practice(4_3).ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "gpuType": "T4"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "## Predicting house prices: A regression example"
      ],
      "metadata": {
        "id": "jqiCHbBcQEzb"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "eivXvKLZOzcN",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "9157d6b6-a9e1-4a4d-c5fd-8aaa9f6fb987"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/boston_housing.npz\n",
            "57026/57026 [==============================] - 0s 0us/step\n"
          ]
        }
      ],
      "source": [
        "from tensorflow.keras.datasets import boston_housing\n",
        "(train_data, train_targets), (test_data, test_targets) = boston_housing.load_data()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "7KTE45wNOzcN",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "96a82688-aff5-4449-f48f-c803cb36935e"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(404, 13)"
            ]
          },
          "metadata": {},
          "execution_count": 2
        }
      ],
      "source": [
        "train_data.shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "xLI5ZQ-FOzcN",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c3064b3e-454d-4c8d-cc87-1a43296a5646"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(102, 13)"
            ]
          },
          "metadata": {},
          "execution_count": 3
        }
      ],
      "source": [
        "test_data.shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "n9E5qcVZOzcN",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "da7a9e4c-0ea0-4f7a-b84f-07379e813f28"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([15.2, 42.3, 50. , 21.1, 17.7, 18.5, 11.3, 15.6, 15.6, 14.4, 12.1,\n",
              "       17.9, 23.1, 19.9, 15.7,  8.8, 50. , 22.5, 24.1, 27.5, 10.9, 30.8,\n",
              "       32.9, 24. , 18.5, 13.3, 22.9, 34.7, 16.6, 17.5, 22.3, 16.1, 14.9,\n",
              "       23.1, 34.9, 25. , 13.9, 13.1, 20.4, 20. , 15.2, 24.7, 22.2, 16.7,\n",
              "       12.7, 15.6, 18.4, 21. , 30.1, 15.1, 18.7,  9.6, 31.5, 24.8, 19.1,\n",
              "       22. , 14.5, 11. , 32. , 29.4, 20.3, 24.4, 14.6, 19.5, 14.1, 14.3,\n",
              "       15.6, 10.5,  6.3, 19.3, 19.3, 13.4, 36.4, 17.8, 13.5, 16.5,  8.3,\n",
              "       14.3, 16. , 13.4, 28.6, 43.5, 20.2, 22. , 23. , 20.7, 12.5, 48.5,\n",
              "       14.6, 13.4, 23.7, 50. , 21.7, 39.8, 38.7, 22.2, 34.9, 22.5, 31.1,\n",
              "       28.7, 46. , 41.7, 21. , 26.6, 15. , 24.4, 13.3, 21.2, 11.7, 21.7,\n",
              "       19.4, 50. , 22.8, 19.7, 24.7, 36.2, 14.2, 18.9, 18.3, 20.6, 24.6,\n",
              "       18.2,  8.7, 44. , 10.4, 13.2, 21.2, 37. , 30.7, 22.9, 20. , 19.3,\n",
              "       31.7, 32. , 23.1, 18.8, 10.9, 50. , 19.6,  5. , 14.4, 19.8, 13.8,\n",
              "       19.6, 23.9, 24.5, 25. , 19.9, 17.2, 24.6, 13.5, 26.6, 21.4, 11.9,\n",
              "       22.6, 19.6,  8.5, 23.7, 23.1, 22.4, 20.5, 23.6, 18.4, 35.2, 23.1,\n",
              "       27.9, 20.6, 23.7, 28. , 13.6, 27.1, 23.6, 20.6, 18.2, 21.7, 17.1,\n",
              "        8.4, 25.3, 13.8, 22.2, 18.4, 20.7, 31.6, 30.5, 20.3,  8.8, 19.2,\n",
              "       19.4, 23.1, 23. , 14.8, 48.8, 22.6, 33.4, 21.1, 13.6, 32.2, 13.1,\n",
              "       23.4, 18.9, 23.9, 11.8, 23.3, 22.8, 19.6, 16.7, 13.4, 22.2, 20.4,\n",
              "       21.8, 26.4, 14.9, 24.1, 23.8, 12.3, 29.1, 21. , 19.5, 23.3, 23.8,\n",
              "       17.8, 11.5, 21.7, 19.9, 25. , 33.4, 28.5, 21.4, 24.3, 27.5, 33.1,\n",
              "       16.2, 23.3, 48.3, 22.9, 22.8, 13.1, 12.7, 22.6, 15. , 15.3, 10.5,\n",
              "       24. , 18.5, 21.7, 19.5, 33.2, 23.2,  5. , 19.1, 12.7, 22.3, 10.2,\n",
              "       13.9, 16.3, 17. , 20.1, 29.9, 17.2, 37.3, 45.4, 17.8, 23.2, 29. ,\n",
              "       22. , 18. , 17.4, 34.6, 20.1, 25. , 15.6, 24.8, 28.2, 21.2, 21.4,\n",
              "       23.8, 31. , 26.2, 17.4, 37.9, 17.5, 20. ,  8.3, 23.9,  8.4, 13.8,\n",
              "        7.2, 11.7, 17.1, 21.6, 50. , 16.1, 20.4, 20.6, 21.4, 20.6, 36.5,\n",
              "        8.5, 24.8, 10.8, 21.9, 17.3, 18.9, 36.2, 14.9, 18.2, 33.3, 21.8,\n",
              "       19.7, 31.6, 24.8, 19.4, 22.8,  7.5, 44.8, 16.8, 18.7, 50. , 50. ,\n",
              "       19.5, 20.1, 50. , 17.2, 20.8, 19.3, 41.3, 20.4, 20.5, 13.8, 16.5,\n",
              "       23.9, 20.6, 31.5, 23.3, 16.8, 14. , 33.8, 36.1, 12.8, 18.3, 18.7,\n",
              "       19.1, 29. , 30.1, 50. , 50. , 22. , 11.9, 37.6, 50. , 22.7, 20.8,\n",
              "       23.5, 27.9, 50. , 19.3, 23.9, 22.6, 15.2, 21.7, 19.2, 43.8, 20.3,\n",
              "       33.2, 19.9, 22.5, 32.7, 22. , 17.1, 19. , 15. , 16.1, 25.1, 23.7,\n",
              "       28.7, 37.2, 22.6, 16.4, 25. , 29.8, 22.1, 17.4, 18.1, 30.3, 17.5,\n",
              "       24.7, 12.6, 26.5, 28.7, 13.3, 10.4, 24.4, 23. , 20. , 17.8,  7. ,\n",
              "       11.8, 24.4, 13.8, 19.4, 25.2, 19.4, 19.4, 29.1])"
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ],
      "source": [
        "train_targets"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "mean = train_data.mean(axis=0)\n",
        "train_data -= mean\n",
        "std = train_data.std(axis=0)\n",
        "train_data /= std\n",
        "test_data -= mean\n",
        "test_data /= std"
      ],
      "metadata": {
        "id": "xPO2GR_IQe7C"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def build_model():\n",
        "    model = keras.Sequential([\n",
        "        layers.Dense(64, activation=\"relu\"),\n",
        "        layers.Dense(64, activation=\"relu\"),\n",
        "        layers.Dense(1)\n",
        "    ])\n",
        "    model.compile(optimizer=\"rmsprop\", loss=\"mse\", metrics=[\"mae\"])\n",
        "    return model"
      ],
      "metadata": {
        "id": "WXH3YoDPQhH6"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "from tensorflow import keras\n",
        "from tensorflow.keras import layers"
      ],
      "metadata": {
        "id": "6E6WpBYCQz0E"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "k = 4\n",
        "num_val_samples = len(train_data) // k\n",
        "num_epochs = 100\n",
        "all_scores = []\n",
        "for i in range(k):\n",
        "    print(f\"Processing fold #{i}\")\n",
        "    val_data = train_data[i * num_val_samples: (i + 1) * num_val_samples]\n",
        "    val_targets = train_targets[i * num_val_samples: (i + 1) * num_val_samples]\n",
        "    partial_train_data = np.concatenate(\n",
        "        [train_data[:i * num_val_samples],\n",
        "         train_data[(i + 1) * num_val_samples:]],\n",
        "        axis=0)\n",
        "    partial_train_targets = np.concatenate(\n",
        "        [train_targets[:i * num_val_samples],\n",
        "         train_targets[(i + 1) * num_val_samples:]],\n",
        "        axis=0)\n",
        "    model = build_model()\n",
        "    model.fit(partial_train_data, partial_train_targets,\n",
        "              epochs=num_epochs, batch_size=16, verbose=0)\n",
        "    val_mse, val_mae = model.evaluate(val_data, val_targets, verbose=0)\n",
        "    all_scores.append(val_mae)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "jxlzwzEnQjUl",
        "outputId": "f4c33e38-e7fa-44ec-b215-7e58b918e2ad"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Processing fold #0\n",
            "Processing fold #1\n",
            "Processing fold #2\n",
            "Processing fold #3\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "859G8hYzTJPY"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "SxlfoCFxOzcO",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "19591c58-5e40-459c-c4a7-0ffb7a3e6ff2"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[2.1003119945526123, 2.473936080932617, 2.4789657592773438, 2.580036163330078]"
            ]
          },
          "metadata": {},
          "execution_count": 16
        }
      ],
      "source": [
        "all_scores"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "mapPA5KxOzcO",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "2190326b-46f7-4dbb-9f34-ca9f3f5322a6"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2.408312499523163"
            ]
          },
          "metadata": {},
          "execution_count": 17
        }
      ],
      "source": [
        "np.mean(all_scores)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "khXgAdbKOzcO",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "99767621-b5ee-414c-df8c-a2f3c4f352bb"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Processing fold #0\n",
            "Processing fold #1\n",
            "Processing fold #2\n",
            "Processing fold #3\n"
          ]
        }
      ],
      "source": [
        "num_epochs = 500\n",
        "all_mae_histories = []\n",
        "for i in range(k):\n",
        "    print(f\"Processing fold #{i}\")\n",
        "    val_data = train_data[i * num_val_samples: (i + 1) * num_val_samples]\n",
        "    val_targets = train_targets[i * num_val_samples: (i + 1) * num_val_samples]\n",
        "    partial_train_data = np.concatenate(\n",
        "        [train_data[:i * num_val_samples],\n",
        "         train_data[(i + 1) * num_val_samples:]],\n",
        "        axis=0)\n",
        "    partial_train_targets = np.concatenate(\n",
        "        [train_targets[:i * num_val_samples],\n",
        "         train_targets[(i + 1) * num_val_samples:]],\n",
        "        axis=0)\n",
        "    model = build_model()\n",
        "    history = model.fit(partial_train_data, partial_train_targets,\n",
        "                        validation_data=(val_data, val_targets),\n",
        "                        epochs=num_epochs, batch_size=16, verbose=0)\n",
        "    mae_history = history.history[\"val_mae\"]\n",
        "    all_mae_histories.append(mae_history)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1XV73kt7OzcO"
      },
      "source": [
        "**Building the history of successive mean K-fold validation scores**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "IliQ7TroOzcO"
      },
      "outputs": [],
      "source": [
        "average_mae_history = [\n",
        "    np.mean([x[i] for x in all_mae_histories]) for i in range(num_epochs)]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vmx-FUVoOzcO"
      },
      "source": [
        "**Plotting validation scores**"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt"
      ],
      "metadata": {
        "id": "X4Hcx4-eT_w_"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "g0rsBiWCOzcO",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 450
        },
