KnowingFly/depression-detection-api

1{2  "cells": [3    {4      "cell_type": "code",5      "execution_count": null,6      "metadata": {7        "id": "Nm0samcYYItm"8      },9      "outputs": [],10      "source": [11        "# !pip install roboflow -q"12      ]13    },14    {15      "cell_type": "code",16      "execution_count": null,17      "metadata": {18        "id": "-IJMSzpIYWQh"19      },20      "outputs": [],21      "source": [22        "import tensorflow as tf\n",23        "import numpy as np\n",24        "import pathlib\n",25        "import matplotlib.pyplot as plt\n",26        "# from roboflow import Roboflow\n",27        "import zipfile\n",28        "import seaborn as sns\n",29        "from tensorflow.keras import layers, Model\n",30        "from sklearn.metrics import confusion_matrix\n",31        "from sklearn.utils import class_weight"32      ]33    },34    {35      "cell_type": "code",36      "execution_count": null,37      "metadata": {38        "colab": {39          "base_uri": "https://localhost:8080/"40        },41        "id": "csjG-tFKYXa7",42        "outputId": "8f0a4d1b-f806-441a-c389-9fd783f86868"43      },44      "outputs": [45        {46          "output_type": "stream",47          "name": "stdout",48          "text": [49            "Mounted at /content/drive\n"50          ]51        }52      ],53      "source": [54        "# 1. Konfigurasi dataset\n",55        "# -------------------------------------------------------------------\n",56        "# rf = Roboflow(api_key=\"XJvBFxKWavgQgkG2i7pF\")\n",57        "# project = rf.workspace(\"jeo\").project(\"calmscope-pvsog\")\n",58        "# version = project.version(1)\n",59        "# dataset = version.download(\"yolov12\")\n",60        "\n",61        "from google.colab import drive\n",62        "drive.mount('/content/drive')\n",63        "\n",64        "zip_path = '/content/drive/MyDrive/calmscope_black_white/calmscope_black_white.zip'\n",65        "with zipfile.ZipFile(zip_path,'r') as zip_ref:\n",66        "    zip_ref.extractall('/content/calmscope_black_white')\n",67        "\n",68        "IMG_SIZE = (224, 224)\n",69        "BATCH_SIZE = 32\n",70        "DATA_DIR = \"calmscope_black_white/calmscope_black_white\"   # folder utama dari Roboflow"71      ]72    },73    {74      "cell_type": "code",75      "execution_count": null,76      "metadata": {77        "id": "ssQpngZ9YZgK"78      },79      "outputs": [],80      "source": [81        "# 2. Fungsi load file dan label\n",82        "#    Label diambil dari prefix: Negative-..., Positive-...\n",83        "# -------------------------------------------------------------------\n",84        "def load_images_yolo_style(directory):\n",85        "    \"\"\"\n",86        "    Membaca gambar dari folder 'images' dan mengambil label dari\n",87        "    prefix nama file (negative / positive).\n",88        "    \"\"\"\n",89        "    data_dir = pathlib.Path(directory)\n",90        "\n",91        "    image_files = []\n",92        "    for ext in [\"*.jpg\", \"*.jpeg\", \"*.png\"]:\n",93        "        image_files += list(data_dir.glob(ext))\n",94        "\n",95        "    file_paths = []\n",96        "    labels = []\n",97        "\n",98        "    for img_path in image_files:\n",99        "        name = img_path.name.lower()\n",100        "\n",101        "        if name.startswith(\"negative\"):\n",102        "            labels.append(0)   # 0 = depressed\n",103        "        elif name.startswith(\"positive\"):\n",104        "            labels.append(1)   # 1 = neutral\n",105        "        else:\n",106        "            # file yang tidak punya prefix akan dilewati\n",107        "            print(\"Peringatan, dilewati:\", img_path.name)\n",108        "            continue\n",109        "\n",110        "        file_paths.append(str(img_path))\n",111        "\n",112        "    return file_paths, labels"113      ]114    },115    {116      "cell_type": "code",117      "execution_count": null,118      "metadata": {119        "id": "SARjeni9YbeC"120      },121      "outputs": [],122      "source": [123        "# 3. Fungsi membuat tf.data.Dataset\n",124        "#    Preprocessing langsung pakai preprocess_input EfficientNet\n",125        "# -------------------------------------------------------------------\n",126        "def create_dataset_from_paths(file_paths, labels,\n",127        "                              batch_size=BATCH_SIZE,\n",128        "                              shuffle=True,\n",129        "                              augment=False):\n",130        "\n",131        "    # fungsi untuk load satu gambar\n",132        "    def load_and_preprocess(path, label):\n",133        "        img = tf.io.read_file(path)\n",134        "        img = tf.image.decode_jpeg(img, channels=3)\n",135        "        img = tf.image.resize(img, IMG_SIZE)\n",136        "\n",137        "        # cast ke float32\n",138        "        img = tf.cast(img, tf.float32)\n",139        "\n",140        "        # preprocessing untuk