5389 - accuracy: 0.7761 - val_loss: 0.4126 - val_accuracy: 0.8572\n",
            "Epoch 2/20\n",
            "30/30 [==============================] - 3s 86ms/step - loss: 0.3398 - accuracy: 0.8898 - val_loss: 0.3237 - val_accuracy: 0.8794\n",
            "Epoch 3/20\n",
            "30/30 [==============================] - 2s 77ms/step - loss: 0.2560 - accuracy: 0.9157 - val_loss: 0.3093 - val_accuracy: 0.8749\n",
            "Epoch 4/20\n",
            "30/30 [==============================] - 2s 65ms/step - loss: 0.2056 - accuracy: 0.9301 - val_loss: 0.2955 - val_accuracy: 0.8812\n",
            "Epoch 5/20\n",
            "30/30 [==============================] - 1s 49ms/step - loss: 0.1741 - accuracy: 0.9419 - val_loss: 0.2777 - val_accuracy: 0.8864\n",
            "Epoch 6/20\n",
            "30/30 [==============================] - 2s 63ms/step - loss: 0.1500 - accuracy: 0.9509 - val_loss: 0.2827 - val_accuracy: 0.8867\n",
            "Epoch 7/20\n",
            "30/30 [==============================] - 1s 50ms/step - loss: 0.1301 - accuracy: 0.9587 - val_loss: 0.2913 - val_accuracy: 0.8857\n",
            "Epoch 8/20\n",
            "30/30 [==============================] - 1s 33ms/step - loss: 0.1129 - accuracy: 0.9649 - val_loss: 0.3080 - val_accuracy: 0.8810\n",
            "Epoch 9/20\n",
            "30/30 [==============================] - 1s 34ms/step - loss: 0.0986 - accuracy: 0.9690 - val_loss: 0.3148 - val_accuracy: 0.8821\n",
            "Epoch 10/20\n",
            "30/30 [==============================] - 1s 51ms/step - loss: 0.0831 - accuracy: 0.9760 - val_loss: 0.3317 - val_accuracy: 0.8832\n",
            "Epoch 11/20\n",
            "30/30 [==============================] - 1s 49ms/step - loss: 0.0755 - accuracy: 0.9796 - val_loss: 0.3463 - val_accuracy: 0.8806\n",
            "Epoch 12/20\n",
            "30/30 [==============================] - 1s 34ms/step - loss: 0.0656 - accuracy: 0.9824 - val_loss: 0.3637 - val_accuracy: 0.8775\n",
            "Epoch 13/20\n",
            "30/30 [==============================] - 1s 32ms/step - loss: 0.0582 - accuracy: 0.9856 - val_loss: 0.3823 - val_accuracy: 0.8776\n",
            "Epoch 14/20\n",
            "30/30 [==============================] - 1s 30ms/step - loss: 0.0486 - accuracy: 0.9881 - val_loss: 0.4067 - val_accuracy: 0.8741\n",
            "Epoch 15/20\n",
            "30/30 [==============================] - 1s 33ms/step - loss: 0.0389 - accuracy: 0.9927 - val_loss: 0.4396 - val_accuracy: 0.8748\n",
            "Epoch 16/20\n",
            "30/30 [==============================] - 1s 33ms/step - loss: 0.0352 - accuracy: 0.9938 - val_loss: 0.4500 - val_accuracy: 0.8739\n",
            "Epoch 17/20\n",
            "30/30 [==============================] - 1s 33ms/step - loss: 0.0308 - accuracy: 0.9944 - val_loss: 0.5026 - val_accuracy: 0.8700\n",
            "Epoch 18/20\n",
            "30/30 [==============================] - 1s 31ms/step - loss: 0.0253 - accuracy: 0.9958 - val_loss: 0.5029 - val_accuracy: 0.8704\n",
            "Epoch 19/20\n",
            "30/30 [==============================] - 1s 31ms/step - loss: 0.0227 - accuracy: 0.9963 - val_loss: 0.5224 - val_accuracy: 0.8718\n",
            "Epoch 20/20\n",
            "30/30 [==============================] - 1s 33ms/step - loss: 0.0176 - accuracy: 0.9988 - val_loss: 0.5422 - val_accuracy: 0.8703\n"
          ]
        }
      ],
      "source": [
        "history = model.fit(partial_x_train,\n",
        "                    partial_y_train,\n",
        "                    epochs=20,\n",
        "                    batch_size=512,\n",
        "                    validation_data=(x_val, y_val))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 14,
      "metadata": {
        "id": "yyNwlIltHa3v",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5c6cf3ee-1994-442f-a3aa-7b41e6637591"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "dict_keys(['loss', 'accuracy', 'val_loss', 'val_accuracy'])"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ],
      "source": [
        "history_dict = history.history\n",
        "history_dict.keys()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 15,
      "metadata": {
        "id": "QZDqnqa1Ha3v",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "outputId": "bffd159b-23d3-4527-843b-86d370b76e43"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "history_dict = history.history\n",