        "outputId": "01386702-310b-4a87-cb6b-de3a631dfd8a"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkAAAAGxCAYAAACKvAkXAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAABNxElEQVR4nO3deVxVZf4H8M9duJf9ArIrCChuqGguhGnLSCKVppmaMaNp5WTa1GhN2aZNTdY0o9Xk0G+alJYpU3Mrl3LF3FfcF0AQlF2997Iv9z6/P5SjVwFB4ZwrfN6v13m9vGe73/uo3A/P85xzVEIIASIiIqJWRK10AURERERyYwAiIiKiVocBiIiIiFodBiAiIiJqdRiAiIiIqNVhACIiIqJWhwGIiIiIWh0GICIiImp1tEoXYI+sViuys7Ph5uYGlUqldDlERETUAEIIFBUVITAwEGp1/X08DEC1yM7ORlBQkNJlEBER0S3IyspCu3bt6t1H0QA0Z84cLFu2DCdPnoSTkxMGDBiADz/8EJ07d5b2KS8vx4wZM7Bo0SJUVFQgNjYW//73v+Hn51fneYUQmDVrFr744gsYjUbcc889SEhIQHh4eIPqcnNzA3C5Ad3d3W/vQxIREZEszGYzgoKCpO/x+qiUfBbY0KFD8cQTT6Bfv36orq7G66+/jqNHj+L48eNwcXEBAEyZMgWrV69GYmIiDAYDpk2bBrVaje3bt9d53g8//BBz5szBV199hdDQULz11ls4cuQIjh8/DkdHx5vWZTabYTAYYDKZGICIiIjuEI35/lY0AF2voKAAvr6+SEpKwr333guTyQQfHx989913ePzxxwEAJ0+eRNeuXbFz507cfffdN5xDCIHAwEDMmDEDL7/8MgDAZDLBz88PiYmJeOKJJ25aBwMQERHRnacx3992dRWYyWQCAHh5eQEA9u/fj6qqKsTExEj7dOnSBcHBwdi5c2et50hPT0dubq7NMQaDAVFRUXUeU1FRAbPZbLMQERFRy2U3AchqteKll17CPffcg+7duwMAcnNzodPp4OHhYbOvn58fcnNzaz1Pzfrr5wjVd8ycOXNgMBikhROgiYiIWja7CUBTp07F0aNHsWjRItnfe+bMmTCZTNKSlZUlew1EREQkH7sIQNOmTcPPP/+MzZs321y25u/vj8rKShiNRpv98/Ly4O/vX+u5atbn5eU1+Bi9Xg93d3ebhYiIiFouRQOQEALTpk3D8uXLsWnTJoSGhtps79OnDxwcHLBx40Zp3alTp5CZmYno6OhazxkaGgp/f3+bY8xmM3bv3l3nMURERNS6KBqApk6dim+//Rbfffcd3NzckJubi9zcXJSVlQG4PHn56aefxvTp07F582bs378fEydORHR0tM0VYF26dMHy5csBACqVCi+99BLee+89rFq1CkeOHMH48eMRGBiIESNGKPExiYiIyM4oeiPEhIQEAMD9999vs37hwoV46qmnAADz5s2DWq3GqFGjbG6EeK1Tp05JV5ABwF/+8heUlJRg8uTJMBqNGDhwINatW9egewARERFRy2dX9wGyF7wPEBER0Z3njr0PEBEREZEcGICIiIio1WEAIiIiolZH0UnQrU1ReRVMZVVw1mnh5aJTuhwiIqJWiz1AMvp651kM/HAzPlx7UulSiIiIWjUGIBmpVSoAgIUX3hERESmKAUhGWvWVAGRlACIiIlISA5CM1AxAREREdoEBSEaay/mHQ2BEREQKYwCSkUZzubktFgYgIiIiJTEAyUjDSdBERER2gQFIRlc6gGDlHCAiIiJFMQDJSKO+3NzVDEBERESKYgCSkdQDxCEwIiIiRTEAyUi6ESJ7gIiIiBTFACQjLYfAiIiI7AIDkIw4CZqIiMg+MADJiM8CIyIisg8MQDLSajgHiIiIyB4wAMmIk6CJiIjsAwOQjDR8GCoREZFdYACSEQMQERGRfWAAkhGfBUZERGQfGIBkVNMDxMvgiYiIlMUAJKOaAMQbIRIRESmLAUhG7AEiIiKyDwxAMuKNEImIiOwDA5CMeCNEIiIi+8AAJCMNb4RIRERkFxiAZKTmfYCIiIjsAgOQjNgDREREZB8YgGQk3Qmak6CJiIgUxQAko6uXwStcCBERUSvHACSjqzdCZAIiIiJSEgOQjKQeIAEIDoMREREphgFIRjWToIHLIYiIiIiUwQAko5rL4AEOgxERESmJAUhG2msCEPMPERGRchQNQFu3bsWwYcMQGBgIlUqFFStW2GxXqVS1Lh999FGd55w9e/YN+3fp0qWZP0nDaK4JQLwUnoiISDmKBqCSkhJERkZi/vz5tW7PycmxWRYsWACVSoVRo0bVe96IiAib47Zt29Yc5Tea+po5QBYLAxAREZFStEq+eVxcHOLi4urc7u/vb/N65cqVeOCBBxAWFlbvebVa7Q3H2gMte4CIiIjswh0zBygvLw+rV6/G008/fdN9U1JSEBgYiLCwMMTHxyMzM7Pe/SsqKmA2m22W5nDtJGg+DoOIiEg5d0wA+uqrr+Dm5obHHnus3v2ioqKQmJiIdevWISEhAenp6Rg0aBCKiorqPGbOnDkwGAzSEhQU1NTlSzR8ICoREZHi7pgAtGDBAsTHx8PR0bHe/eLi4jB69Gj07NkTsbGxWLNmDYxGIxYvXlznMTNnzoTJZJKWrKyspi5fwueBERERKU/ROUAN9dtvv+HUqVP44YcfGn2sh4cHOnXqhNTU1Dr30ev10Ov1t1Nig9XcDNHKHiAiIiLF3BE9QF9++SX69OmDyMjIRh9bXFyMtLQ0BAQENENljXf1eWAMQEREREpRNAAVFxcjOTkZycnJAID09HQkJyfbTFo2m81YsmQJnnnmmVrPMXjwYHz22WfS65dffhlJSUnIyMjAjh07MHLkSGg0GowbN65ZP0tDcQ4QERGR8hQdAtu3bx8eeOAB6fX06dMBABMmTEBiYiIAYNGiRRBC1Blg0tLSUFhYKL0+d+4cxo0bhwsXLsDHxwcDBw7Erl274OPj03wfpBGuPhCVAYiIiEgpKsHHkt/AbDbDYDDAZDLB3d29Sc/d970NKCyuwJo/DUK3wKY9NxERUWvWmO/vO2IOUEuiZQ8QERGR4hiAZMY5QERERMpjAJKZ+kqL8yowIiIi5TAAyUy6DxCHwIiIiBTDACQzDoEREREpjwFIZgxAREREymMAkplaxQBERESkNAYgmWk1fBgqERGR0hiAZFYzCdpiYQAiIiJSCgOQzNRq9gAREREpjQFIZtKdoDkHiIiISDEMQDKrmQTNGyESEREphwFIZnwaPBERkfIYgGTG+wAREREpjwFIZjUBiENgREREymEAkpn0LDAGICIiIsUwAMlMw8vgiYiIFMcAJDPOASIiIlIeA5DM1AxAREREimMAkpmWAYiIiEhxDEAy0/Bp8ERERIpjAJIZnwVGRESkPAYgmfFZYERERMpjAJKZmjdCJCIiUhwDkMx4I0QiIiLlMQDJjDdCJCIiUh4DkMzU0lVgChdCRETUijEAyUxzpcWt7AEiIiJSDAOQzHgnaCIiIuUxAMmMN0IkIiJSHgOQzGomQXMIjIiISDkMQDJTsweIiIhIcQxAMmMPEBERkfIYgGSm4SRoIiIixTEAyYz3ASIiIlIeA5DMeB8gIiIi5TEAyYyToImIiJTHACQzPguMiIhIeYoGoK1bt2LYsGEIDAyESqXCihUrbLY/9dRTUKlUNsvQoUNvet758+cjJCQEjo6OiIqKwp49e5rpEzSedBUYe4CIiIgUo2gAKikpQWRkJObPn1/nPkOHDkVOTo60fP/99/We84cffsD06dMxa9YsHDhwAJGRkYiNjUV+fn5Tl39LOARGRESkPK2Sbx4XF4e4uLh699Hr9fD392/wOefOnYtnn30WEydOBAB8/vnnWL16NRYsWIDXXnvttuptCrwPEBERkfLsfg7Qli1b4Ovri86dO2PKlCm4cOFCnftWVlZi//79iImJkdap1WrExMRg586ddR5XUVEBs9lsszQXPguMiIhIeXYdgIYOHYqvv/4aGzduxIcffoikpCTExcXBYrHUun9hYSEsFgv8/Pxs1vv5+SE3N7fO95kzZw4MBoO0BAUFNennuJb0NHjmHyIiIsUoOgR2M0888YT05x49eqBnz57o0KEDtmzZgsGDBzfZ+8ycORPTp0+XXpvN5mYLQdJ9gNgDREREpBi77gG6XlhYGLy9vZGamlrrdm9vb2g0GuTl5dmsz8vLq3cekV6vh7u7u83SXDgJmoiISHl3VAA6d+4cLly4gICAgFq363Q69OnTBxs3bpTWWa1WbNy4EdHR0XKVWS/eB4iIiEh5igag4uJiJCcnIzk5GQCQnp6O5ORkZGZmori4GK+88gp27dqFjIwMbNy4EY8++ig6duyI2NhY6RyDBw/GZ599Jr2ePn06vvjiC3z11Vc4ceIEpkyZgpKSEumqMKXVTILmEBgREZFyFJ0DtG/fPjzwwAPS65p5OBMmTEBCQgIOHz6Mr776CkajEYGBgRgyZAjeffdd6PV66Zi0tDQUFhZKr8eOHYuCggK8/fbbyM3NRa9evbBu3bobJkYrRc0eICIiIsWphOA38fXMZjMMBgNMJlOTzwfacDwPz3y9D5HtDFg5bWCTnpuIiKg1a8z39x01B6gl4BwgIiIi5TEAyUwaArMqXAgREVErxgAkM06CJiIiUh4DkMzUV1qcQ2BERETKYQCSGXuAiIiIlMcAJDNOgiYiIlIeA5DMrk6CZgAiIiJSCgOQzGqGwNgBREREpBwGIJlp2ANERESkOAYgmUlPg2cXEBERkWIYgGRW0wPEq8CIiIiUwwAkMw3vA0RERKQ4BiCZSUNg7AEiIiJSDAOQzDgERkREpDwGIJlxEjQREZHyGIBkppZ6gBQuhIiIqBVjAJKZhj1AREREimMAkpn0NHjOASIiIlIMA5DManqAAE6EJiIiUgoDkMxqrgIDOAxGRESkFAYgmamvDUDsASIiIlIEA5DMbIbA2ANERESkCAYgmWnYA0RERKQ4BiCZqW0mQStYCBERUSvGACQzToImIiJSXoMD0N///neUlZVJr7dv346KigrpdVFREZ5//vmmra4Fuib/cAiMiIhIIQ0OQDNnzkRRUZH0Oi4uDufPn5del5aW4v/+7/+atroWSKVSSSGIk6CJiIiU0eAAJK77sr7+NTVczTAYe4CIiIiUwTlACpCeCM8AREREpAgGIAXU9ABxCIyIiEgZ2sbs/N///heurq4AgOrqaiQmJsLb2xsAbOYHUf007AEiIiJSVIMDUHBwML744gvptb+/P7755psb9qGbU7MHiIiISFENDkAZGRnNWEbrcnUStMKFEBERtVJNNgfIaDTis88+a6rTtWicBE1ERKSs2w5AGzduxJNPPomAgADMmjWrKWpq8TRXWp1DYERERMq4pQCUlZWFv/71rwgNDcWQIUOgUqmwfPly5ObmNnV9LRInQRMRESmrwQGoqqoKS5YsQWxsLDp37ozk5GR89NFHUKvVeOONNzB06FA4ODg0Z60tRs0kaD4LjIiISBkNngTdtm1bdOnSBb///e+xaNEieHp6AgDGjRvXbMW1VNJ9gNgDREREpIgG9wBVV1dDpVJBpVJBo9E0yZtv3boVw4YNQ2BgIFQqFVasWCFtq6qqwquvvooePXrAxcUFgYGBGD9+PLKzs+s95+zZs6U6a5YuXbo0Sb1NhUNgREREympwAMrOzsbkyZPx/fffw9/fH6NGjcLy5cuhUqlufnAdSkpKEBkZifnz59+wrbS0FAcOHMBbb72FAwcOYNmyZTh16hSGDx9+0/NGREQgJydHWrZt23bLNTYHDoEREREpq8FDYI6OjoiPj0d8fDzS0tKwcOFC/OlPf0J1dTX+9re/4amnnsLvfve7RvUOxcXFIS4urtZtBoMB69evt1n32WefoX///sjMzKz3potarRb+/v4NrkNuNT1AVt4HiIiISBG3dBVYhw4d8N577+Hs2bNYvXo1Kioq8Mgjj8DPz6+p67NhMpmgUqng4eFR734pKSkIDAxEWFgY4uPjkZmZWe/+FRUVMJvNNktzYg8QERGRsm7rPkBqtRpxcXFYunQpzp07h9dff72p6rpBeXk5Xn31VYwbNw7u7u517hcVFYXExESsW7cOCQkJSE9Px6BBg+p9VtmcOXNgMBikJSgoqDk+goT3ASIiIlJWk90J2sfHB9OnT2+q09moqqrCmDFjIIRAQkJCvfvGxcVh9OjR6NmzJ2JjY7FmzRoYjUYsXry4zmNmzpwJk8kkLVlZWU39EWxcHQJjACIiIlJCg+cAhYWFNWi/M2fO3HIxtakJP2fPnsWmTZvq7f2pjYeHBzp16oTU1NQ699Hr9dDr9bdbaoNJQ2AMQERERIpo1MNQ27dvjyeffBK+vr7NWZOkJvykpKRg8+bNaNOmTaPPUVxcjLS0NPzhD39ohgpvjdQDxCEwIiIiRTQ4AP3www9YsGAB5s6di7i4OEyaNAkPPfQQ1OpbH0UrLi626ZlJT09HcnIyvLy8EBAQgMcffxwHDhzAzz//DIvFIj1qw8vLCzqdDgAwePBgjBw5EtOmTQMAvPzyyxg2bBjat2+P7OxszJo1CxqNxq5u2Kjm0+CJiIgU1eD0Mnr0aKxduxapqano06cP/vznPyMoKAivvfYaUlJSbunN9+3bh969e6N3794AgOnTp6N37954++23cf78eaxatQrnzp1Dr169EBAQIC07duyQzpGWlobCwkLp9blz5zBu3Dh07twZY8aMQZs2bbBr1y74+PjcUo3NQboRInuAiIiIFKES4ta/hZOSkjB79mxs3boVhYWF0uMx7nRmsxkGgwEmk6nRc44a4vf/3Y1tqYX4eGwvjOjdtsnPT0RE1Bo15vu7wUNg1yovL8fSpUuxYMEC7N69G6NHj4azs/MtFdsacRI0ERGRshoVgHbv3o0vv/wSixcvRlhYGCZNmoQff/yxxfT8yEVz5ekhHAIjIiJSRoMDUEREBPLz8/Hkk08iKSkJkZGRzVlXi8anwRMRESmrwQHoxIkTcHFxwddff41vvvmmzv0uXrzYJIW1ZGpOgiYiIlJUgwPQwoULm7OOVoU9QERERMpqcACaMGFCc9bRqnASNBERkbKa7Flg1HBX7wOkcCFEREStFAOQAjgERkREpCwGIAVwEjQREZGyGIAUoLnS6pwDREREpAwGIAXU9ABxCIyIiEgZjX4UhsViQWJiIjZu3Ij8/HxYrbaPNN+0aVOTFddSSVeBcQiMiIhIEY0OQC+++CISExPx8MMPo3v37lBd6c2ghtOwB4iIiEhRjQ5AixYtwuLFi/HQQw81Rz2tgoY9QERERIpq9BwgnU6Hjh07NkctrYZ0FZj1JjsSERFRs2h0AJoxYwY++eQTCPZe3LKaq8CsbEMiIiJFNHoIbNu2bdi8eTPWrl2LiIgIODg42GxftmxZkxXXUvFRGERERMpqdADy8PDAyJEjm6OWVkN6FAYDEBERkSIaHYD4VPjbJz0Kg0NgREREimh0AKpRUFCAU6dOAQA6d+4MHx+fJiuqpVOzB4iIiEhRjZ4EXVJSgkmTJiEgIAD33nsv7r33XgQGBuLpp59GaWlpc9TY4rAHiIiISFmNDkDTp09HUlISfvrpJxiNRhiNRqxcuRJJSUmYMWNGc9TY4mg4CZqIiEhRjR4C+/HHH7F06VLcf//90rqHHnoITk5OGDNmDBISEpqyvhaJ9wEiIiJSVqN7gEpLS+Hn53fDel9fXw6BNRDvA0RERKSsRgeg6OhozJo1C+Xl5dK6srIyvPPOO4iOjm7S4loqToImIiJSVqOHwD755BPExsaiXbt2iIyMBAAcOnQIjo6O+OWXX5q8wJaIzwIjIiJSVqMDUPfu3ZGSkoL//e9/OHnyJABg3LhxiI+Ph5OTU5MX2BJJV4GxB4iIiEgRt3QfIGdnZzz77LNNXUurwSEwIiIiZTUoAK1atQpxcXFwcHDAqlWr6t13+PDhTVJYS3b1PkAKF0JERNRKNSgAjRgxArm5ufD19cWIESPq3E+lUsFisTRVbS1WzbPAeBUYERGRMhoUgKxWa61/plvDp8ETEREpq9GXwX/99deoqKi4YX1lZSW+/vrrJimqpeN9gIiIiJTV6AA0ceJEmEymG9YXFRVh4sSJTVJUS8dJ0ERERMpqdAASQkB15Qv8WufOnYPBYGiSolo6PguMiIhIWQ2+DL53795QqVRQqVQYPHgwtNqrh1osFqSnp2Po0KHNUmRLw0nQREREympwAKq5+is5ORmxsbFwdXWVtul0OoSEhGDUqFFNXmBLxEnQREREympwAJo1axYAICQkBGPHjoWjo2OzFdXS1fQAWZh/iIiIFNHoO0FPmDChOepoVfgoDCIiImU1ehK0xWLBP/7xD/Tv3x/+/v7w8vKyWRpj69atGDZsGAIDA6FSqbBixQqb7UIIvP322wgICICTkxNiYmKQkpJy0/POnz8fISEhcHR0RFRUFPbs2dOoupobh8CIiIiU1egA9M4772Du3LkYO3YsTCYTpk+fjsceewxqtRqzZ89u1LlKSkoQGRmJ+fPn17r973//Oz799FN8/vnn2L17N1xcXBAbG4vy8vI6z/nDDz9g+vTpmDVrFg4cOIDIyEjExsYiPz+/UbU1J06CJiIiUphopLCwMPHzzz8LIYRwdXUVqampQgghPvnkEzFu3LjGnk4CQCxfvlx6bbVahb+/v/joo4+kdUajUej1evH999/XeZ7+/fuLqVOnSq8tFosIDAwUc+bMaXAtJpNJABAmk6lxH6KBtqcWiPav/ixi/rmlWc5PRETUGjXm+7vRPUC5ubno0aMHAMDV1VW6KeIjjzyC1atXN1kwS09PR25uLmJiYqR1BoMBUVFR2LlzZ63HVFZWYv/+/TbHqNVqxMTE1HkMAFRUVMBsNtsszenqJGj2ABERESmh0QGoXbt2yMnJAQB06NABv/76KwBg79690Ov1TVZYbm4uAMDPz89mvZ+fn7TteoWFhbBYLI06BgDmzJkDg8EgLUFBQbdZff04CZqIiEhZjQ5AI0eOxMaNGwEAL7zwAt566y2Eh4dj/PjxmDRpUpMXKIeZM2fCZDJJS1ZWVrO+nzQJmj1AREREimj0ZfAffPCB9OexY8ciODgYO3fuRHh4OIYNG9Zkhfn7+wMA8vLyEBAQIK3Py8tDr169aj3G29sbGo0GeXl5Nuvz8vKk89VGr9c3ae/VzUiToK2yvSURERFdo9E9QNeLjo7G9OnTmzT8AEBoaCj8/f2l3iYAMJvN2L17N6Kjo2s9RqfToU+fPjbHWK1WbNy4sc5jlMBngRERESmrQT1Aq1atavAJhw8f3uB9i4uLkZqaKr1OT09HcnIyvLy8EBwcjJdeegnvvfcewsPDERoairfeeguBgYHSYzkAYPDgwRg5ciSmTZsGAJg+fTomTJiAvn37on///vj4449RUlJiV0+qV3MSNBERkaIaFICuDRwAoFKpIK778q55QrzFYmnwm+/btw8PPPCA9Hr69OkALt9tOjExEX/5y19QUlKCyZMnw2g0YuDAgVi3bp3NYzjS0tJQWFgovR47diwKCgrw9ttvIzc3F7169cK6detumBitJE6CJiIiUpZKXJ9kbmLDhg149dVX8f7770vDSjt37sSbb76J999/Hw8++GCzFCons9kMg8EAk8kEd3f3Jj9/an4RYuZuhYezA5LfHtLk5yciImqNGvP93ehJ0C+99BI+//xzDBw4UFoXGxsLZ2dnTJ48GSdOnGh8xa2MNATGHiAiIiJFNHoSdFpaGjw8PG5YbzAYkJGR0QQltXwcAiMiIlJWowNQv379MH36dJtLzfPy8vDKK6+gf//+TVpcS8VJ0ERERMpqdABasGABcnJyEBwcjI4dO6Jjx44IDg7G+fPn8eWXXzZHjS3O1R4ghQshIiJqpRo9B6hjx444fPgw1q9fj5MnTwIAunbtipiYGOlKMKqfhneCJiIiUlSjAxBw+ZL3IUOGYMgQXsF0K2pyIidBExERKaNBAejTTz/F5MmT4ejoiE8//bTeff/0pz81SWEtmeaanjKrVUjPBiMiIiJ5NCgAzZs3D/Hx8XB0dMS8efPq3E+lUjEANYDmmsBjEQJqMAARERHJqUEBKD09vdY/0625tsfHYhVw0ChYDBERUSt02w9DpcazGQLjRGgiIiLZNagHqOYZXQ0xd+7cWy6mtdBc1wNERERE8mpQADp48GCDTsbL4BtGbTMJWsFCiIiIWqkGBaDNmzc3dx2tyvWToImIiEhenAOkgGuveucQGBERkfxu6UaI+/btw+LFi5GZmYnKykqbbcuWLWuSwloylUoFtQqwCkCwB4iIiEh2je4BWrRoEQYMGIATJ05g+fLlqKqqwrFjx7Bp0yYYDIbmqLFF4uMwiIiIlNPoAPT+++9j3rx5+Omnn6DT6fDJJ5/g5MmTGDNmDIKDg5ujxhZJeiI8h8CIiIhk1+gAlJaWhocffhgAoNPpUFJSApVKhT//+c/4z3/+0+QFtlR8IjwREZFyGh2APD09UVRUBABo27Ytjh49CgAwGo0oLS1t2upasJqbIXIIjIiISH6NngR97733Yv369ejRowdGjx6NF198EZs2bcL69esxePDg5qixRap5HAaHwIiIiOTX4AB09OhRdO/eHZ999hnKy8sBAG+88QYcHBywY8cOjBo1Cm+++WazFdrSSENg7AEiIiKSXYMDUM+ePdGvXz8888wzeOKJJwAAarUar732WrMV15JxEjQREZFyGjwHKCkpCREREZgxYwYCAgIwYcIE/Pbbb81ZW4umudLyDEBERETya3AAGjRoEBYsWICcnBz861//QkZGBu677z506tQJH374IXJzc5uzzhanZhI0h8CIiIjk1+irwFxcXDBx4kQkJSXh9OnTGD16NObPn4/g4GAMHz68OWpskTgJmoiISDm39Sywjh074vXXX8ebb74JNzc3rF69uqnqavE4CZqIiEg5t/QsMADYunUrFixYgB9//BFqtRpjxozB008/3ZS1tWjSfYB4I0QiIiLZNSoAZWdnIzExEYmJiUhNTcWAAQPw6aefYsyYMXBxcWmuGlskDoEREREpp8EBKC4uDhs2bIC3tzfGjx+PSZMmoXPnzs1ZW4vGSdBERETKaXAAcnBwwNKlS/HII49Ao9E0Z02tAnuAiIiIlNPgALRq1armrKPVke4DxB4gIiIi2d3WVWB066QhMPYAERERyY4BSCEcAiMiIlIOA5BCOAmaiIhIOQxACqm5EWI1e4CIiIhkxwCkEJ32ctNX8U6IREREsmMAUoj2Sg9QlYU9QERERHJjAFKI9sp18NUMQERERLKz+wAUEhIClUp1wzJ16tRa909MTLxhX0dHR5mrvjmdhkNgRERESrnlh6HKZe/evbBYLNLro0eP4sEHH8To0aPrPMbd3R2nTp2SXquuXHFlT7SamiEwBiAiIiK52X0A8vHxsXn9wQcfoEOHDrjvvvvqPEalUsHf37+5S7stWvWVITBeBUZERCQ7ux8Cu1ZlZSW+/fZbTJo0qd5eneLiYrRv3x5BQUF49NFHcezYsXrPW1FRAbPZbLM0N532Sg9QNXuAiIiI5HZHBaAVK1bAaDTiqaeeqnOfzp07Y8GCBVi5ciW+/fZbWK1WDBgwAOfOnavzmDlz5sBgMEhLUFBQM1Rvq6YHqIo9QERERLJTCXHn3Io4NjYWOp0OP/30U4OPqaqqQteuXTFu3Di8++67te5TUVGBiooK6bXZbEZQUBBMJhPc3d1vu+7avPPTMSzcnoHn7++Avwzt0izvQURE1JqYzWYYDIYGfX/b/RygGmfPnsWGDRuwbNmyRh3n4OCA3r17IzU1tc599Ho99Hr97ZbYKLwKjIiISDl3zBDYwoUL4evri4cffrhRx1ksFhw5cgQBAQHNVNmtuXoV2B3TAUdERNRi3BEByGq1YuHChZgwYQK0WttOq/Hjx2PmzJnS67/+9a/49ddfcebMGRw4cAC///3vcfbsWTzzzDNyl12vq1eBsQeIiIhIbnfEENiGDRuQmZmJSZMm3bAtMzMTavXVHHfp0iU8++yzyM3NhaenJ/r06YMdO3agW7ducpZ8U9KzwKrZA0RERCS3OyIADRkyBHXN1d6yZYvN63nz5mHevHkyVHV7pGeBsQeIiIhIdnfEEFhL5MBngRERESmGAUghDnwUBhERkWIYgBSilS6DZw8QERGR3BiAFCINgXEOEBERkewYgBTCITAiIiLlMAApRHoWGIfAiIiIZMcApJCaHqBq9gARERHJjgFIIQ6cBE1ERKQYBiCFaDkHiIiISDEMQAq5ehUYe4CIiIjkxgCkEF4FRkREpBwGIIVIT4PnHCAiIiLZMQAppGYIrJI9QERERLJjAFIIL4MnIiJSDgOQQrR8GjwREZFiGIAUUtMDxCEwIiIi+TEAKYSXwRMRESmHAUghWvXlHiCLVcDKEERERCQrBiCFOGivNn2VlcNgREREcmIAUoiD+mrTcyI0ERGRvBiAFFLzLDCAAYiIiEhuDEAKqZkDBPBKMCIiIrkxAClEpVJdvRki5wARERHJigFIQXweGBERkTIYgBTEmyESEREpgwFIQQ58HAYREZEiGIAUVHMlWBV7gIiIiGTFAKSgmh4gBiAiIiJ5MQApSHclAFVWMwARERHJiQFIQU46DQCgtMqicCVEREStCwOQgpyvBKCySgYgIiIiOTEAKchJpwUAlDIAERERyYoBSEEuUg9QtcKVEBERtS4MQAqqmQNUwh4gIiIiWTEAKahmDhCHwIiIiOTFAKQg5ytzgDgERkREJC8GIAWxB4iIiEgZDEAKYgAiIiJShl0HoNmzZ0OlUtksXbp0qfeYJUuWoEuXLnB0dESPHj2wZs0amaptvKuXwXMIjIiISE52HYAAICIiAjk5OdKybdu2OvfdsWMHxo0bh6effhoHDx7EiBEjMGLECBw9elTGihvO2YE9QEREREqw+wCk1Wrh7+8vLd7e3nXu+8knn2Do0KF45ZVX0LVrV7z77ru466678Nlnn8lYccO56BmAiIiIlGD3ASglJQWBgYEICwtDfHw8MjMz69x3586diImJsVkXGxuLnTt31vseFRUVMJvNNosceCdoIiIiZdh1AIqKikJiYiLWrVuHhIQEpKenY9CgQSgqKqp1/9zcXPj5+dms8/PzQ25ubr3vM2fOHBgMBmkJCgpqss9QH2feCZqIiEgRdh2A4uLiMHr0aPTs2ROxsbFYs2YNjEYjFi9e3KTvM3PmTJhMJmnJyspq0vPXxYlzgIiIiBShVbqAxvDw8ECnTp2Qmppa63Z/f3/k5eXZrMvLy4O/v3+959Xr9dDr9U1WZ0O56DkERkREpAS77gG6XnFxMdLS0hAQEFDr9ujoaGzcuNFm3fr16xEdHS1HeY129T5A1RBCKFwNERFR62HXAejll19GUlISMjIysGPHDowcORIajQbjxo0DAIwfPx4zZ86U9n/xxRexbt06/POf/8TJkycxe/Zs7Nu3D9OmTVPqI9Sr5mGoVgFUVFsVroaIiKj1sOshsHPnzmHcuHG4cOECfHx8MHDgQOzatQs+Pj4AgMzMTKjVVzPcgAED8N133+HNN9/E66+/jvDwcKxYsQLdu3dX6iPUq+Y+QABQVmmB4zWviYiIqPmoBMdebmA2m2EwGGAymeDu7t6s79XpzbWorLZi26sPoJ2nc7O+FxERUUvWmO9vux4Caw3cHR0AAMbSKoUrISIiaj0YgBTWxkUHALhUWqlwJURERK0HA5DCvK4EoIslDEBERERyYQBSmJfr5