EfficientNet\n",141        "        img = tf.keras.applications.efficientnet.preprocess_input(img)\n",142        "\n",143        "        return img, label\n",144        "\n",145        "    dataset = tf.data.Dataset.from_tensor_slices((file_paths, labels))\n",146        "\n",147        "    if shuffle:\n",148        "        dataset = dataset.shuffle(len(file_paths), seed=42)\n",149        "\n",150        "    dataset = dataset.map(load_and_preprocess,\n",151        "                          num_parallel_calls=tf.data.AUTOTUNE)\n",152        "\n",153        "    # augmentasi sederhana\n",154        "    if augment:\n",155        "        aug = tf.keras.Sequential([\n",156        "            layers.RandomFlip(\"horizontal\"),\n",157        "            layers.RandomRotation(0.1),\n",158        "            layers.RandomZoom(0.1),\n",159        "            layers.RandomTranslation(0.05, 0.05),\n",160        "        ])\n",161        "        dataset = dataset.map(lambda x, y: (aug(x, training=True), y),\n",162        "                              num_parallel_calls=tf.data.AUTOTUNE)\n",163        "\n",164        "    dataset = dataset.batch(batch_size)\n",165        "    dataset = dataset.prefetch(tf.data.AUTOTUNE)\n",166        "    return dataset"167      ]168    },169    {170      "cell_type": "code",171      "execution_count": null,172      "metadata": {173        "colab": {174          "base_uri": "https://localhost:8080/"175        },176        "id": "ZT5ZpzUwYdnm",177        "outputId": "6d3844d8-7cb1-4a02-d081-1af0c20d01c3"178      },179      "outputs": [180        {181          "output_type": "stream",182          "name": "stdout",183          "text": [184            "--- Memuat dataset ---\n",185            "Train dataset: 3293 images\n",186            "Valid dataset: 952 images\n",187            "Test dataset : 995 images\n",188            "Train labels: {0, 1}\n",189            "Valid labels: {0, 1}\n",190            "Test labels : {0, 1}\n"191          ]192        }193      ],194      "source": [195        "# 4. Memuat dataset train, valid, test\n",196        "# -------------------------------------------------------------------\n",197        "print(\"--- Memuat dataset ---\")\n",198        "\n",199        "train_paths, train_labels = load_images_yolo_style(f\"{DATA_DIR}/train/images\")\n",200        "val_paths,   val_labels   = load_images_yolo_style(f\"{DATA_DIR}/valid/images\")\n",201        "test_paths,  test_labels  = load_images_yolo_style(f\"{DATA_DIR}/test/images\")\n",202        "\n",203        "print(f\"Train dataset: {len(train_paths)} images\")\n",204        "print(f\"Valid dataset: {len(val_paths)} images\")\n",205        "print(f\"Test dataset : {len(test_paths)} images\")\n",206        "\n",207        "print(\"Train labels:\", set(train_labels))\n",208        "print(\"Valid labels:\", set(val_labels))\n",209        "print(\"Test labels :\", set(test_labels))\n",210        "\n",211        "train_ds = create_dataset_from_paths(train_paths, train_labels,\n",212        "                                     batch_size=BATCH_SIZE,\n",213        "                                     shuffle=True,\n",214        "                                     augment=True)\n",215        "\n",216        "val_ds = create_dataset_from_paths(val_paths, val_labels,\n",217        "                                   batch_size=BATCH_SIZE,\n",218        "                                   shuffle=False,\n",219        "                                   augment=False)\n",220        "\n",221        "test_ds = create_dataset_from_paths(test_paths, test_labels,\n",222        "                                    batch_size=BATCH_SIZE,\n",223        "                                    shuffle=False,\n",224        "                                    augment=False)\n",225        "\n",226        "class_names = [\"depressed\", \"neutral\"]\n"227      ]228    },229    {230      "cell_type": "code",231      "execution_count": null,232      "metadata": {233        "colab": {234          "base_uri": "https://localhost:8080/"235        },236        "id": "amvWpJYbYfhQ",237        "outputId": "2104e4c5-2d8d-4caf-be96-9c6fc3b2a860"238      },239      "outputs": [240        {241          "output_type": "stream",242          "name": "stdout",243          "text": [244            "Class weights: {np.int64(0): np.float64(0.9229260089686099), np.int64(1): np.float64(1.0911199469847581)}\n"245          ]246        }247      ],248      "source": [249        "# 5. Hitung class weight untuk mengatasi imbalance\n",250        "# -------------------------------------------------------------------\n",251        "unique_classes = np.unique(train_labels)\n",252        "cw = class_weight.compute_class_weight(\n",253        "    class_weight=\"balanced\",\n",254        "    classes=unique_classes,\n",255        "    y=train_labels\n",256        ")\n",257        "class_weight_dict = dict(zip(unique_classes, cw))\n",258        "print(\"Class weights:\", class_weight_dict)"259      ]260    },261    {262      "cell_type": "code",263      "execution_count": null,264      "metadata": {265        "colab": {266          "base_uri": "https://localhost:8080/",267          "height": 425268        },269        "id": "sGWa68JzYhYd",270        "outputId": "e1c8200d-fede-4bc8-a1aa-c83a5ba3d80b"271      },272      "outputs": [273        {274          "output_type": "stream",275          "name": "stdout",276          "text": [277            "\n",278            "--- Membangun EfficientNetB0 ---\n",279            "Downloading data from https://storage.googleapis.com/keras-applications/efficientnetb0_notop.h5\n",280            "\u001b[1m16705208/16705208\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 0us/step\n"281          ]282        },283        {284          "output_type": "display_data",285          "data": {286            "text/plain": [287              "\u001b[1mModel: \"EfficientNetB0_depression\"\u001b[0m\n"288            ],289            "text/html": [290              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Model: \"EfficientNetB0_depression\"</span>\n",291              "</pre>\n"292            ]293          },294          "metadata": {}295        },296        {297          "output_type": "display_data",298          "data": {299            "text/plain": [300              "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",301              "┃\u001b[1m \u001b[0m\u001b[1mLayer (type)                   \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape          \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m      Param #\u001b[0m\u001b[1m \u001b[0m┃\n",302              "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",303              "│ input (\u001b[38;5;33mInputLayer\u001b[0m)              │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m224\u001b[0m, \u001b[38;5;34m224\u001b[0m, \u001b[38;5;34m3\u001b[0m)    │             \u001b[38;5;34m0\u001b[0m │\n",304              "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",305              "│ efficientnetb0 (\u001b[38;5;33mFunctional\u001b[0m)     │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m7\u001b[0m, \u001b[38;5;34m7\u001b[0m, \u001b[38;5;34m1280\u001b[0m)     │     \u001b[38;5;34m4,049,571\u001b[0m │\n",306              "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",307              "│ gap (\u001b[38;5;33mGlobalAveragePooling2D\u001b[0m)    │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1280\u001b[0m)           │             \u001b[38;5;34m0\u001b[0m │\n",308              "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",309              "│ dropout1 (\u001b[38;5;33mDropout\u001b[0m)              │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1280\u001b[0m)           │             \u001b[38;5;34m0\u001b[0m │\n",310              "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",311              "│ fc1 (\u001b[38;5;33mDense\u001b[0m)                     │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)            │       \u001b[38;5;34m327,936\u001b[0m │\n",312              "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",313              "│ dropout2 (\u001b[38;5;33mDropout\u001b[0m)              │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",314              "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",315              "│ output (\u001b[38;5;33mDense\u001b[0m)                  │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1\u001b[0m)              │           \u001b[38;5;34m257\u001b[0m │\n",316              "└─────────────────────────────────┴────────────────────────┴───────────────┘\n"317            ],318            "text/html": [319              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",320              "┃<span style=\"font-weight: bold\"> Layer (type)                    </span>┃<span style=\"font-weight: bold\"> Output Shape           </span>┃<span style=\"font-weight: bold\">       Param # </span>┃\n",321              "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",322              "│ input (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">InputLayer</span>)              │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">224</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">224</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">3</span>)    │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",323              "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",324              "│ efficientnetb0 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Functional</span>)     │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">7</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">7</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1280</span>)     │     <span style=\"color: #00af00; text-decoration-color: #00af00\">4,049,571</span> │\n",325              "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",326              "│ gap (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">GlobalAveragePooling2D</span>)    │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1280</span>)           │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",327              "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",328              "│ dropout1 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dropout</span>)              │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1280</span>)           │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",329              "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",330              "│ fc1 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                     │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>)            │       <span style=\"color: #00af00; text-decoration-color: #00af00\">327,936</span> │\n",331              "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",332              "│ dropout2 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dropout</span>)              │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",333              "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",334              "│ output (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                  │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>)              │           <span style=\"color: #00af00; text-decoration-color: #00af00\">257</span> │\n",335              "└─────────────────────────────────┴────────────────────────┴───────────────┘\n",336              "</pre>\n"337            ]338          },339          "metadata": {}340        },341        {342          "output_type": "display_data",343          "data": {344            "text/plain": [345              "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m4,377,764\u001b[0m (16.70 MB)\n"346            ],347            "text/html": [348              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Total params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">4,377,764</span> (16.70 MB)\n",349              "</pre>\n"350            ]351          },352          "metadata": {}353        },354        {355          "output_type": "display_data",356          "data": {357            "text/plain": [358              "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m328,193\u001b[0m (1.25 MB)\n"359            ],360            "text/html": [361              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">328,193</span> (1.25 MB)\n",362              "</pre>\n"363            ]364          },365          "metadata": {}366        },367        {368          "output_type": "display_data",369          "data": {370            "text/plain": [371              "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m4,049,571\u001b[0m (15.45 MB)\n"372            ],373            "text/html": [374              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Non-trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">4,049,571</span> (15.45 MB)\n",375              "</pre>\n"376            ]377          },378          "metadata": {}379        }380      ],381      "source": [382        "# 6. Membangun model EfficientNetB0\n",383        "# -------------------------------------------------------------------\n",384        "print(\"\\n--- Membangun EfficientNetB0 ---\")\n",385        "\n",386        "base_model = tf.keras.applications.EfficientNetB0(\n",387        "    include_top=False,\n",388        "    weights=\"imagenet\",\n",389        "    input_shape=(IMG_SIZE[0], IMG_SIZE[1], 3) # Changed from 1 to 3\n",390        ")\n",391        "\n",392        "base_model.trainable = False   # tahap pertama, freeze\n",393        "\n",394        "inputs = tf.keras.Input(shape=(IMG_SIZE[0], IMG_SIZE[1], 3), name=\"input\") # Changed from 1 to 3\n",395        "\n",396        "# input sudah di-preprocess di dataset, jadi langsung ke base_model\n",397        "x = base_model(inputs, training=False)\n",398        "x = layers.GlobalAveragePooling2D(name=\"gap\")(x)\n",399        "x = layers.Dropout(0.2, name=\"dropout1\")(x)\n",400        "x = layers.Dense(256, activation=\"relu\", name=\"fc1\")(x)\n",401        "x = layers.Dropout(0.2, name=\"dropout2\")(x)\n",402        "outputs = layers.Dense(1, activation=\"sigmoid\", name=\"output\")(x)\n",403        "\n",404        "model = Model(inputs, outputs, name=\"EfficientNetB0_depression\")\n",405        "\n",406        "model.compile(\n",407        "    optimizer=tf.keras.optimizers.Adam(learning_rate=1e-4),\n",408        "    loss=\"binary_crossentropy\",\n",409        "    metrics=[\"accuracy\"]\n",410        ")\n",411        "\n",412        "model.summary()\n",413        "\n",414        "# -------------------------------------------------------------------\n",415        "# 7. Callback\n",416        "# -------------------------------------------------------------------\n",417        "early_stop = tf.keras.callbacks.EarlyStopping(\n",418        "    monitor=\"val_loss\",\n",419        "    patience=5,\n",420        "    