        "loss_values = history_dict[\"loss\"]\n",
        "val_loss_values = history_dict[\"val_loss\"]\n",
        "epochs = range(1, len(loss_values) + 1)\n",
        "plt.plot(epochs, loss_values, \"bo\", label=\"Training loss\")\n",
        "plt.plot(epochs, val_loss_values, \"b\", label=\"Validation loss\")\n",
        "plt.title(\"Training and validation loss\")\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Loss\")\n",
        "plt.legend()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 16,
      "metadata": {
        "id": "vwgnzE05Ha3v",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "outputId": "ee21d7c3-0a28-451a-9d8b-732f026cb6a9"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "plt.clf()\n",
        "acc = history_dict[\"accuracy\"]\n",
        "val_acc = history_dict[\"val_accuracy\"]\n",
        "plt.plot(epochs, acc, \"bo\", label=\"Training acc\")\n",
        "plt.plot(epochs, val_acc, \"b\", label=\"Validation acc\")\n",
        "plt.title(\"Training and validation accuracy\")\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.legend()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 17,
      "metadata": {
        "id": "NHTCK2yQHa3w",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e2eadbff-67a2-4831-b19d-aad2d1b24a6e"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/4\n",
            "49/49 [==============================] - 3s 34ms/step - loss: 0.4415 - accuracy: 0.8222\n",
            "Epoch 2/4\n",
            "49/49 [==============================] - 1s 25ms/step - loss: 0.2606 - accuracy: 0.9054\n",
            "Epoch 3/4\n",
            "49/49 [==============================] - 1s 23ms/step - loss: 0.2065 - accuracy: 0.9248\n",
            "Epoch 4/4\n",
            "49/49 [==============================] - 1s 22ms/step - loss: 0.1768 - accuracy: 0.9359\n",
            "782/782 [==============================] - 2s 2ms/step - loss: 0.2976 - accuracy: 0.8810\n"
          ]
        }
      ],
      "source": [
        "model = keras.Sequential([\n",
        "    layers.Dense(16, activation=\"relu\"),\n",
        "    layers.Dense(16, activation=\"relu\"),\n",
        "    layers.Dense(1, activation=\"sigmoid\")\n",
        "])\n",
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"binary_crossentropy\",\n",
        "              metrics=[\"accuracy\"])\n",
        "model.fit(x_train, y_train, epochs=4, batch_size=512)\n",
        "results = model.evaluate(x_test, y_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 18,
      "metadata": {
        "id": "X69igiY5Ha3w",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "0971aca9-da3c-43fd-ff44-6f7a9674c9e3"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.2976357936859131, 0.8809999823570251]"
            ]
          },
          "metadata": {},
          "execution_count": 18
        }
      ],
      "source": [
        "results"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yEzYMHF0Ha3w"
      },
      "source": [
        "### Using a trained model to generate predictions on new data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 19,
      "metadata": {
        "id": "VuHsEYtQHa3w",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "2b2935bc-2b03-4c3f-a065-b7e0196d92bf"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "782/782 [==============================] - 3s 3ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0.20705025],\n",
              "       [0.99972236],\n",
              "       [0.92243826],\n",
              "       ...,\n",
              "       [0.11894033],\n",
              "       [0.11072411],\n",
              "       [0.7817268 ]], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 19
        }
      ],
      "source": [
        "model.predict(x_test)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4CRCuTFMHa3w"
      },
      "source": [
        "## Classifying newswires: A multiclass classification example"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wsc5QoyCHa3x"
      },
      "source": [
        "### The Reuters dataset"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 20,
      "metadata": {
        "id": "JnmpmrbPHa3x",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "9645ae1e-ed57-46a2-cfd8-0a9c13f35656"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/reuters.npz\n",