QB0oZgBiIiISC4MQArzcuYQGBERkdwYgBRWMwR2gUNgREREsmEAUlibK0NgFzkERkREJBsGIIVxEjQREZH8GIAUVjMH6CLnABEREcmGAUhhNVeBsQeIiIhIPgxACvO65kaIFiufSkJERCQHBiCFtXHRw0GjghBArrlc6XKIiIhaBQYghWnUKgR5XX4IanpBicLVEBERtQ4MQHYgtI0LACD9AgMQERGRHBiA7ECI9+UAlFHIAERERCQHBiA7EMoAREREJCsGIDtQE4DSGYCIiIhkwQBkBzr6ugIAMi6UwFxepXA1RERELR8DkB3wc3dE+zbOsApgX8ZFpcshIiJq8RiA7ER0WBsAwK4zDEBERETNjQHITtx9JQBtTy1UuBIiIqKWjwHITgwM94ZKBRzLNuO8sUzpcoiIiFo0BiA74e2qR7/2XgCAX4/lKlwNERFRy8YAZEeGRPgBAH5hACIiImpWDEB2JDbCHwCwJ/0iLhRXKFwNERFRy8UAZEeCvJwREegOqwA2nshXuhwiIqIWiwHIzgy90gv0xW9nUFFtUbgaIiKilokByM78/u728HbVISW/GAu2ZShdDhERUYvEAGRnPF10eCW2MwBg8b4sCCEUroiIiKjlYQCyQ4/0DISzToP0whIcyDQqXQ4REVGLwwBkh1z0Wgztfnku0LID5xSuhoiIqOVhALJTo+5qBwD46VA2J0MTERE1MQYgO3V3WBsEGBxhLq/Gkn3sBSIiImpKDEB2SqNW4ZlBYQCA99ecQHKWUdmCiIiIWhAGIDv21IAQDOjQBqWVFjz5xS6k5hcrXRIREVGLwABkxzRqFb4Y3xf9Q7xQWmnBH7/Zh5O5ZqXLIiIiuuMxANk5F70Wn47rDU9nB6QVlGB0wk5kG8uULouIiOiOxgB0B/A3OOKnFwaiW4A7iiqq8fz/DjAEERGRosqrLCivathVykfPm5C4PR1JpwtgtdrHDX7tOgDNmTMH/fr1g5ubG3x9fTFixAicOnWq3mMSExOhUqlsFkdHR5kqbj7tPJ3x6bjecNVrkZxlxGP/3oGfD2ej2mJVujQioltiLq9CcUW1Yu8vhEBhccUdecf9xv7s/3bXWUxYsAf55vI69xFCYNmBc/j7upOw1BFSam7LYi6vwpB5WzH4n0koLK5Aan4xKqsv12S1Cpu/19N5RXh0/nbM/uk4JizYgw/XnWxU7c1Fq3QB9UlKSsLUqVPRr18/VFdX4/XXX8eQIUNw/PhxuLi41Hmcu7u7TVBSqVRylNvsOvq64qcXBuLpxL04U1iCad8dxP2dffDF+L5w0Nh1liUiAnD5S7ai2gqLVWDI3K3QO6ix/s/3Qae9/DMsx1SGnw/l4A/R7eHooMHW0wUor7JgyJUHRdfFVFYFRwc1dBo11h/PQ78QL3i66Oo9ZvG+LLz64xHMeawHxvUPvuXPVFFtwb6MS+gb4gmdRo1/b0nDsWwT/jaiB9wctdBe+flcZbFCBcAiBHafuQitWoUBHb1RUlENjVoFRwcNrFaBfWcvIdzX9Yb6LVaBpfuzsP54HracKkBHX1e8N6I7+oZ4ITW/GCdyzOgX4gV/gyOKyquQdLoA93TwhrNegw/XnURReTVmLDmENx/uhn9tSoGLTovjOWb0DvbAxHtCsXR/FuZvTgMARAZ5IDbCHyl5RTiWbUaf9p7416YUrEzOxu+6+GLDiTxUWS6HpL7vbQAAeLvqENLGBRkXSnCptAofjuqJTn6uePKL3TaB6j+/nUFFtRV/GdoZzjrlYohK3EHRt6CgAL6+vkhKSsK9995b6z6JiYl46aWXYDQab/l9zGYzDAYDTCYT3N3db/k8zSXHVIYP1p7EyuRsAEDf9p6YH38XtGoV2rjqFa7u9hVXVEOvVTPUUbMqq7Qgv6gc7dvU/ctUS5NWUAxPZx28bhIMrmexCpRXWXDeWAZTWRX6hXjVut9vKQXYl3EJk+8Ng4v+xi82IQRe+P4gNp7Ix5AIP+ln2DdP98egcB8AwINzk5CSX4zn7++ASQNDcff7G1FtFRjdpx1eju0MP3dHXCyphMHJAe+vOYE1R3LwzvAIvL78KKxC4HddfLF0/+V7pz0ZFYypD3REWw8nAEBhcQVe+/EwLFaBSQND8Ycv90i19Q/1wrQHOsJJp0E7Tyc4OWhw9kIpXlt2BPd39sH0BztJP5MsVoGXlxzCpdJKeDg5YOeZC8gzV+Cx3m3h7uSAxB0Z0nndHbUYEuGPiyWV2H3mAkoqbYeMnhkYiuUHz6PSYsWfYzoh61IpFm7PgFatQqDH5Truau+BMwUl2J1+8YY2ddFpsOCpfnju2/24VFoFAOjg44JcUzlKKi1o5+mEfiFeWH7wfIP+rmuEertgQIc2+N/uzEYdV58lz0Vj+uJkZF28PIVjRK9AfPxE7yY7P9C47+87KgClpqYiPDwcR44cQffu3WvdJzExEc888wzatm0Lq9WKu+66C++//z4iIiLqPG9FRQUqKiqk12azGUFBQXYbgGpsPpmPF74/aNPV6OWiw0M9/DEzris2nMhDVGgbHD5nxDe7zmJCdAhiuvkpWPHNpRUU4+FPf8OD3fzxr3FN8x8jNb8Y7Tyd4OigaZLztVY5pjI4O2hhcHZQupQmMfW7A1hzJAdLnxuAPu09G3xctcWKlcnZ6BXsgQ4+rrXuc/S8CZ4uOrT1cMKlkkp4ODvY9ETnmcuRYyrHrFXHcHeYF2bGdW3Qe5vKqiCEwK4zF/BgN38IIVBpsdb6W3SVxWrzS8Thc5eHzjv5uWH1nwbW2jN+IPMS/vHLKVisAn+Ibo+QNi7416YUXCypxN6MS9J+c8dEwstFhwOZRmQUlmB4ZCD2ZlzEf7elw2IViAzywLwxkQi7pn3OXijBmyuO4reUwhveNzbCD8FezigsrpS+qB0d1Pjr8O74y4+Hpf2CvZzxh7vb429rTjSovQDggc4++Hd8H1RarBiVsOOWbyfSxd8NzjoNisqrYS6vQp654uYHNRO1Cnh6YCiOZZuxI+1Cg48L9nJG5sXSG9aHersgvbAEAOCm16LoumFJvVaNimordFq1NMzV1sMJT/QLwj/XnwYALH0uGueNZdCoVfB3d8SSfefww74sad8hEX54+5Fu+HRjKuZtOI02Ljp8/oc+dYbpW9UiA5DVasXw4cNhNBqxbdu2OvfbuXMnUlJS0LNnT5hMJvzjH//A1q1bcezYMbRr167WY2bPno133nnnhvX2HoCAy4FhVMIOGK8k/5uZ/mAndPJzRXQHbxicrn6RnbtUigCDE7KNZTh63oQhEf44kHkJ7/58HJXVVkx/sFOtXdBWq8D/bT2Dtp5OGB4ZWO97CyGwLbUQ/u6OCPF2qbWH568/HceC7ekAgD1vDIav2+3N3/r1WC4mf7Mf46Pb46+P1h6a7c2xbBNcdFqEeCvbMyGEwLTvDuLweSM+faI3nvxiN9q3ccbaFwfVOawshMDifVno6OvWqFDRGMUV1VhzOAelldU4byyDn7ujdNPQ6+WYylBRZUWItwtKK6vx6cZUPNwjAO08ndD73fUAgAEd2uCZQaHoFmCAv8ER2cYy/HtLKsb2DUaPdgYAwGebUnAitwhP9g/Gkn1ZWJGcjQ4+Ltgw/T4knS7A4n1ZeP2hrmjn6YxDWUaM/Pd2eLno8Xifdvg8KQ2Du/ji4yd6IcdUDotV4Mkvdkm/rQPAphn3wVhWdeVLVgtTaRVKq6oRYLjcc3HuUileXJSM/WevhpCBHb1xOq8I1VaB75+9Gx18XJB0ugC55nKk5BVj0d5MzHiwM84by7DrzAWczC2yaZsno4IR190fPdt5wFmnwScbUvDvLaloqvmpOo0aw3sFoluAOzIvluK73ZmovMU5iwM6tMGp3CJcKKm87boMTg4I9nLGkfOmBu0faHBEcUU1zOW1z1Vq5+mEpweG4oe9WVIbv/FQVymkDenmh0APJ6w5koPYCH8MifDDtpRC/P7u9khISsOiPZlwd3LAyN5t8fXOs7BYBf40OBxDuvkhOcuIC8WVKCguR6CHE7IulqFbgBv+EB0CADCWVmLox78h98q8nvdGdL88x0elwpBuflCpgI83pEj/D/4zvi9+2JsFU1kV/vxgOP75y2kMDPfGsMhA/HQoG+uO5eLV2C6YsHAPMi6UYEg3P4zo1RYPdvPDnoyLiAg04HReEX45mosXBofD3VGL+ZtT4emiQ3xU+xva5kxBMUxlVegdfPVnQXmVBauSs3FfZx/4uTf9/NwWGYCmTJmCtWvXYtu2bXUGmdpUVVWha9euGDduHN59991a97lTe4BqpBUUY3/GJQzu6otvdp3FxxtSbnqMVq1CvxAvjI9uj9T8Yvxz/Wm4OWpRdOU/+aBwb+xOvyilfeDy+O7fRvbAPR294arX4kSOGUv3n8OX2y4Hlsf7tEOvIA8M6xkIg7MDqi1WlFVZ4ObogGxjGRK2pOGbXWel9x8fHYKZD3VBRbUVu89cwM60C/jvlXMBwKxh3TDxnlDpdUW1BQVFFfB3d5TG1K9V85t5aZUFHX1c4arX4sn/7pI+09i+Qegd7IG2nk4Y2NG71i/xg5mXsOpQNqbc1wG+dfznrLZYa33/ppBrKse9f98MlQp465FuGNG7LVyvGUr46VA2jp43oYOvK45nm3F3WBsM7e6PtIJi/Lj/HMb2C5KGdIQQ2HKqAObyKni56PDKksP40+BwPBkVjJO5ZizclgEnnQZvPNwVOcZyPPv1PliEgLujFvd18kW11Yp/bUoFAKhUQM1PiqcGhGDK/R1u+OGVby7Hkv3n8NEvp6DTqPHFhL64r5MPCooqoFYBh84ZkZpfjNgIf1wqrcK3u86irYcT9p29iDBvV0y5vwP+vSUVKXnF6BXkgbMXStHGVYdB4d6wWIFFezORklcs/bC/1tuPdMPyg+dxd5gXXozphGxjGb7bnYnEHRnQa9X45aV78eOBc9LnqY1eq8a0Bzrim11nkV90+efBk1HB6B5owOvLj9z0787HTY/IdgZsOJF/033r4qzTIMDgiMyLpaiyCESFeuGRyEB8viUN5+u58lOjVsFVr4WprGG/CF1LrQKcHDTS0MyQbn5o6+mEhdszbth33thI/HI0D+uO5QIA+rT3tAllf3+8J/qFeGH2qmNIOl1ww/FBXk74x+OR6B/qhW93Z8JcVoUj50xYdywXgQZHPNE/GEIAp/OLsPpwjnTcLy/di/PGUjz79X6phynI0wnrjubi60n9cSzbDAGBEb3bYvaqY/hdFz8MjwzErFXH8P2eq0M4Oq0ai/8YjR5tDfjf7rNYfzwPPm56/JZSiAc6+6CsyoqMwhJUVFsQ6u2Cth7OmHxvGCxC4NdjufBx08NiFfjb6hPQqFVY/MdoBBgu/zw6byxD4vZ0DI9six7tDFh3NAc5pnI8NSCk3nmoptIqqNSAu6MDTGVVsFrFTecuXSvzQim+25MJN0ct/nhvWJP8bLpQXIFKi1UK4HeSFheApk2bhpUrV2Lr1q0IDQ29+QHXGT16NLRaLb7//vsG7W/vc4DqI4TAiuTz0KjV6OLvhoQtaegdfDmUPDgvCYXFlfB21aGwuGG/SUUGeaC0ohop13UbXxuWrtfO0wmRQR7Yn3EJBcUVaN/GGWcKSmrdt5OfK4rLq5FtuvFLzd/dEf+d0BdF5dUoLK7AOz8dQ2Hx5bH/Lv5uaOOqQ2W1QN8QT3g565CQlCZ1496Mm6MWUaFt0KOtAQPD2+BMQQkyL5YiYUsaqq/8CvxgNz8EeTrjhd91hJujFhq1Cv+39Qzm/noaMd18MequdvhdF1+cyitCtUWgrYcT5qw9AV83R3Rva0Cf9p44d6kU7du4IM9cjgCDI1YcPI/0whJkm8qxN+Mi/nhvB3QLdMehLCO8XHSwWAVmrTom1enp7IBHe7VF+zbOiAzywJjPd0r1AZd/y547NhJvrjgKY2kVPJwdENPVDy46DdIKSrAt9cbhhhVT78HEhXukHoiPx/bChhN5+PmaL5yb8XLR4e1HukFAIL2gBLvSL2LPdfMTdBo1wnxcbuh5aG5eLjpcbIKegtp4ODvA392xwZ9Jo1ZhULg3Dp8z3VDTnwaHI9DgiNeXH2lQr4uHswOm3NcBm0/lY9eZy20dHdYG54yl0pwKT2cHlFRYpJ4WnUaNSosV8VHBcNVrUVBUgWVXhph6tjPAVFaFsxdKpWPfG9EDD/cMAADsTLuAL7edwfFsM7JN5fj7qJ4Y0y8I5VUWvLf6OAI9nDDlvg44dM6ED9eexIsx4bg7rA2Ayz+HtqYUYsupfCw7cB6d/d1wXycfjO7TrtZfLE7kmNHW0wnujg7S8XnmCiRnGaF3UOOBzr4ALv+ydyjLiPs7+8LLRYfyKku9Q9vVFivOXiyFr5seeeZy6LUaBHk537yxb6LaYkVFtbXWeU6krBYTgIQQeOGFF7B8+XJs2bIF4eHhjT6HxWJBREQEHnroIcydO7dBx9zJAag+ZwqKcSKnCEO7+yPrYikW7c3CF7+dgcUq8FAPfxicdEg6lY9n7w3DJxtToALw0wsDoVap8O7Px7H2aG6d5/ZwdkBIGxecN5ahoKj2sXFXvRZDuvnh/cd6YMOJPMxaeUzq0nbQqBDs5Yy0ghKE+7oqPsZeGweNSrrqoUZEoDuOZTf93bl1GjUCPBylL6frOTqoUV7V9LdAiAh0h16rxqXSKrT1cEKYjwu+3nm2UecI9nJGR19XbDp56z0hABAfFQy9VoO1R3NQUW3FiF5t8UhkAMK8XeCi1+KZr/Zhd/oFqR1CvV1QWFRhM3+hrrDv5KDBjCGdsGBbOrJN5Xj30Qik5hcjOcuIboHusFgFFl/zEOKuAe5Y+ly09IWXUViCF39IhreLDr7uegwK98Hao7lQAegb4om7gj3hpNPAzVELXzdHnDeWYdPJfDzSIwDJWUaYy6swPDIQKpUK+89ewrID5/Bor7YwODngQnEFAjycUFpZjQkL9sAqgOGRgfj93cHo6OsG4PJlxbvPXMCYfkGwWAX++1s63B21GNMvCJXVVmw8kY+h3f1RUlmNiyWV6OJ/9efYoSwjwnxc4HYlbJw3liHfXI5uge7Qa28ME+VVFhzLNuOuYI9buqJWCNFirsQl+9diAtDzzz+P7777DitXrkTnzp2l9QaDAU5Ol7vmxo8fj7Zt22LOnDkAgL/+9a+4++670bFjRxiNRnz00UdYsWIF9u/fj27dujXofVtqAKrNxZJKXCypQKi3KzTqqz+kyiotqKi2wMP5aldslcWKfRmX0MHHBSdyi+CoVcPg7ACtWo2Ovq7S+X46lA2rEPB1c4SDRoU8cznu7eRzw9U2xtJKzFt/GqfyivD3UZEIbuOMC8UVcNFr8VtKIZ7/336oVSpUXBmGm3hPCF4a3AlfbjuDI+dN2Hf2ktQLFezlDK1ahfcf64F/bUpBtrEcbVx08HB2wKXSKuw/ewkhbZyx+eX7kXS6AMlZRqigwq/Hc3Es2wyVCniwqx/6h3rBx00vdY0v3pt1w1UbMV39oHdQ23TRXyuynQEZF0prHY7oFuCOB7r4wNNZh58O5+BQlhHernp09nfF9tSrkxnX/GkQOvm54stt6fhyW7o0JOPlokPixH4I8XZB1sVSPPbvHaiotmJQuDf+OSYS//zlNHacKUTvIE/4Gxwxuk87aDVqvL3yKNwctdiWUgiruBwM3n+sByYl7pUCxLODQvHGwzf+H/nb6uPIuliG9x/rgZXJ55FjKseX29IR5u0CTxcdLpVU4lJpFdq46PDR6J5o38YFjg5qrEzORo6xHA/39EfWpTIczzajo68rKqut6BfihaLyKry/5gTG9guGm6MWi/dl4fn7O8JZp4GprArd2xrq+Fdr68g5E1z0GoT5uCKjsAQLtqejuKIarw7tAj93RyRnGfHb6QJ8tycTcd0D0D/UEwPDfeCq1+JUbhHSCooR193f5kvaYhU4ct4Ei1Vgw4k8PDMwVJErLMurLNCoVbwikqiBWkwAquu3hoULF+Kpp54CANx///0ICQlBYmIiAODPf/4zli1bhtzcXHh6eqJPnz5477330Lt3w68oak0ByJ5VW6zQqFUwl1fj3KVSdAtwt/k3UfObpcUqbMLb9UylVfhy2xk8GdUe/gbb7vcqixXLDpxDRKCh1i/cC8UVOJZtxvtrTqBbgDveeTRC+s35f7vPIulUAf54Xxh+SylEedXloYYgL2cYSyuxLbUQWrUah84Z8WivQKigQic/V+kzCCFwKq8IHXxc4aBRY/eZC0jckYFADye8+XBXm89aVH55rkSvYA+bK36KyqtgKrvcW3Mrv2UfzLyEzSfz0SXA/YYQUJ/KaiscNCr+Zk9EdqXFBCClMAARERHdeRrz/c1+VSIiImp1GICIiIio1WEAIiIiolaHAYiIiIhaHQYgIiIianUYgIiIiKjVYQAiIiKiVocBiIiIiFodBiAiIiJqdRiAiIiIqNVhACIiIqJWhwGIiIiIWh0GICIiImp1GICIiIio1dEqXYA9EkIAAMxms8KVEBERUUPVfG/XfI/XhwGoFkVFRQCAoKAghSshIiKixioqKoLBYKh3H5VoSExqZaxWK7Kzs+Hm5gaVStVk5zWbzQgKCkJWVhbc3d2b7Lxki+0sD7azfNjW8mA7y6M521kIgaKiIgQGBkKtrn+WD3uAaqFWq9GuXbtmO7+7uzv/c8mA7SwPtrN82NbyYDvLo7na+WY9PzU4CZqIiIhaHQYgIiIianUYgGSk1+sxa9Ys6PV6pUtp0djO8mA7y4dtLQ+2szzspZ05CZqIiIhaHfYAERERUavDAEREREStDgMQERERtToMQERERNTqMADJZP78+QgJCYGjoyOioqKwZ88epUu6o2zduhXDhg1DYGAgVCoVVqxYYbNdCIG3334bAQEBcHJyQkxMDFJSUmz2uXjxIuLj4+Hu7g4PDw88/fTTKC4ulvFT2L85c+agX79+cHNzg6+vL0aMGIFTp07Z7FNeXo6pU6eiTZs2cHV1xahRo5CXl2ezT2ZmJh5++GE4OzvD19cXr7zyCqqrq+X8KHYvISEBPXv2lG4GFx0djbVr10rb2c7N44MPPoBKpcJLL70krWNb377Zs2dDpVLZLF26dJG222UbC2p2ixYtEjqdTixYsEAcO3ZMPPvss8LDw0Pk5eUpXdodY82aNeKNN94Qy5YtEwDE8uXLbbZ/8MEHwmAwiBUrVohDhw6J4cOHi9DQUFFWVibtM3ToUBEZGSl27dolfvvtN9GxY0cxbtw4mT+JfYuNjRULFy4UR48eFcnJyeKhhx4SwcHBori4WNrnueeeE0FBQWLjxo1i37594u677xYDBgyQtldXV4vu3buLmJgYcfDgQbFmzRrh7e0tZs6cqcRHslurVq0Sq1evFqdPnxanTp0Sr7/+unBwcBBHjx4VQrCdm8OePXtESEiI6Nmzp3jxxRel9Wzr2zdr1iwREREhcnJypKWgoEDabo9tzAAkg/79+4upU6dKry0WiwgMDBRz5sxRsKo71/UByGq1Cn9/f/HRRx9J64xGo9Dr9eL7778XQghx/PhxAUDs3btX2mft2rVCpVKJ8+fPy1b7nSY/P18AEElJSUKIy+3q4OAglixZIu1z4sQJAUDs3LlTCHE5rKrVapGbmyvtk5CQINzd3UVFRYW8H+AO4+npKf773/+ynZtBUVGRCA8PF+vXrxf33XefFIDY1k1j1qxZIjIystZt9trGHAJrZpWVldi/fz9iYmKkdWq1GjExMdi5c6eClbUc6enpyM3NtWljg8GAqKgoqY137twJDw8P9O3bV9onJiYGarUau3fvlr3mO4XJZAIAeHl5AQD279+Pqqoqm7bu0qULgoODbdq6R48e8PPzk/aJjY2F2WzGsWPHZKz+zmGxWLBo0SKUlJQgOjqa7dwMpk6diocfftimTQH+m25KKSkpCAwMRFhYGOLj45GZmQnAftuYD0NtZoWFhbBYLDZ/qQDg5+eHkydPKlRVy5KbmwsAtbZxzbbc3Fz4+vrabNdqtfDy8pL2IVtWqxUvvfQS7rnnHnTv3h3A5XbU6XTw8PCw2ff6tq7t76JmG1115MgRREdHo7y8HK6urli+fDm6deuG5ORktnMTWrRoEQ4cOIC9e/fesI3/pptGVFQUEhMT0blzZ+Tk5OCdd97BoEGDcPToUbttYwYgIqrV1KlTcfToUWzbtk3pUlqszp07Izk5GSaTCUuXLsWECROQlJSkdFktSlZWFl588UWsX78ejo6OSpfTYsXFxUl/7tmzJ6KiotC+fXssXrwYTk5OClZWNw6BNTNvb29oNJobZrvn5eXB399foapalpp2rK+N/f39kZ+fb7O9uroaFy9e5N9DLaZNm4aff/4ZmzdvRrt27aT1/v7+qKyshNFotNn/+rau7e+iZhtdpdPp0LFjR/Tp0wdz5sxBZGQkPvnkE7ZzE9q/fz/y8/Nx1113QavVQqvVIikpCZ9++im0Wi38/PzY1s3Aw8MDnTp1Qmpqqt3+e2YAamY6nQ59+vTBxo0bpXVWqxUbN25EdHS0gpW1HKGhofD397dpY7PZjN27d0ttHB0dDaPRiP3790v7bNq0CVarFVFRUbLXbK+EEJg2bRqWL1+OTZs2ITQ01GZ7nz594ODgYNPWp06dQmZmpk1bHzlyxCZwrl+/Hu7u7ujWrZs8H+QOZbVaUVFRwXZuQoMHD8aRI0eQnJwsLX379kV8fLz0Z7Z10ysuLkZaWhoCAgLs999zs0ytJhuLFi0Ser1eJCYmiuPHj4vJkycLDw8Pm9nuVL+ioiJx8OBBcfDgQQFAzJ07Vxw8eFCcPXtWCHH5MngPDw+xcuVKcfjwYfHoo4/Wehl87969xe7du8W2bdtEeHg4L4O/zpQpU4TBYBBbtmyxuZy1tLRU2ue5554TwcHBYtOmTWLfvn0iOjpaREdHS9trLmcdMmSISE5OFuvWrRM+Pj68ZPg6r732mkhKShLp6eni8OHD4rXXXhMqlUr8+uuvQgi2c3O69iowIdjWTWHGjBliy5YtIj09XWzfvl3ExMQIb29vkZ+fL4SwzzZmAJLJv/71LxEcHCx0Op3o37+/2LVrl9Il3VE2b94sANywTJgwQQhx+VL4t956S/j5+Qm9Xi8GDx4sTp06ZXOOCxcuiHHjxglXV1fh7u4uJk6cKIqKihT4NPartjYGIBYuXCjtU1ZWJp5//nnh6ekpnJ2dxciRI0VOTo7NeTIyMkRcXJxwcnIS3t7eYsaMGaKqqkrmT2PfJk2aJNq3by90Op3w8fERgwcPlsKPEGzn5nR9AGJb376xY8eKgIAAodPpRNu2bcXYsWNFamqqtN0e21glhBDN07dEREREZJ84B4iIiIhaHQYgIiIianUYgIiIiKjVYQAiIiKiVocBiIiIiFodBiAiIiJqdRiAiIiIqNVhACIiIqJWhwGIiKgOKpUKK1asULoMImoGDEBEZJeeeuopqFSqG5ahQ4cqXRoRtQBapQsgIqrL0KFDsXDhQpt1er1eoWqIqCVhDxAR2S29Xg9/f3+bxdPTE8Dl4amEhATExcXByckJYWFhWLp0qc3xR44cwe9+9zs4OTmhTZs2mDx5MoqLi232WbBgASIiIqDX6xEQEIBp06bZbC8sLMTIkSPh7OyM8PBwrFq1Stp26dIlxMfHw8fHB05OTggPD78hsBGRfWIAIqI71ltvvYVRo0bh0KFDiI+PxxNPPIETJ04AAEpKShAbGwtPT0/s3bsXS5YswYYNG2wCTkJCAqZOnYrJkyfjyJEjWLVqFTp27GjzHu+88w7GjBmDw4cP46GHHkJ8fDwuXrwovf/x48exdu1anDhxAgkJCfD29pavAYjo1jXbc+aJiG7DhAkThEajES4uLjbL3/72NyGEEADEc889Z3NMVFSUmDJlihBCiP/85z/C09NTFBcXS9tXr14t1Gq1yM3NFUIIERgYKN544406awAg3nzzTel1cXGxACDWrl0rhBBi2LBhYuLEiU3zgYlIVpwDRER264EHHkBCQoLNOi8vL+nP0dHRNtuio6ORnJwMADhx4gQiIyPh4uIibb/nnntgtVpx6tQpqFQqZGdnY/DgwfXW0LNnT+nPLi4ucHd3R35+PgBgypQpGDVqFA4cOIAhQ4ZgxIgRGDBgwC19ViKSFwMQEdktFxeXG4akmoqTk1OD9nNwcLB5rVKpYLVaAQBxcXE4e/Ys1qxZg/Xr12Pw4MGYOnUq/vGPfzR5vUTUtDgHiIjuWLt27brhddeuXQEAXbt2xaFDh1BSUiJt3759O9RqNTp37gw3NzeEhIRg48aNt1WDj48PJkyYgG+//RYff/wx/vOf/9zW+YhIHuwBIiK7VVFRgdzcXJt1Wq1Wmmi8ZMkS9O3bFwMHDsT//vc/7NmzB19++SUAID4+HrNmzcKECRMwe/ZsFBQU4IUXXsAf/vAH+Pn5AQBmz56N5557Dr6+voiLi0NRURG2b9+OF154oUH1vf322+jTpw8iIiJQUVGBn3/+WQpgRGTfGICIyG6tW7cOAQEBNus6d+6MkydPArh8hdaiRYvw/PPPIyAgAN9//z26desGAHB2dsYvv/yCF198Ef369YOzszNGjRqFuXPnSueaMGECysvLMW/ePLz88svw9vbG448/3uD6dDodZs6ciYyMDDg5OWHQoEFYtGhRE3xyImpuKiGEULoIIqLGUqlUWL58OUaMGKF0KUR0B+IcICIiImp1GICIiIio1eEcICK6I3H0nohuB3uAiIiIqNVhACIiIqJWhwGIiIiIWh0GICIiImp1GICIiIio1WEAIiIiolaHAYiIiIhaHQYgIiIianX+Hyjgh/EzThSCAAAAAElFTkSuQmCC\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "plt.plot(range(1, len(average_mae_history) + 1), average_mae_history)\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Validation MAE\")\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WQV7HiSsOzcO"
      },
      "source": [
        "**Plotting validation scores, excluding the first 10 data points**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Ugcb5cLWOzcP",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 449
        },
        "outputId": "8e3b0441-51da-4b02-86df-255822809cbb"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "truncated_mae_history = average_mae_history[10:]\n",
        "plt.plot(range(1, len(truncated_mae_history) + 1), truncated_mae_history)\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Validation MAE\")\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QSpqLS4NOzcP"
      },
      "source": [
        "**Training the final model**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Y3Y7qO2DOzcP",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d083e058-9b3d-4e6c-b073-3bf558774121"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "4/4 [==============================] - 0s 4ms/step - loss: 16.1113 - mae: 2.5235\n"
          ]
        }
      ],
      "source": [
        "model = build_model()\n",
        "model.fit(train_data, train_targets,\n",
        "          epochs=130, batch_size=16, verbose=0)\n",
        "test_mse_score, test_mae_score = model.evaluate(test_data, test_targets)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "hMPTyvzyOzcP",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c51aae27-a29e-4657-d53f-e72b40fdedb6"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2.5235164165496826"
            ]
          },
          "metadata": {},
          "execution_count": 25
        }
      ],
      "source": [
        "test_mae_score"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DcC8hDurOzcP"
      },
      "source": [
        "### Generating predictions on new data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "fH0GDCQPOzcP",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b40c1cab-c15b-4289-cb6f-7b32c128e765"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "4/4 [==============================] - 0s 3ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([7.647238], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 26
        }
      ],
      "source": [
        "predictions = model.predict(test_data)\n",
        "predictions[0]"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## chapter05_fundamentals"
      ],
      "metadata": {
        "id": "qMBJbJFjVMMd"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow.keras.datasets import mnist\n",
        "import numpy as np\n",
        "\n",
        "(train_images, train_labels), _ = mnist.load_data()\n",
        "train_images = train_images.reshape((60000, 28 * 28))\n",
        "train_images = train_images.astype(\"float32\") / 255\n",
        "\n",
        "train_images_with_noise_channels = np.concatenate(\n",
        "    [train_images, np.random.random((len(train_images), 784))], axis=1)\n",
        "\n",
        "train_images_with_zeros_channels = np.concatenate(\n",
        "    [train_images, np.zeros((len(train_images), 784))], axis=1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HBySrCb4VGYJ",
        "outputId": "3bca7d22-9e02-42f7-f293-8605cebd9174"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/mnist.npz\n",
            "11490434/11490434 [==============================] - 0s 0us/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow import keras\n",
        "from tensorflow.keras import layers\n",
        "\n",
        "def get_model():\n",
        "    model = keras.Sequential([\n",
        "        layers.Dense(512, activation=\"relu\"),\n",
        "        layers.Dense(10, activation=\"softmax\")\n",
        "    ])\n",
        "    model.compile(optimizer=\"rmsprop\",\n",
        "                  loss=\"sparse_categorical_crossentropy\",\n",
        "                  metrics=[\"accuracy\"])\n",
        "    return model\n",
        "\n",
        "model = get_model()\n",
        "history_noise = model.fit(\n",
        "    train_images_with_noise_channels, train_labels,\n",
        "    epochs=10,\n",
        "    batch_size=128,\n",
        "    validation_split=0.2)\n",
        "\n",
        "model = get_model()\n",
        "history_zeros = model.fit(\n",
        "    train_images_with_zeros_channels, train_labels,\n",
        "    epochs=10,\n",
        "    batch_size=128,\n",
        "    validation_split=0.2)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_JHLKnV_W7gb",
        "outputId": "df4a0567-a11a-41bb-cdf0-946c2d6d5b97"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "375/375 [==============================] - 7s 7ms/step - loss: 0.6400 - accuracy: 0.8076 - val_loss: 0.2915 - val_accuracy: 0.9068\n",
            "Epoch 2/10\n",
            "375/375 [==============================] - 2s 7ms/step - loss: 0.2623 - accuracy: 0.9183 - val_loss: 0.2409 - val_accuracy: 0.9243\n",
            "Epoch 3/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.1751 - accuracy: 0.9467 - val_loss: 0.1857 - val_accuracy: 0.9467\n",
            "Epoch 4/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.1260 - accuracy: 0.9617 - val_loss: 0.1330 - val_accuracy: 0.9605\n",
            "Epoch 5/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0929 - accuracy: 0.9710 - val_loss: 0.1473 - val_accuracy: 0.9567\n",
            "Epoch 6/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0700 - accuracy: 0.9783 - val_loss: 0.1187 - val_accuracy: 0.9654\n",
            "Epoch 7/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0525 - accuracy: 0.9831 - val_loss: 0.1152 - val_accuracy: 0.9673\n",
            "Epoch 8/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0388 - accuracy: 0.9884 - val_loss: 0.1429 - val_accuracy: 0.9584\n",
            "Epoch 9/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0292 - accuracy: 0.9910 - val_loss: 0.1795 - val_accuracy: 0.9542\n",
            "Epoch 10/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0214 - accuracy: 0.9935 - val_loss: 0.1247 - val_accuracy: 0.9693\n",
            "Epoch 1/10\n",
            "375/375 [==============================] - 2s 4ms/step - loss: 0.2901 - accuracy: 0.9171 - val_loss: 0.1599 - val_accuracy: 0.9551\n",
            "Epoch 2/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.1230 - accuracy: 0.9641 - val_loss: 0.1020 - val_accuracy: 0.9701\n",
            "Epoch 3/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0815 - accuracy: 0.9771 - val_loss: 0.0942 - val_accuracy: 0.9718\n",
            "Epoch 4/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0590 - accuracy: 0.9827 - val_loss: 0.0847 - val_accuracy: 0.9743\n",
            "Epoch 5/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0444 - accuracy: 0.9869 - val_loss: 0.0806 - val_accuracy: 0.9767\n",
            "Epoch 6/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0337 - accuracy: 0.9904 - val_loss: 0.0831 - val_accuracy: 0.9761\n",
            "Epoch 7/10\n",
            "375/375 [==============================] - 2s 4ms/step - loss: 0.0261 - accuracy: 0.9924 - val_loss: 0.0888 - val_accuracy: 0.9758\n",
            "Epoch 8/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0189 - accuracy: 0.9949 - val_loss: 0.0822 - val_accuracy: 0.9768\n",
            "Epoch 9/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0147 - accuracy: 0.9963 - val_loss: 0.0744 - val_accuracy: 0.9793\n",
            "Epoch 10/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0108 - accuracy: 0.9971 - val_loss: 0.0765 - val_accuracy: 0.9792\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "val_acc_noise = history_noise.history[\"val_accuracy\"]\n",
        "val_acc_zeros = history_zeros.history[\"val_accuracy\"]\n",
        "epochs = range(1, 11)\n",
        "plt.plot(epochs, val_acc_noise, \"b-\",\n",
        "         label=\"Validation accuracy with noise channels\")\n",
        "plt.plot(epochs, val_acc_zeros, \"b--\",\n",
        "         label=\"Validation accuracy with zeros channels\")\n",
        "plt.title(\"Effect of noise channels on validation accuracy\")\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.legend()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 490
        },
        "id": "dtIuKxZWXIyU",
        "outputId": "d3b19ce3-53f1-4b18-8135-7bb0607d98b7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x78980cf5a650>"
            ]
          },
          "metadata": {},
          "execution_count": 3
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "(train_images, train_labels), _ = mnist.load_data()\n",
        "train_images = train_images.reshape((60000, 28 * 28))\n",
        "train_images = train_images.astype(\"float32\") / 255\n",
        "\n",
        "random_train_labels = train_labels[:]\n",
        "np.random.shuffle(random_train_labels)\n",
        "\n",
        "model = keras.Sequential([\n",
        "    layers.Dense(512, activation=\"relu\"),\n",
        "    layers.Dense(10, activation=\"softmax\")\n",
        "])\n",
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"sparse_categorical_crossentropy\",\n",
        "              metrics=[\"accuracy\"])\n",
        "model.fit(train_images, random_train_labels,\n",
        "          epochs=100,\n",
        "          batch_size=128,\n",
        "          validation_split=0.2)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4C7Y4ZCbX4Uy",
        "outputId": "3461cff9-4634-4835-bd96-69db6d09f59a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "375/375 [==============================] - 2s 4ms/step - loss: 2.3152 - accuracy: 0.1026 - val_loss: 2.3053 - val_accuracy: 0.1058\n",
            "Epoch 2/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 2.2996 - accuracy: 0.1137 - val_loss: 2.3109 - val_accuracy: 0.1115\n",
            "Epoch 3/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 2.2910 - accuracy: 0.1265 - val_loss: 2.3166 - val_accuracy: 0.1003\n",
            "Epoch 4/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 2.2789 - accuracy: 0.1394 - val_loss: 2.3232 - val_accuracy: 0.1013\n",
            "Epoch 5/100\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 2.2642 - accuracy: 0.1510 - val_loss: 2.3401 - val_accuracy: 0.0991\n",
            "Epoch 6/100\n",
            "375/375 [==============================] - 2s 4ms/step - loss: 2.2431 - accuracy: 0.1664 - val_loss: 2.3499 - val_accuracy: 0.1028\n",
            "Epoch 7/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 2.2205 - accuracy: 0.1787 - val_loss: 2.3677 - val_accuracy: 0.0954\n",
            "Epoch 8/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 2.1945 - accuracy: 0.1965 - val_loss: 2.3901 - val_accuracy: 0.0946\n",
            "Epoch 9/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 2.1645 - accuracy: 0.2128 - val_loss: 2.3992 - val_accuracy: 0.0960\n",
            "Epoch 10/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 2.1328 - accuracy: 0.2286 - val_loss: 2.4315 - val_accuracy: 0.0967\n",
            "Epoch 11/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 2.0989 - accuracy: 0.2450 - val_loss: 2.4462 - val_accuracy: 0.0998\n",
            "Epoch 12/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 2.0625 - accuracy: 0.2615 - val_loss: 2.4839 - val_accuracy: 0.1033\n",
            "Epoch 13/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 2.0254 - accuracy: 0.2792 - val_loss: 2.5160 - val_accuracy: 0.0987\n",
            "Epoch 14/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.9871 - accuracy: 0.2959 - val_loss: 2.5396 - val_accuracy: 0.0999\n",