            "2110848/2110848 [==============================] - 0s 0us/step\n"
          ]
        }
      ],
      "source": [
        "from tensorflow.keras.datasets import reuters\n",
        "(train_data, train_labels), (test_data, test_labels) = reuters.load_data(\n",
        "    num_words=10000)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 21,
      "metadata": {
        "id": "QryxPW8aHa3x",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c4560379-f209-4542-e2ff-8a3f4be0e95a"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "8982"
            ]
          },
          "metadata": {},
          "execution_count": 21
        }
      ],
      "source": [
        "len(train_data)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 22,
      "metadata": {
        "id": "clSOJd7lHa3x",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5d9b1cb4-dfad-40ce-81d6-324914e8e9fc"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2246"
            ]
          },
          "metadata": {},
          "execution_count": 22
        }
      ],
      "source": [
        "len(test_data)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 23,
      "metadata": {
        "id": "OjAgYWI6Ha3x",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6a7efcf9-a1ba-4c3c-f2f6-2cf3f6092dad"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[1,\n",
              " 245,\n",
              " 273,\n",
              " 207,\n",
              " 156,\n",
              " 53,\n",
              " 74,\n",
              " 160,\n",
              " 26,\n",
              " 14,\n",
              " 46,\n",
              " 296,\n",
              " 26,\n",
              " 39,\n",
              " 74,\n",
              " 2979,\n",
              " 3554,\n",
              " 14,\n",
              " 46,\n",
              " 4689,\n",
              " 4329,\n",
              " 86,\n",
              " 61,\n",
              " 3499,\n",
              " 4795,\n",
              " 14,\n",
              " 61,\n",
              " 451,\n",
              " 4329,\n",
              " 17,\n",
              " 12]"
            ]
          },
          "metadata": {},
          "execution_count": 23
        }
      ],
      "source": [
        "train_data[10]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 24,
      "metadata": {
        "id": "SgC6byfLHa3x",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e8338d3f-a7c2-4293-eb88-440ac8b94d7b"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/reuters_word_index.json\n",
            "550378/550378 [==============================] - 0s 0us/step\n"
          ]
        }
      ],
      "source": [
        "word_index = reuters.get_word_index()\n",
        "reverse_word_index = dict([(value, key) for (key, value) in word_index.items()])\n",
        "decoded_newswire = \" \".join([reverse_word_index.get(i - 3, \"?\") for i in\n",
        "    train_data[0]])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 25,
      "metadata": {
        "id": "nLoQntTMHa3y",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f7c85627-3e82-48dd-805f-d03b7b8f615b"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "3"
            ]
          },
          "metadata": {},
          "execution_count": 25
        }
      ],
      "source": [
        "train_labels[10]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 26,
      "metadata": {
        "id": "jzZGmQ_FHa3y"
      },
      "outputs": [],
      "source": [
        "x_train = vectorize_sequences(train_data)\n",
        "x_test = vectorize_sequences(test_data)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 27,
      "metadata": {
        "id": "iXbN9mztHa3y"
      },
      "outputs": [],
      "source": [
        "def to_one_hot(labels, dimension=46):\n",
        "    results = np.zeros((len(labels), dimension))\n",
        "    for i, label in enumerate(labels):\n",
        "        results[i, label] = 1.\n",
        "    return results\n",
        "y_train = to_one_hot(train_labels)\n",
        "y_test = to_one_hot(test_labels)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 28,
      "metadata": {
        "id": "GxsktTatHa3y"
      },
      "outputs": [],
      "source": [
        "from tensorflow.keras.utils import to_categorical\n",
        "y_train = to_categorical(train_labels)\n",