            "Epoch 15/100\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 1.9504 - accuracy: 0.3130 - val_loss: 2.5660 - val_accuracy: 0.1016\n",
            "Epoch 16/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.9120 - accuracy: 0.3266 - val_loss: 2.6115 - val_accuracy: 0.1007\n",
            "Epoch 17/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.8731 - accuracy: 0.3426 - val_loss: 2.6529 - val_accuracy: 0.0993\n",
            "Epoch 18/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.8344 - accuracy: 0.3569 - val_loss: 2.6719 - val_accuracy: 0.1002\n",
            "Epoch 19/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.7967 - accuracy: 0.3735 - val_loss: 2.7138 - val_accuracy: 0.1029\n",
            "Epoch 20/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.7578 - accuracy: 0.3895 - val_loss: 2.7503 - val_accuracy: 0.0951\n",
            "Epoch 21/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.7194 - accuracy: 0.4041 - val_loss: 2.8154 - val_accuracy: 0.1000\n",
            "Epoch 22/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.6845 - accuracy: 0.4175 - val_loss: 2.8463 - val_accuracy: 0.0987\n",
            "Epoch 23/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.6480 - accuracy: 0.4324 - val_loss: 2.8730 - val_accuracy: 0.1025\n",
            "Epoch 24/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.6131 - accuracy: 0.4443 - val_loss: 2.9269 - val_accuracy: 0.0984\n",
            "Epoch 25/100\n",
            "375/375 [==============================] - 2s 4ms/step - loss: 1.5791 - accuracy: 0.4588 - val_loss: 2.9710 - val_accuracy: 0.0963\n",
            "Epoch 26/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.5442 - accuracy: 0.4717 - val_loss: 3.0254 - val_accuracy: 0.1024\n",
            "Epoch 27/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.5114 - accuracy: 0.4845 - val_loss: 3.0700 - val_accuracy: 0.0998\n",
            "Epoch 28/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.4774 - accuracy: 0.4941 - val_loss: 3.1023 - val_accuracy: 0.0988\n",
            "Epoch 29/100\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 1.4477 - accuracy: 0.5044 - val_loss: 3.1566 - val_accuracy: 0.1027\n",
            "Epoch 30/100\n",
            "375/375 [==============================] - 2s 6ms/step - loss: 1.4157 - accuracy: 0.5194 - val_loss: 3.2208 - val_accuracy: 0.0961\n",
            "Epoch 31/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.3849 - accuracy: 0.5295 - val_loss: 3.2682 - val_accuracy: 0.0989\n",
            "Epoch 32/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.3553 - accuracy: 0.5394 - val_loss: 3.3164 - val_accuracy: 0.1013\n",
            "Epoch 33/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.3270 - accuracy: 0.5513 - val_loss: 3.3512 - val_accuracy: 0.1049\n",
            "Epoch 34/100\n",
            "375/375 [==============================] - 2s 4ms/step - loss: 1.2983 - accuracy: 0.5598 - val_loss: 3.4030 - val_accuracy: 0.1047\n",
            "Epoch 35/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.2701 - accuracy: 0.5704 - val_loss: 3.4754 - val_accuracy: 0.0996\n",
            "Epoch 36/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.2450 - accuracy: 0.5814 - val_loss: 3.5376 - val_accuracy: 0.1014\n",
            "Epoch 37/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.2190 - accuracy: 0.5913 - val_loss: 3.5953 - val_accuracy: 0.1000\n",
            "Epoch 38/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.1933 - accuracy: 0.5982 - val_loss: 3.6646 - val_accuracy: 0.0980\n",
            "Epoch 39/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.1672 - accuracy: 0.6077 - val_loss: 3.6974 - val_accuracy: 0.0986\n",
            "Epoch 40/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.1441 - accuracy: 0.6177 - val_loss: 3.7938 - val_accuracy: 0.1014\n",
            "Epoch 41/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.1187 - accuracy: 0.6254 - val_loss: 3.8093 - val_accuracy: 0.0993\n",
            "Epoch 42/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.0972 - accuracy: 0.6324 - val_loss: 3.8628 - val_accuracy: 0.1008\n",
            "Epoch 43/100\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 1.0738 - accuracy: 0.6424 - val_loss: 3.9449 - val_accuracy: 0.0968\n",
            "Epoch 44/100\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 1.0526 - accuracy: 0.6512 - val_loss: 3.9927 - val_accuracy: 0.0991\n",
            "Epoch 45/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.0287 - accuracy: 0.6583 - val_loss: 4.0598 - val_accuracy: 0.1002\n",
            "Epoch 46/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 1.0072 - accuracy: 0.6654 - val_loss: 4.1245 - val_accuracy: 0.1035\n",
            "Epoch 47/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.9892 - accuracy: 0.6716 - val_loss: 4.2011 - val_accuracy: 0.0978\n",
            "Epoch 48/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.9654 - accuracy: 0.6807 - val_loss: 4.2369 - val_accuracy: 0.0994\n",
            "Epoch 49/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.9478 - accuracy: 0.6855 - val_loss: 4.3095 - val_accuracy: 0.1000\n",
            "Epoch 50/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.9280 - accuracy: 0.6946 - val_loss: 4.3979 - val_accuracy: 0.0988\n",
            "Epoch 51/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.9090 - accuracy: 0.6985 - val_loss: 4.4335 - val_accuracy: 0.1036\n",
            "Epoch 52/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.8904 - accuracy: 0.7076 - val_loss: 4.4981 - val_accuracy: 0.1018\n",
            "Epoch 53/100\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 0.8705 - accuracy: 0.7134 - val_loss: 4.5602 - val_accuracy: 0.1002\n",
            "Epoch 54/100\n",
            "375/375 [==============================] - 2s 4ms/step - loss: 0.8535 - accuracy: 0.7208 - val_loss: 4.6411 - val_accuracy: 0.1003\n",
            "Epoch 55/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.8380 - accuracy: 0.7251 - val_loss: 4.6807 - val_accuracy: 0.0993\n",
            "Epoch 56/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.7717 - accuracy: 0.7469 - val_loss: 4.9465 - val_accuracy: 0.1002\n",
            "Epoch 60/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.7552 - accuracy: 0.7538 - val_loss: 5.0043 - val_accuracy: 0.1002\n",
            "Epoch 61/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.7424 - accuracy: 0.7584 - val_loss: 5.0918 - val_accuracy: 0.0986\n",
            "Epoch 62/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.7270 - accuracy: 0.7652 - val_loss: 5.1587 - val_accuracy: 0.1032\n",
            "Epoch 63/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.7105 - accuracy: 0.7709 - val_loss: 5.2760 - val_accuracy: 0.0998\n",
            "Epoch 64/100\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 0.6998 - accuracy: 0.7725 - val_loss: 5.3350 - val_accuracy: 0.0991\n",
            "Epoch 65/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.6816 - accuracy: 0.7786 - val_loss: 5.4249 - val_accuracy: 0.0983\n",
            "Epoch 66/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.6697 - accuracy: 0.7840 - val_loss: 5.4610 - val_accuracy: 0.1010\n",
            "Epoch 67/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.6579 - accuracy: 0.7884 - val_loss: 5.5396 - val_accuracy: 0.1009\n",
            "Epoch 68/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.6478 - accuracy: 0.7909 - val_loss: 5.5822 - val_accuracy: 0.1035\n",
            "Epoch 69/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.6316 - accuracy: 0.7965 - val_loss: 5.6077 - val_accuracy: 0.0984\n",
            "Epoch 70/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.6185 - accuracy: 0.8002 - val_loss: 5.7277 - val_accuracy: 0.1022\n",
            "Epoch 71/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.6075 - accuracy: 0.8036 - val_loss: 5.7946 - val_accuracy: 0.1022\n",
            "Epoch 72/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.5961 - accuracy: 0.8083 - val_loss: 5.8635 - val_accuracy: 0.1020\n",
            "Epoch 73/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.5824 - accuracy: 0.8126 - val_loss: 5.9243 - val_accuracy: 0.1048\n",
            "Epoch 74/100\n",
            "375/375 [==============================] - 2s 4ms/step - loss: 0.5744 - accuracy: 0.8150 - val_loss: 6.0136 - val_accuracy: 0.1014\n",
            "Epoch 75/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.5626 - accuracy: 0.8212 - val_loss: 6.0398 - val_accuracy: 0.1017\n",
            "Epoch 76/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.5508 - accuracy: 0.8250 - val_loss: 6.1716 - val_accuracy: 0.1012\n",
            "Epoch 77/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.5392 - accuracy: 0.8285 - val_loss: 6.2367 - val_accuracy: 0.0992\n",
            "Epoch 78/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.5291 - accuracy: 0.8298 - val_loss: 6.3005 - val_accuracy: 0.1018\n",
            "Epoch 79/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.5245 - accuracy: 0.8336 - val_loss: 6.3880 - val_accuracy: 0.1022\n",
            "Epoch 80/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.5094 - accuracy: 0.8383 - val_loss: 6.4711 - val_accuracy: 0.1012\n",
            "Epoch 81/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.5016 - accuracy: 0.8403 - val_loss: 6.5454 - val_accuracy: 0.1018\n",
            "Epoch 82/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.4916 - accuracy: 0.8449 - val_loss: 6.5921 - val_accuracy: 0.1016\n",
            "Epoch 83/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.4821 - accuracy: 0.8455 - val_loss: 6.6692 - val_accuracy: 0.1006\n",
            "Epoch 84/100\n",
            "375/375 [==============================] - 2s 4ms/step - loss: 0.4739 - accuracy: 0.8507 - val_loss: 6.7310 - val_accuracy: 0.1015\n",
            "Epoch 85/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.4648 - accuracy: 0.8526 - val_loss: 6.8060 - val_accuracy: 0.1022\n",
            "Epoch 86/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.4558 - accuracy: 0.8575 - val_loss: 6.8730 - val_accuracy: 0.1018\n",
            "Epoch 87/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.4462 - accuracy: 0.8593 - val_loss: 7.0519 - val_accuracy: 0.1021\n",
            "Epoch 88/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.4365 - accuracy: 0.8622 - val_loss: 7.0728 - val_accuracy: 0.1005\n",
            "Epoch 89/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.4324 - accuracy: 0.8640 - val_loss: 7.0757 - val_accuracy: 0.1038\n",
            "Epoch 90/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.4224 - accuracy: 0.8668 - val_loss: 7.1941 - val_accuracy: 0.1038\n",
            "Epoch 91/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.4165 - accuracy: 0.8701 - val_loss: 7.2420 - val_accuracy: 0.1031\n",
            "Epoch 92/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.4052 - accuracy: 0.8745 - val_loss: 7.3725 - val_accuracy: 0.1023\n",
            "Epoch 93/100\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 0.3995 - accuracy: 0.8758 - val_loss: 7.3557 - val_accuracy: 0.1029\n",
            "Epoch 94/100\n",
            "375/375 [==============================] - 2s 4ms/step - loss: 0.3935 - accuracy: 0.8769 - val_loss: 7.4539 - val_accuracy: 0.1041\n",
            "Epoch 95/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.3861 - accuracy: 0.8795 - val_loss: 7.5518 - val_accuracy: 0.1035\n",
            "Epoch 96/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.3768 - accuracy: 0.8831 - val_loss: 7.5684 - val_accuracy: 0.1005\n",
            "Epoch 97/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.3724 - accuracy: 0.8836 - val_loss: 7.7463 - val_accuracy: 0.1039\n",
            "Epoch 98/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.3652 - accuracy: 0.8884 - val_loss: 7.7741 - val_accuracy: 0.1010\n",
            "Epoch 99/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.3578 - accuracy: 0.8896 - val_loss: 7.8898 - val_accuracy: 0.1021\n",
            "Epoch 100/100\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.3511 - accuracy: 0.8913 - val_loss: 7.8734 - val_accuracy: 0.1052\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.src.callbacks.History at 0x7897b1a64cd0>"
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "(train_images, train_labels), _ = mnist.load_data()\n",
        "train_images = train_images.reshape((60000, 28 * 28))\n",
        "train_images = train_images.astype(\"float32\") / 255\n",
        "\n",
        "model = keras.Sequential([\n",
        "    layers.Dense(512, activation=\"relu\"),\n",
        "    layers.Dense(10, activation=\"softmax\")\n",
        "])\n",
        "model.compile(optimizer=keras.optimizers.RMSprop(1.),\n",
        "              loss=\"sparse_categorical_crossentropy\",\n",
        "              metrics=[\"accuracy\"])\n",
        "model.fit(train_images, train_labels,\n",
        "          epochs=10,\n",
        "          batch_size=128,\n",
        "          validation_split=0.2)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dTtF9nphZ6JT",
        "outputId": "590afc7c-8553-4e38-d99b-2eb2123028de"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "375/375 [==============================] - 3s 5ms/step - loss: 828.1812 - accuracy: 0.3001 - val_loss: 2.2160 - val_accuracy: 0.1865\n",
            "Epoch 2/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 3.1598 - accuracy: 0.1858 - val_loss: 2.0929 - val_accuracy: 0.1910\n",
            "Epoch 3/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 2.6904 - accuracy: 0.1611 - val_loss: 2.3525 - val_accuracy: 0.1631\n",
            "Epoch 4/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 2.4511 - accuracy: 0.1549 - val_loss: 2.2915 - val_accuracy: 0.1543\n",
            "Epoch 5/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 2.4735 - accuracy: 0.1907 - val_loss: 2.2177 - val_accuracy: 0.1810\n",
            "Epoch 6/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 2.4619 - accuracy: 0.1870 - val_loss: 2.1838 - val_accuracy: 0.1866\n",
            "Epoch 7/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 2.2848 - accuracy: 0.1992 - val_loss: 2.1233 - val_accuracy: 0.1937\n",
            "Epoch 8/10\n",
            "375/375 [==============================] - 2s 5ms/step - loss: 2.3333 - accuracy: 0.2106 - val_loss: 2.0473 - val_accuracy: 0.2387\n",
            "Epoch 9/10\n",
            "375/375 [==============================] - 2s 4ms/step - loss: 2.3394 - accuracy: 0.2126 - val_loss: 2.2500 - val_accuracy: 0.2065\n",
            "Epoch 10/10\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 2.4071 - accuracy: 0.2155 - val_loss: 2.1478 - val_accuracy: 0.2186\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.src.callbacks.History at 0x7898260af2e0>"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model = keras.Sequential([\n",
        "    layers.Dense(512, activation=\"relu\"),\n",
        "    layers.Dense(10, activation=\"softmax\")\n",
        "])\n",
        "model.compile(optimizer=keras.optimizers.RMSprop(1e-2),\n",
        "              loss=\"sparse_categorical_crossentropy\",\n",
        "              metrics=[\"accuracy\"])\n",
        "model.fit(train_images, train_labels,\n",
        "          epochs=10,\n",
        "          batch_size=128,\n",
        "          validation_split=0.2)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0Y9fYhniaAIh",
        "outputId": "561d7f54-a77b-4722-ba59-067dc47ecbb7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "375/375 [==============================] - 2s 4ms/step - loss: 0.3428 - accuracy: 0.9135 - val_loss: 0.1508 - val_accuracy: 0.9548\n",
            "Epoch 2/10\n",
            "375/375 [==============================] - 2s 4ms/step - loss: 0.1256 - accuracy: 0.9646 - val_loss: 0.1466 - val_accuracy: 0.9656\n",
            "Epoch 3/10\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 0.0973 - accuracy: 0.9742 - val_loss: 0.1608 - val_accuracy: 0.9652\n",
            "Epoch 4/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0780 - accuracy: 0.9791 - val_loss: 0.1631 - val_accuracy: 0.9678\n",
            "Epoch 5/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0697 - accuracy: 0.9827 - val_loss: 0.2001 - val_accuracy: 0.9672\n",
            "Epoch 6/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0610 - accuracy: 0.9851 - val_loss: 0.1892 - val_accuracy: 0.9720\n",
            "Epoch 7/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0572 - accuracy: 0.9867 - val_loss: 0.1915 - val_accuracy: 0.9736\n",
            "Epoch 8/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0480 - accuracy: 0.9890 - val_loss: 0.1953 - val_accuracy: 0.9718\n",
            "Epoch 9/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0422 - accuracy: 0.9909 - val_loss: 0.2048 - val_accuracy: 0.9746\n",
            "Epoch 10/10\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0390 - accuracy: 0.9913 - val_loss: 0.2376 - val_accuracy: 0.9718\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.src.callbacks.History at 0x789826127760>"
            ]
          },
          "metadata": {},
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "KrJYRNdROWan",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c6dc4321-206e-4ddc-ff60-801a50be0085"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "375/375 [==============================] - 2s 3ms/step - loss: 0.6761 - accuracy: 0.8331 - val_loss: 0.3605 - val_accuracy: 0.9047\n",
            "Epoch 2/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.3539 - accuracy: 0.9021 - val_loss: 0.3082 - val_accuracy: 0.9141\n",
            "Epoch 3/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.3181 - accuracy: 0.9117 - val_loss: 0.2916 - val_accuracy: 0.9207\n",
            "Epoch 4/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.3020 - accuracy: 0.9159 - val_loss: 0.2833 - val_accuracy: 0.9212\n",
            "Epoch 5/20\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 0.2921 - accuracy: 0.9183 - val_loss: 0.2766 - val_accuracy: 0.9251\n",
            "Epoch 6/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2856 - accuracy: 0.9207 - val_loss: 0.2761 - val_accuracy: 0.9242\n",
            "Epoch 7/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2804 - accuracy: 0.9217 - val_loss: 0.2720 - val_accuracy: 0.9268\n",
            "Epoch 8/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2766 - accuracy: 0.9232 - val_loss: 0.2703 - val_accuracy: 0.9273\n",
            "Epoch 9/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2735 - accuracy: 0.9240 - val_loss: 0.2693 - val_accuracy: 0.9275\n",
            "Epoch 10/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2709 - accuracy: 0.9247 - val_loss: 0.2677 - val_accuracy: 0.9276\n",
            "Epoch 11/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2685 - accuracy: 0.9256 - val_loss: 0.2664 - val_accuracy: 0.9283\n",
            "Epoch 12/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2665 - accuracy: 0.9265 - val_loss: 0.2639 - val_accuracy: 0.9284\n",
            "Epoch 13/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2651 - accuracy: 0.9270 - val_loss: 0.2634 - val_accuracy: 0.9291\n",
            "Epoch 14/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2632 - accuracy: 0.9276 - val_loss: 0.2649 - val_accuracy: 0.9290\n",
            "Epoch 15/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2619 - accuracy: 0.9276 - val_loss: 0.2645 - val_accuracy: 0.9295\n",
            "Epoch 16/20\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 0.2603 - accuracy: 0.9274 - val_loss: 0.2624 - val_accuracy: 0.9294\n",
            "Epoch 17/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2592 - accuracy: 0.9280 - val_loss: 0.2645 - val_accuracy: 0.9298\n",
            "Epoch 18/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2583 - accuracy: 0.9295 - val_loss: 0.2630 - val_accuracy: 0.9290\n",
            "Epoch 19/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2573 - accuracy: 0.9287 - val_loss: 0.2616 - val_accuracy: 0.9299\n",
            "Epoch 20/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.2564 - accuracy: 0.9296 - val_loss: 0.2619 - val_accuracy: 0.9300\n"
          ]
        }
      ],
      "source": [
        "model = keras.Sequential([layers.Dense(10, activation=\"softmax\")])\n",
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"sparse_categorical_crossentropy\",\n",
        "              metrics=[\"accuracy\"])\n",
        "history_small_model = model.fit(\n",
        "    train_images, train_labels,\n",
        "    epochs=20,\n",
        "    batch_size=128,\n",
        "    validation_split=0.2)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "PN4-Zi5BOWan",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 490
        },
        "outputId": "c217a7aa-daa8-4552-cac5-b499270e9979"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x789824d5eef0>"
            ]
          },
          "metadata": {},
          "execution_count": 9
        },
        {
          "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",
        "val_loss = history_small_model.history[\"val_loss\"]\n",
        "epochs = range(1, 21)\n",
        "plt.plot(epochs, val_loss, \"b--\",\n",
        "         label=\"Validation loss\")\n",
        "plt.title(\"Effect of insufficient model capacity on validation loss\")\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Loss\")\n",
        "plt.legend()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "kK6FMT1FOWan",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a816fd4b-2b2f-467f-aa09-59ba66baa165"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "375/375 [==============================] - 2s 4ms/step - loss: 0.3639 - accuracy: 0.8979 - val_loss: 0.1781 - val_accuracy: 0.9485\n",
            "Epoch 2/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.1591 - accuracy: 0.9530 - val_loss: 0.1252 - val_accuracy: 0.9642\n",
            "Epoch 3/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.1121 - accuracy: 0.9669 - val_loss: 0.1123 - val_accuracy: 0.9665\n",
            "Epoch 4/20\n",
            "375/375 [==============================] - 2s 5ms/step - loss: 0.0852 - accuracy: 0.9750 - val_loss: 0.0988 - val_accuracy: 0.9712\n",
            "Epoch 5/20\n",
            "375/375 [==============================] - 2s 5ms/step - loss: 0.0685 - accuracy: 0.9796 - val_loss: 0.1022 - val_accuracy: 0.9705\n",
            "Epoch 6/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0564 - accuracy: 0.9828 - val_loss: 0.0912 - val_accuracy: 0.9731\n",
            "Epoch 7/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0454 - accuracy: 0.9862 - val_loss: 0.0952 - val_accuracy: 0.9749\n",
            "Epoch 8/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0398 - accuracy: 0.9885 - val_loss: 0.0899 - val_accuracy: 0.9737\n",
            "Epoch 9/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0330 - accuracy: 0.9899 - val_loss: 0.0943 - val_accuracy: 0.9758\n",
            "Epoch 10/20\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 0.0271 - accuracy: 0.9915 - val_loss: 0.0908 - val_accuracy: 0.9765\n",
            "Epoch 11/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0233 - accuracy: 0.9925 - val_loss: 0.1031 - val_accuracy: 0.9734\n",
            "Epoch 12/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0198 - accuracy: 0.9941 - val_loss: 0.1026 - val_accuracy: 0.9759\n",
            "Epoch 13/20\n",
            "375/375 [==============================] - 2s 4ms/step - loss: 0.0167 - accuracy: 0.9950 - val_loss: 0.1006 - val_accuracy: 0.9761\n",
            "Epoch 14/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0144 - accuracy: 0.9959 - val_loss: 0.1171 - val_accuracy: 0.9735\n",
            "Epoch 15/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0128 - accuracy: 0.9964 - val_loss: 0.1148 - val_accuracy: 0.9744\n",
            "Epoch 16/20\n",
            "375/375 [==============================] - 1s 4ms/step - loss: 0.0107 - accuracy: 0.9966 - val_loss: 0.1194 - val_accuracy: 0.9747\n",
            "Epoch 17/20\n",
            "375/375 [==============================] - 2s 5ms/step - loss: 0.0089 - accuracy: 0.9975 - val_loss: 0.1074 - val_accuracy: 0.9766\n",