        "y_test = to_categorical(test_labels)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NVUkoGdJHa3y"
      },
      "source": [
        "### Building your model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 29,
      "metadata": {
        "id": "6urZHQHXHa3y"
      },
      "outputs": [],
      "source": [
        "model = keras.Sequential([\n",
        "    layers.Dense(64, activation=\"relu\"),\n",
        "    layers.Dense(64, activation=\"relu\"),\n",
        "    layers.Dense(46, activation=\"softmax\")\n",
        "])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PNppiYb2Ha3y"
      },
      "source": [
        "**Compiling the model**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 30,
      "metadata": {
        "id": "jh3AON5HHa3y"
      },
      "outputs": [],
      "source": [
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"categorical_crossentropy\",\n",
        "              metrics=[\"accuracy\"])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5S9FpigjHa3z"
      },
      "source": [
        "### Validating your approach"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "n5-m7bT3Ha3z"
      },
      "source": [
        "**Setting aside a validation set**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 31,
      "metadata": {
        "id": "a_nHCS5iHa3z"
      },
      "outputs": [],
      "source": [
        "x_val = x_train[:1000]\n",
        "partial_x_train = x_train[1000:]\n",
        "y_val = y_train[:1000]\n",
        "partial_y_train = y_train[1000:]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RU8bqEMvHa3z"
      },
      "source": [
        "**Training the model**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 32,
      "metadata": {
        "id": "75Lgu5FtHa3z",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "49fe79c0-b1d9-4437-d66b-daed9daf4012"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "16/16 [==============================] - 2s 82ms/step - loss: 2.7321 - accuracy: 0.4902 - val_loss: 1.8483 - val_accuracy: 0.6150\n",
            "Epoch 2/20\n",
            "16/16 [==============================] - 1s 80ms/step - loss: 1.5583 - accuracy: 0.6699 - val_loss: 1.4066 - val_accuracy: 0.7070\n",
            "Epoch 3/20\n",
            "16/16 [==============================] - 1s 63ms/step - loss: 1.1831 - accuracy: 0.7462 - val_loss: 1.2138 - val_accuracy: 0.7330\n",
            "Epoch 4/20\n",
            "16/16 [==============================] - 1s 46ms/step - loss: 0.9518 - accuracy: 0.8001 - val_loss: 1.0753 - val_accuracy: 0.7630\n",
            "Epoch 5/20\n",
            "16/16 [==============================] - 1s 47ms/step - loss: 0.7841 - accuracy: 0.8296 - val_loss: 0.9977 - val_accuracy: 0.7850\n",
            "Epoch 6/20\n",
            "16/16 [==============================] - 1s 48ms/step - loss: 0.6513 - accuracy: 0.8586 - val_loss: 0.9564 - val_accuracy: 0.8000\n",
            "Epoch 7/20\n",
            "16/16 [==============================] - 1s 45ms/step - loss: 0.5408 - accuracy: 0.8869 - val_loss: 0.9100 - val_accuracy: 0.8060\n",
            "Epoch 8/20\n",
            "16/16 [==============================] - 1s 47ms/step - loss: 0.4551 - accuracy: 0.9040 - val_loss: 0.9023 - val_accuracy: 0.8130\n",
            "Epoch 9/20\n",
            "16/16 [==============================] - 1s 46ms/step - loss: 0.3852 - accuracy: 0.9188 - val_loss: 0.8817 - val_accuracy: 0.8120\n",
            "Epoch 10/20\n",
            "16/16 [==============================] - 1s 47ms/step - loss: 0.3306 - accuracy: 0.9290 - val_loss: 0.8713 - val_accuracy: 0.8170\n",
            "Epoch 11/20\n",
            "16/16 [==============================] - 1s 45ms/step - loss: 0.2812 - accuracy: 0.9381 - val_loss: 0.8875 - val_accuracy: 0.8160\n",
            "Epoch 12/20\n",
            "16/16 [==============================] - 1s 48ms/step - loss: 0.2464 - accuracy: 0.9442 - val_loss: 0.8809 - val_accuracy: 0.8180\n",
            "Epoch 13/20\n",
            "16/16 [==============================] - 1s 47ms/step - loss: 0.2191 - accuracy: 0.9468 - val_loss: 0.8841 - val_accuracy: 0.8200\n",
            "Epoch 14/20\n",
            "16/16 [==============================] - 1s 46ms/step - loss: 0.1993 - accuracy: 0.9495 - val_loss: 0.8954 - val_accuracy: 0.8150\n",
            "Epoch 15/20\n",