            "Epoch 18/20\n",
            "375/375 [==============================] - 3s 7ms/step - loss: 0.0076 - accuracy: 0.9977 - val_loss: 0.1175 - val_accuracy: 0.9772\n",
            "Epoch 19/20\n",
            "375/375 [==============================] - 1s 3ms/step - loss: 0.0068 - accuracy: 0.9978 - val_loss: 0.1216 - val_accuracy: 0.9754\n",
            "Epoch 20/20\n",
            "375/375 [==============================] - 2s 4ms/step - loss: 0.0059 - accuracy: 0.9985 - val_loss: 0.1221 - val_accuracy: 0.9758\n"
          ]
        }
      ],
      "source": [
        "model = keras.Sequential([\n",
        "    layers.Dense(96, activation=\"relu\"),\n",
        "    layers.Dense(96, activation=\"relu\"),\n",
        "    layers.Dense(10, activation=\"softmax\"),\n",
        "])\n",
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"sparse_categorical_crossentropy\",\n",
        "              metrics=[\"accuracy\"])\n",
        "history_large_model = model.fit(\n",
        "    train_images, train_labels,\n",
        "    epochs=20,\n",
        "    batch_size=128,\n",
        "    validation_split=0.2)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Improving generalization"
      ],
      "metadata": {
        "id": "mlp1DRD5b5J_"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "YZlmndrCOWat",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a5970636-2e8b-43db-e952-1b85f63eaaf5"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/imdb.npz\n",
            "17464789/17464789 [==============================] - 0s 0us/step\n",
            "Epoch 1/20\n",
            "30/30 [==============================] - 4s 80ms/step - loss: 0.5164 - accuracy: 0.7801 - val_loss: 0.3992 - val_accuracy: 0.8520\n",
            "Epoch 2/20\n",
            "30/30 [==============================] - 1s 44ms/step - loss: 0.3168 - accuracy: 0.8943 - val_loss: 0.3058 - val_accuracy: 0.8865\n",
            "Epoch 3/20\n",
            "30/30 [==============================] - 2s 54ms/step - loss: 0.2391 - accuracy: 0.9205 - val_loss: 0.2799 - val_accuracy: 0.8907\n",
            "Epoch 4/20\n",
            "30/30 [==============================] - 1s 39ms/step - loss: 0.1920 - accuracy: 0.9375 - val_loss: 0.2778 - val_accuracy: 0.8899\n",
            "Epoch 5/20\n",
            "30/30 [==============================] - 1s 23ms/step - loss: 0.1617 - accuracy: 0.9464 - val_loss: 0.2772 - val_accuracy: 0.8900\n",
            "Epoch 6/20\n",
            "30/30 [==============================] - 1s 25ms/step - loss: 0.1404 - accuracy: 0.9539 - val_loss: 0.2854 - val_accuracy: 0.8877\n",
            "Epoch 7/20\n",
            "30/30 [==============================] - 1s 22ms/step - loss: 0.1187 - accuracy: 0.9628 - val_loss: 0.3257 - val_accuracy: 0.8781\n",
            "Epoch 8/20\n",
            "30/30 [==============================] - 1s 22ms/step - loss: 0.1035 - accuracy: 0.9693 - val_loss: 0.3346 - val_accuracy: 0.8769\n",
            "Epoch 9/20\n",
            "30/30 [==============================] - 1s 22ms/step - loss: 0.0911 - accuracy: 0.9729 - val_loss: 0.3258 - val_accuracy: 0.8813\n",
            "Epoch 10/20\n",
            "30/30 [==============================] - 1s 23ms/step - loss: 0.0793 - accuracy: 0.9777 - val_loss: 0.3442 - val_accuracy: 0.8803\n",
            "Epoch 11/20\n",
            "30/30 [==============================] - 1s 21ms/step - loss: 0.0680 - accuracy: 0.9825 - val_loss: 0.3646 - val_accuracy: 0.8779\n",
            "Epoch 12/20\n",
            "30/30 [==============================] - 1s 24ms/step - loss: 0.0610 - accuracy: 0.9844 - val_loss: 0.3868 - val_accuracy: 0.8770\n",
            "Epoch 13/20\n",
            "30/30 [==============================] - 1s 22ms/step - loss: 0.0532 - accuracy: 0.9862 - val_loss: 0.4071 - val_accuracy: 0.8751\n",
            "Epoch 14/20\n",
            "30/30 [==============================] - 1s 24ms/step - loss: 0.0442 - accuracy: 0.9903 - val_loss: 0.4277 - val_accuracy: 0.8740\n",
            "Epoch 15/20\n",
            "30/30 [==============================] - 1s 24ms/step - loss: 0.0395 - accuracy: 0.9910 - val_loss: 0.4532 - val_accuracy: 0.8732\n",
            "Epoch 16/20\n",
            "30/30 [==============================] - 1s 21ms/step - loss: 0.0371 - accuracy: 0.9921 - val_loss: 0.4750 - val_accuracy: 0.8725\n",
            "Epoch 17/20\n",
            "30/30 [==============================] - 1s 22ms/step - loss: 0.0296 - accuracy: 0.9947 - val_loss: 0.4985 - val_accuracy: 0.8696\n",
            "Epoch 18/20\n",
            "30/30 [==============================] - 1s 21ms/step - loss: 0.0240 - accuracy: 0.9961 - val_loss: 0.5181 - val_accuracy: 0.8707\n",
            "Epoch 19/20\n",
            "30/30 [==============================] - 1s 26ms/step - loss: 0.0231 - accuracy: 0.9958 - val_loss: 0.5443 - val_accuracy: 0.8700\n",
            "Epoch 20/20\n",
            "30/30 [==============================] - 1s 28ms/step - loss: 0.0209 - accuracy: 0.9960 - val_loss: 0.5704 - val_accuracy: 0.8680\n"
          ]
        }
      ],
      "source": [
        "from tensorflow.keras.datasets import imdb\n",
        "(train_data, train_labels), _ = imdb.load_data(num_words=10000)\n",
        "\n",
        "def vectorize_sequences(sequences, dimension=10000):\n",
        "    results = np.zeros((len(sequences), dimension))\n",
        "    for i, sequence in enumerate(sequences):\n",
        "        results[i, sequence] = 1.\n",
        "    return results\n",
        "train_data = vectorize_sequences(train_data)\n",
        "\n",
        "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",
        "history_original = model.fit(train_data, train_labels,\n",
        "                             epochs=20, batch_size=512, validation_split=0.4)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GnXm77hXOWat"
      },
      "source": [
        "**Version of the model with lower capacity**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "kI4KzYG_OWat",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e7412e51-1c1a-4c2d-a028-a9d77d0e1595"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "30/30 [==============================] - 3s 76ms/step - loss: 0.6521 - accuracy: 0.6493 - val_loss: 0.6122 - val_accuracy: 0.7033\n",
            "Epoch 2/20\n",
            "30/30 [==============================] - 1s 29ms/step - loss: 0.5790 - accuracy: 0.7654 - val_loss: 0.5453 - val_accuracy: 0.8433\n",
            "Epoch 3/20\n",
            "30/30 [==============================] - 1s 22ms/step - loss: 0.4976 - accuracy: 0.8558 - val_loss: 0.4737 - val_accuracy: 0.8635\n",
            "Epoch 4/20\n",
            "30/30 [==============================] - 1s 23ms/step - loss: 0.4222 - accuracy: 0.8911 - val_loss: 0.4140 - val_accuracy: 0.8785\n",
            "Epoch 5/20\n",
            "30/30 [==============================] - 1s 21ms/step - loss: 0.3579 - accuracy: 0.9078 - val_loss: 0.3671 - val_accuracy: 0.8844\n",
            "Epoch 6/20\n",
            "30/30 [==============================] - 1s 24ms/step - loss: 0.3075 - accuracy: 0.9171 - val_loss: 0.3323 - val_accuracy: 0.8852\n",
            "Epoch 7/20\n",
            "30/30 [==============================] - 1s 22ms/step - loss: 0.2689 - accuracy: 0.9231 - val_loss: 0.3120 - val_accuracy: 0.8844\n",
            "Epoch 8/20\n",
            "30/30 [==============================] - 1s 24ms/step - loss: 0.2391 - accuracy: 0.9292 - val_loss: 0.2952 - val_accuracy: 0.8877\n",
            "Epoch 9/20\n",
            "30/30 [==============================] - 1s 24ms/step - loss: 0.2155 - accuracy: 0.9359 - val_loss: 0.2836 - val_accuracy: 0.8915\n",
            "Epoch 10/20\n",
            "30/30 [==============================] - 1s 24ms/step - loss: 0.1957 - accuracy: 0.9411 - val_loss: 0.2805 - val_accuracy: 0.8894\n",
            "Epoch 11/20\n",
            "30/30 [==============================] - 1s 22ms/step - loss: 0.1801 - accuracy: 0.9445 - val_loss: 0.2746 - val_accuracy: 0.8929\n",
            "Epoch 12/20\n",
            "30/30 [==============================] - 1s 21ms/step - loss: 0.1653 - accuracy: 0.9497 - val_loss: 0.2747 - val_accuracy: 0.8912\n",
            "Epoch 13/20\n",
            "30/30 [==============================] - 1s 25ms/step - loss: 0.1526 - accuracy: 0.9543 - val_loss: 0.2765 - val_accuracy: 0.8904\n",
            "Epoch 14/20\n",
            "30/30 [==============================] - 1s 21ms/step - loss: 0.1411 - accuracy: 0.9583 - val_loss: 0.2781 - val_accuracy: 0.8883\n",
            "Epoch 15/20\n",
            "30/30 [==============================] - 1s 24ms/step - loss: 0.1312 - accuracy: 0.9620 - val_loss: 0.2810 - val_accuracy: 0.8891\n",
            "Epoch 16/20\n",
            "30/30 [==============================] - 1s 22ms/step - loss: 0.1226 - accuracy: 0.9641 - val_loss: 0.2858 - val_accuracy: 0.8882\n",
            "Epoch 17/20\n",
            "30/30 [==============================] - 1s 38ms/step - loss: 0.1139 - accuracy: 0.9682 - val_loss: 0.2924 - val_accuracy: 0.8872\n",
            "Epoch 18/20\n",
            "30/30 [==============================] - 1s 32ms/step - loss: 0.1061 - accuracy: 0.9701 - val_loss: 0.2973 - val_accuracy: 0.8870\n",
            "Epoch 19/20\n",
            "30/30 [==============================] - 1s 24ms/step - loss: 0.0985 - accuracy: 0.9739 - val_loss: 0.3035 - val_accuracy: 0.8851\n",
            "Epoch 20/20\n",
            "30/30 [==============================] - 1s 20ms/step - loss: 0.0920 - accuracy: 0.9757 - val_loss: 0.3112 - val_accuracy: 0.8841\n"
          ]
        }
      ],
      "source": [
        "model = keras.Sequential([\n",
        "    layers.Dense(4, activation=\"relu\"),\n",
        "    layers.Dense(4, activation=\"relu\"),\n",
        "    layers.Dense(1, activation=\"sigmoid\")\n",
        "])\n",
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"binary_crossentropy\",\n",
        "              metrics=[\"accuracy\"])\n",
        "history_smaller_model = model.fit(\n",
        "    train_data, train_labels,\n",
        "    epochs=20, batch_size=512, validation_split=0.4)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BweNeMdUOWat"
      },
      "source": [
        "**Version of the model with higher capacity**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "15PZwcKUOWat",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5cc86365-7efe-4cdb-8988-9615fa64505e"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "30/30 [==============================] - 3s 86ms/step - loss: 0.5732 - accuracy: 0.7107 - val_loss: 0.3606 - val_accuracy: 0.8633\n",
            "Epoch 2/20\n",
            "30/30 [==============================] - 1s 31ms/step - loss: 0.3214 - accuracy: 0.8664 - val_loss: 0.3571 - val_accuracy: 0.8437\n",
            "Epoch 3/20\n",
            "30/30 [==============================] - 1s 28ms/step - loss: 0.2379 - accuracy: 0.9050 - val_loss: 0.2881 - val_accuracy: 0.8789\n",
            "Epoch 4/20\n",
            "30/30 [==============================] - 1s 26ms/step - loss: 0.1861 - accuracy: 0.9266 - val_loss: 0.2741 - val_accuracy: 0.8901\n",
            "Epoch 5/20\n",
            "30/30 [==============================] - 1s 25ms/step - loss: 0.1403 - accuracy: 0.9491 - val_loss: 0.3509 - val_accuracy: 0.8500\n",
            "Epoch 6/20\n",
            "30/30 [==============================] - 1s 26ms/step - loss: 0.0974 - accuracy: 0.9664 - val_loss: 0.3302 - val_accuracy: 0.8873\n",
            "Epoch 7/20\n",
            "30/30 [==============================] - 1s 27ms/step - loss: 0.0846 - accuracy: 0.9743 - val_loss: 0.3635 - val_accuracy: 0.8845\n",
            "Epoch 8/20\n",
            "30/30 [==============================] - 1s 28ms/step - loss: 0.0816 - accuracy: 0.9763 - val_loss: 0.3631 - val_accuracy: 0.8841\n",
            "Epoch 9/20\n",
            "30/30 [==============================] - 1s 25ms/step - loss: 0.0594 - accuracy: 0.9805 - val_loss: 0.3197 - val_accuracy: 0.8837\n",
            "Epoch 10/20\n",
            "30/30 [==============================] - 1s 25ms/step - loss: 0.0118 - accuracy: 0.9989 - val_loss: 0.4754 - val_accuracy: 0.8802\n",
            "Epoch 11/20\n",