            "16/16 [==============================] - 1s 46ms/step - loss: 0.1799 - accuracy: 0.9531 - val_loss: 0.9139 - val_accuracy: 0.8140\n",
            "Epoch 16/20\n",
            "16/16 [==============================] - 1s 55ms/step - loss: 0.1685 - accuracy: 0.9538 - val_loss: 0.9105 - val_accuracy: 0.8170\n",
            "Epoch 17/20\n",
            "16/16 [==============================] - 1s 83ms/step - loss: 0.1587 - accuracy: 0.9545 - val_loss: 0.9297 - val_accuracy: 0.8170\n",
            "Epoch 18/20\n",
            "16/16 [==============================] - 1s 75ms/step - loss: 0.1484 - accuracy: 0.9570 - val_loss: 0.9551 - val_accuracy: 0.8180\n",
            "Epoch 19/20\n",
            "16/16 [==============================] - 1s 45ms/step - loss: 0.1407 - accuracy: 0.9555 - val_loss: 0.9689 - val_accuracy: 0.8170\n",
            "Epoch 20/20\n",
            "16/16 [==============================] - 1s 44ms/step - loss: 0.1318 - accuracy: 0.9580 - val_loss: 0.9674 - val_accuracy: 0.8080\n"
          ]
        }
      ],
      "source": [
        "history = model.fit(partial_x_train,\n",
        "                    partial_y_train,\n",
        "                    epochs=20,\n",
        "                    batch_size=512,\n",
        "                    validation_data=(x_val, y_val))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BLU0IKFiHa3z"
      },
      "source": [
        "**Plotting the training and validation loss**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 33,
      "metadata": {
        "id": "3WwCwIaWHa3z",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "outputId": "945b596c-e7a3-4773-f757-6e90ea8ed728"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "loss = history.history[\"loss\"]\n",
        "val_loss = history.history[\"val_loss\"]\n",
        "epochs = range(1, len(loss) + 1)\n",
        "plt.plot(epochs, loss, \"bo\", label=\"Training loss\")\n",
        "plt.plot(epochs, val_loss, \"b\", label=\"Validation loss\")\n",
        "plt.title(\"Training and validation loss\")\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Loss\")\n",
        "plt.legend()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JCJSORZsHa3z"
      },
      "source": [
        "**Plotting the training and validation accuracy**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 34,
      "metadata": {
        "id": "SxdWWMyvHa3z",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "outputId": "14680ba2-09d0-495f-d5a6-6b94aa440201"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "plt.clf()\n",
        "acc = history.history[\"accuracy\"]\n",
        "val_acc = history.history[\"val_accuracy\"]\n",
        "plt.plot(epochs, acc, \"bo\", label=\"Training accuracy\")\n",
        "plt.plot(epochs, val_acc, \"b\", label=\"Validation accuracy\")\n",
        "plt.title(\"Training and validation accuracy\")\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.legend()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kNiJ7iCDHa3z"
      },
      "source": [
        "**Retraining a model from scratch**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 35,
      "metadata": {
        "id": "32P2wqsVHa3z",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f8719837-d647-4b4a-e385-93d59afe37ba"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/9\n",
            "18/18 [==============================] - 2s 73ms/step - loss: 2.8104 - accuracy: 0.4874\n",
            "Epoch 2/9\n",
            "18/18 [==============================] - 1s 63ms/step - loss: 1.5099 - accuracy: 0.6945\n",
            "Epoch 3/9\n",
            "18/18 [==============================] - 1s 42ms/step - loss: 1.1323 - accuracy: 0.7563\n",
            "Epoch 4/9\n",
            "18/18 [==============================] - 1s 42ms/step - loss: 0.9218 - accuracy: 0.8048\n",
            "Epoch 5/9\n",
            "18/18 [==============================] - 1s 41ms/step - loss: 0.7661 - accuracy: 0.8372\n",
            "Epoch 6/9\n",
            "18/18 [==============================] - 1s 43ms/step - loss: 0.6344 - accuracy: 0.8632\n",
            "Epoch 7/9\n",
            "18/18 [==============================] - 1s 42ms/step - loss: 0.5362 - accuracy: 0.8849\n",
            "Epoch 8/9\n",
            "18/18 [==============================] - 1s 39ms/step - loss: 0.4569 - accuracy: 0.9040\n",