            "30/30 [==============================] - 1s 27ms/step - loss: 0.0043 - accuracy: 0.9997 - val_loss: 0.5661 - val_accuracy: 0.8780\n",
            "Epoch 12/20\n",
            "30/30 [==============================] - 1s 25ms/step - loss: 0.0827 - accuracy: 0.9835 - val_loss: 0.4678 - val_accuracy: 0.8843\n",
            "Epoch 13/20\n",
            "30/30 [==============================] - 1s 26ms/step - loss: 0.0024 - accuracy: 0.9999 - val_loss: 0.5456 - val_accuracy: 0.8839\n",
            "Epoch 14/20\n",
            "30/30 [==============================] - 1s 25ms/step - loss: 0.0827 - accuracy: 0.9852 - val_loss: 0.4171 - val_accuracy: 0.8802\n",
            "Epoch 15/20\n",
            "30/30 [==============================] - 1s 40ms/step - loss: 0.0041 - accuracy: 1.0000 - val_loss: 0.4901 - val_accuracy: 0.8824\n",
            "Epoch 16/20\n",
            "30/30 [==============================] - 1s 33ms/step - loss: 0.0014 - accuracy: 1.0000 - val_loss: 0.5715 - val_accuracy: 0.8830\n",
            "Epoch 17/20\n",
            "30/30 [==============================] - 1s 28ms/step - loss: 7.5910e-04 - accuracy: 1.0000 - val_loss: 0.6286 - val_accuracy: 0.8836\n",
            "Epoch 18/20\n",
            "30/30 [==============================] - 1s 28ms/step - loss: 4.7790e-04 - accuracy: 1.0000 - val_loss: 0.6659 - val_accuracy: 0.8835\n",
            "Epoch 19/20\n",
            "30/30 [==============================] - 1s 26ms/step - loss: 3.3583e-04 - accuracy: 1.0000 - val_loss: 0.6959 - val_accuracy: 0.8828\n",
            "Epoch 20/20\n",
            "30/30 [==============================] - 1s 27ms/step - loss: 2.5053e-04 - accuracy: 1.0000 - val_loss: 0.7178 - val_accuracy: 0.8835\n"
          ]
        }
      ],
      "source": [
        "model = keras.Sequential([\n",
        "    layers.Dense(512, activation=\"relu\"),\n",
        "    layers.Dense(512, activation=\"relu\"),\n",
        "    layers.Dense(1, activation=\"sigmoid\")\n",
        "])\n",
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"binary_crossentropy\",\n",
        "              metrics=[\"accuracy\"])\n",
        "history_larger_model = model.fit(\n",
        "    train_data, train_labels,\n",
        "    epochs=20, batch_size=512, validation_split=0.4)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nl2q_d9zOWat"
      },
      "source": [
        "#### Adding weight regularization"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XtYTC8LGOWat"
      },
      "source": [
        "**Adding L2 weight regularization to the model**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "pUi0SbNKOWat",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a78d34ed-0663-454e-980d-ff02d7dcf6d7"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "30/30 [==============================] - 3s 67ms/step - loss: 0.5910 - accuracy: 0.7749 - val_loss: 0.4611 - val_accuracy: 0.8610\n",
            "Epoch 2/20\n",
            "30/30 [==============================] - 1s 22ms/step - loss: 0.3918 - accuracy: 0.8953 - val_loss: 0.3882 - val_accuracy: 0.8831\n",
            "Epoch 3/20\n",
            "30/30 [==============================] - 1s 22ms/step - loss: 0.3236 - accuracy: 0.9195 - val_loss: 0.3649 - val_accuracy: 0.8884\n",
            "Epoch 4/20\n",
            "30/30 [==============================] - 1s 22ms/step - loss: 0.2918 - accuracy: 0.9269 - val_loss: 0.3824 - val_accuracy: 0.8755\n",
            "Epoch 5/20\n",
            "30/30 [==============================] - 1s 25ms/step - loss: 0.2708 - accuracy: 0.9343 - val_loss: 0.3602 - val_accuracy: 0.8856\n",
            "Epoch 6/20\n",
            "30/30 [==============================] - 1s 29ms/step - loss: 0.2576 - accuracy: 0.9389 - val_loss: 0.3596 - val_accuracy: 0.8829\n",
            "Epoch 7/20\n",
            "30/30 [==============================] - 1s 41ms/step - loss: 0.2462 - accuracy: 0.9433 - val_loss: 0.3721 - val_accuracy: 0.8808\n",
            "Epoch 8/20\n",
            "30/30 [==============================] - 1s 26ms/step - loss: 0.2386 - accuracy: 0.9456 - val_loss: 0.3693 - val_accuracy: 0.8819\n",
            "Epoch 9/20\n",
            "30/30 [==============================] - 1s 24ms/step - loss: 0.2311 - accuracy: 0.9488 - val_loss: 0.3826 - val_accuracy: 0.8817\n",
            "Epoch 10/20\n",
            "30/30 [==============================] - 1s 24ms/step - loss: 0.2283 - accuracy: 0.9499 - val_loss: 0.3813 - val_accuracy: 0.8806\n",
            "Epoch 11/20\n",
            "30/30 [==============================] - 1s 23ms/step - loss: 0.2241 - accuracy: 0.9506 - val_loss: 0.3851 - val_accuracy: 0.8814\n",
            "Epoch 12/20\n",
            "30/30 [==============================] - 1s 30ms/step - loss: 0.2204 - accuracy: 0.9523 - val_loss: 0.4100 - val_accuracy: 0.8730\n",
            "Epoch 13/20\n",
            "30/30 [==============================] - 1s 25ms/step - loss: 0.2133 - accuracy: 0.9545 - val_loss: 0.4655 - val_accuracy: 0.8633\n",
            "Epoch 14/20\n",
            "30/30 [==============================] - 1s 24ms/step - loss: 0.2119 - accuracy: 0.9529 - val_loss: 0.4430 - val_accuracy: 0.8695\n",
            "Epoch 15/20\n",
            "30/30 [==============================] - 1s 22ms/step - loss: 0.2097 - accuracy: 0.9565 - val_loss: 0.4080 - val_accuracy: 0.8756\n",
            "Epoch 16/20\n",
            "30/30 [==============================] - 1s 22ms/step - loss: 0.2080 - accuracy: 0.9593 - val_loss: 0.4054 - val_accuracy: 0.8770\n",
            "Epoch 17/20\n",
            "30/30 [==============================] - 1s 23ms/step - loss: 0.2041 - accuracy: 0.9595 - val_loss: 0.4189 - val_accuracy: 0.8748\n",
            "Epoch 18/20\n",
            "30/30 [==============================] - 1s 21ms/step - loss: 0.2019 - accuracy: 0.9608 - val_loss: 0.4283 - val_accuracy: 0.8726\n",
            "Epoch 19/20\n",
            "30/30 [==============================] - 1s 23ms/step - loss: 0.2002 - accuracy: 0.9584 - val_loss: 0.4158 - val_accuracy: 0.8750\n",
            "Epoch 20/20\n",
            "30/30 [==============================] - 1s 21ms/step - loss: 0.2020 - accuracy: 0.9580 - val_loss: 0.4563 - val_accuracy: 0.8682\n"
          ]
        }
      ],
      "source": [
        "from tensorflow.keras import regularizers\n",
        "model = keras.Sequential([\n",
        "    layers.Dense(16,\n",
        "                 kernel_regularizer=regularizers.l2(0.002),\n",
        "                 activation=\"relu\"),\n",
        "    layers.Dense(16,\n",
        "                 kernel_regularizer=regularizers.l2(0.002),\n",
        "                 activation=\"relu\"),\n",
        "    layers.Dense(1, activation=\"sigmoid\")\n",
        "])\n",
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"binary_crossentropy\",\n",
        "              metrics=[\"accuracy\"])\n",
        "history_l2_reg = model.fit(\n",
        "    train_data, train_labels,\n",
        "    epochs=20, batch_size=512, validation_split=0.4)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2daccZuCOWau"
      },
      "source": [
        "**Different weight regularizers available in Keras**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "9cWORZCVOWau",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "61fb6bf8-2f40-4e1d-c207-a087fb4c095a"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.src.regularizers.L1L2 at 0x78981cd07a30>"
            ]
          },
          "metadata": {},
          "execution_count": 15
        }
      ],
      "source": [
        "from tensorflow.keras import regularizers\n",
        "regularizers.l1(0.001)\n",
        "regularizers.l1_l2(l1=0.001, l2=0.001)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nu2TKDMuOWau"
      },
      "source": [
        "#### Adding dropout"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "y4_PC3yLOWau"
      },
      "source": [
        "**Adding dropout to the IMDB model**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "m2BeW2yLOWau",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d5d1d2fe-68b0-414b-c4b3-91840859cc9e"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "30/30 [==============================] - 5s 67ms/step - loss: 0.6482 - accuracy: 0.6225 - val_loss: 0.5744 - val_accuracy: 0.8390\n",
            "Epoch 2/20\n",
            "30/30 [==============================] - 1s 25ms/step - loss: 0.5553 - accuracy: 0.7251 - val_loss: 0.4719 - val_accuracy: 0.8679\n",
            "Epoch 3/20\n",
            "30/30 [==============================] - 1s 22ms/step - loss: 0.4721 - accuracy: 0.7917 - val_loss: 0.3784 - val_accuracy: 0.8785\n",
            "Epoch 4/20\n",
            "30/30 [==============================] - 1s 23ms/step - loss: 0.4021 - accuracy: 0.8315 - val_loss: 0.3221 - val_accuracy: 0.8822\n",
            "Epoch 5/20\n",
            "30/30 [==============================] - 1s 23ms/step - loss: 0.3527 - accuracy: 0.8580 - val_loss: 0.2900 - val_accuracy: 0.8880\n",
            "Epoch 6/20\n",
            "30/30 [==============================] - 1s 25ms/step - loss: 0.3044 - accuracy: 0.8852 - val_loss: 0.2839 - val_accuracy: 0.8833\n",
            "Epoch 7/20\n",
            "30/30 [==============================] - 1s 22ms/step - loss: 0.2707 - accuracy: 0.9001 - val_loss: 0.2884 - val_accuracy: 0.8806\n",
            "Epoch 8/20\n",
            "30/30 [==============================] - 1s 24ms/step - loss: 0.2438 - accuracy: 0.9155 - val_loss: 0.2751 - val_accuracy: 0.8925\n",
            "Epoch 9/20\n",
            "30/30 [==============================] - 1s 30ms/step - loss: 0.2134 - accuracy: 0.9235 - val_loss: 0.2767 - val_accuracy: 0.8926\n",
            "Epoch 10/20\n",
            "30/30 [==============================] - 1s 42ms/step - loss: 0.1983 - accuracy: 0.9325 - val_loss: 0.3221 - val_accuracy: 0.8818\n",
            "Epoch 11/20\n",
            "30/30 [==============================] - 1s 25ms/step - loss: 0.1754 - accuracy: 0.9420 - val_loss: 0.3089 - val_accuracy: 0.8865\n",
            "Epoch 12/20\n",
            "30/30 [==============================] - 1s 22ms/step - loss: 0.1582 - accuracy: 0.9473 - val_loss: 0.3140 - val_accuracy: 0.8891\n",
            "Epoch 13/20\n",
            "30/30 [==============================] - 1s 24ms/step - loss: 0.1440 - accuracy: 0.9529 - val_loss: 0.3444 - val_accuracy: 0.8894\n",
            "Epoch 14/20\n",
            "30/30 [==============================] - 1s 22ms/step - loss: 0.1304 - accuracy: 0.9597 - val_loss: 0.3594 - val_accuracy: 0.8868\n",
            "Epoch 15/20\n",
            "30/30 [==============================] - 1s 22ms/step - loss: 0.1172 - accuracy: 0.9629 - val_loss: 0.4044 - val_accuracy: 0.8813\n",
            "Epoch 16/20\n",
            "30/30 [==============================] - 1s 25ms/step - loss: 0.1145 - accuracy: 0.9621 - val_loss: 0.3956 - val_accuracy: 0.8869\n",
            "Epoch 17/20\n",
            "30/30 [==============================] - 1s 23ms/step - loss: 0.1003 - accuracy: 0.9661 - val_loss: 0.4290 - val_accuracy: 0.8855\n",
            "Epoch 18/20\n",
            "30/30 [==============================] - 1s 22ms/step - loss: 0.0949 - accuracy: 0.9690 - val_loss: 0.4598 - val_accuracy: 0.8853\n",
            "Epoch 19/20\n",
            "30/30 [==============================] - 1s 23ms/step - loss: 0.0907 - accuracy: 0.9711 - val_loss: 0.4906 - val_accuracy: 0.8855\n",
            "Epoch 20/20\n",
            "30/30 [==============================] - 1s 22ms/step - loss: 0.0851 - accuracy: 0.9715 - val_loss: 0.5089 - val_accuracy: 0.8853\n"
          ]
        }
      ],
      "source": [
        "model = keras.Sequential([\n",
        "    layers.Dense(16, activation=\"relu\"),\n",
        "    layers.Dropout(0.5),\n",
        "    layers.Dense(16, activation=\"relu\"),\n",
        "    layers.Dropout(0.5),\n",
        "    layers.Dense(1, activation=\"sigmoid\")\n",
        "])\n",
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"binary_crossentropy\",\n",
        "              metrics=[\"accuracy\"])\n",
        "history_dropout = model.fit(\n",
        "    train_data, train_labels,\n",
        "    epochs=20, batch_size=512, validation_split=0.4)"
      ]
    }
  ]
}