            "Epoch 9/9\n",
            "18/18 [==============================] - 1s 40ms/step - loss: 0.3848 - accuracy: 0.9173\n",
            "71/71 [==============================] - 0s 3ms/step - loss: 0.9424 - accuracy: 0.7898\n"
          ]
        }
      ],
      "source": [
        "model = keras.Sequential([\n",
        "  layers.Dense(64, activation=\"relu\"),\n",
        "  layers.Dense(64, activation=\"relu\"),\n",
        "  layers.Dense(46, activation=\"softmax\")\n",
        "])\n",
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"categorical_crossentropy\",\n",
        "              metrics=[\"accuracy\"])\n",
        "model.fit(x_train,\n",
        "          y_train,\n",
        "          epochs=9,\n",
        "          batch_size=512)\n",
        "results = model.evaluate(x_test, y_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 36,
      "metadata": {
        "id": "FjG8pkDLHa30",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "2426c0af-5b16-4af3-bf2d-142d694234d0"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.9423660635948181, 0.7898486256599426]"
            ]
          },
          "metadata": {},
          "execution_count": 36
        }
      ],
      "source": [
        "results"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 37,
      "metadata": {
        "id": "BPO2eXLqHa30",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4446c57b-802f-41b1-9461-15356b3e655c"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.18121104185218165"
            ]
          },
          "metadata": {},
          "execution_count": 37
        }
      ],
      "source": [
        "import copy\n",
        "test_labels_copy = copy.copy(test_labels)\n",
        "np.random.shuffle(test_labels_copy)\n",
        "hits_array = np.array(test_labels) == np.array(test_labels_copy)\n",
        "hits_array.mean()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "X96h9koOHa30"
      },
      "source": [
        "### Generating predictions on new data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 38,
      "metadata": {
        "id": "AM5cWJCGHa30",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "1418fe57-27ac-40b5-92c4-85f2518055e6"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "71/71 [==============================] - 0s 3ms/step\n"
          ]
        }
      ],
      "source": [
        "predictions = model.predict(x_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 39,
      "metadata": {
        "id": "R1sG73WAHa30",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "0271a8ba-7240-4259-854a-258c68151dd4"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(46,)"
            ]
          },
          "metadata": {},
          "execution_count": 39
        }
      ],
      "source": [
        "predictions[0].shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 40,
      "metadata": {
        "id": "w-s3Bt8iHa30",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "691e1704-2af8-4b41-c28f-6f7139d1f8b4"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "1.0"
            ]
          },
          "metadata": {},
          "execution_count": 40
        }
      ],
      "source": [
        "np.sum(predictions[0])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 41,
      "metadata": {
        "id": "QLP_E7diHa30",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "73a94d50-e2ce-42fa-e647-47869f6c11b6"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "3"
            ]
          },
          "metadata": {},
          "execution_count": 41
        }
      ],
      "source": [
        "np.argmax(predictions[0])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ea2VSyzpHa30"
      },
      "source": [
        "### A different way to handle the labels and the loss"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 42,
      "metadata": {
        "id": "xaB_4wm2Ha30"
      },
      "outputs": [],
      "source": [
        "y_train = np.array(train_labels)\n",
        "y_test = np.array(test_labels)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 43,
      "metadata": {
        "id": "SdJyHLjhHa30"
      },
      "outputs": [],
      "source": [
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"sparse_categorical_crossentropy\",\n",
        "              metrics=[\"accuracy\"])"
      ]
    }
  ]
}