restore_best_weights=True,\n",421        "    verbose=1\n",422        ")\n",423        "\n",424        "reduce_lr = tf.keras.callbacks.ReduceLROnPlateau(\n",425        "    monitor=\"val_loss\",\n",426        "    factor=0.2,\n",427        "    patience=3,\n",428        "    verbose=1\n",429        ")"430      ]431    },432    {433      "cell_type": "code",434      "execution_count": null,435      "metadata": {436        "colab": {437          "base_uri": "https://localhost:8080/"438        },439        "id": "4M-qdcMwYjFh",440        "outputId": "a583002e-e223-45ee-fb47-2f0c4f403a8e"441      },442      "outputs": [443        {444          "output_type": "stream",445          "name": "stdout",446          "text": [447            "\n",448            "--- Training tahap 1 (transfer learning) ---\n",449            "Epoch 1/50\n",450            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m112s\u001b[0m 776ms/step - accuracy: 0.5458 - loss: 0.6877 - val_accuracy: 0.6481 - val_loss: 0.6285 - learning_rate: 1.0000e-04\n",451            "Epoch 2/50\n",452            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m46s\u001b[0m 444ms/step - accuracy: 0.6579 - loss: 0.6133 - val_accuracy: 0.6733 - val_loss: 0.5903 - learning_rate: 1.0000e-04\n",453            "Epoch 3/50\n",454            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m82s\u001b[0m 449ms/step - accuracy: 0.6775 - loss: 0.5901 - val_accuracy: 0.6733 - val_loss: 0.5803 - learning_rate: 1.0000e-04\n",455            "Epoch 4/50\n",456            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m46s\u001b[0m 446ms/step - accuracy: 0.6729 - loss: 0.5801 - val_accuracy: 0.6996 - val_loss: 0.5731 - learning_rate: 1.0000e-04\n",457            "Epoch 5/50\n",458            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m46s\u001b[0m 448ms/step - accuracy: 0.7073 - loss: 0.5573 - val_accuracy: 0.6922 - val_loss: 0.5694 - learning_rate: 1.0000e-04\n",459            "Epoch 6/50\n",460            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m48s\u001b[0m 461ms/step - accuracy: 0.7101 - loss: 0.5639 - val_accuracy: 0.7069 - val_loss: 0.5644 - learning_rate: 1.0000e-04\n",461            "Epoch 7/50\n",462            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m46s\u001b[0m 446ms/step - accuracy: 0.7079 - loss: 0.5413 - val_accuracy: 0.7090 - val_loss: 0.5572 - learning_rate: 1.0000e-04\n",463            "Epoch 8/50\n",464            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m47s\u001b[0m 452ms/step - accuracy: 0.6922 - loss: 0.5682 - val_accuracy: 0.7017 - val_loss: 0.5610 - learning_rate: 1.0000e-04\n",465            "Epoch 9/50\n",466            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m46s\u001b[0m 445ms/step - accuracy: 0.7111 - loss: 0.5408 - val_accuracy: 0.6985 - val_loss: 0.5598 - learning_rate: 1.0000e-04\n",467            "Epoch 10/50\n",468            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 430ms/step - accuracy: 0.7248 - loss: 0.5497\n",469            "Epoch 10: ReduceLROnPlateau reducing learning rate to 1.9999999494757503e-05.\n",470            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m45s\u001b[0m 441ms/step - accuracy: 0.7248 - loss: 0.5496 - val_accuracy: 0.7038 - val_loss: 0.5575 - learning_rate: 1.0000e-04\n",471            "Epoch 11/50\n",472            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m47s\u001b[0m 459ms/step - accuracy: 0.7039 - loss: 0.5503 - val_accuracy: 0.7069 - val_loss: 0.5571 - learning_rate: 2.0000e-05\n",473            "Epoch 12/50\n",474            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m46s\u001b[0m 444ms/step - accuracy: 0.7143 - loss: 0.5371 - val_accuracy: 0.7048 - val_loss: 0.5549 - learning_rate: 2.0000e-05\n",475            "Epoch 13/50\n",476            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m46s\u001b[0m 445ms/step - accuracy: 0.7162 - loss: 0.5389 - val_accuracy: 0.7059 - val_loss: 0.5536 - learning_rate: 2.0000e-05\n",477            "Epoch 14/50\n",478            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m84s\u001b[0m 459ms/step - accuracy: 0.7445 - loss: 0.5143 - val_accuracy: 0.7017 - val_loss: 0.5530 - learning_rate: 2.0000e-05\n",479            "Epoch 15/50\n",480            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m46s\u001b[0m 444ms/step - accuracy: 0.7146 - loss: 0.5394 - val_accuracy: 0.7111 - val_loss: 0.5573 - learning_rate: 2.0000e-05\n",481            "Epoch 16/50\n",482            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m47s\u001b[0m 457ms/step - accuracy: 0.7188 - loss: 0.5315 - val_accuracy: 0.7048 - val_loss: 0.5522 - learning_rate: 2.0000e-05\n",483            "Epoch 17/50\n",484            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m81s\u001b[0m 445ms/step - accuracy: 0.7351 - loss: 0.5206 - val_accuracy: 0.7111 - val_loss: 0.5535 - learning_rate: 2.0000e-05\n",485            "Epoch 18/50\n",486            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m82s\u001b[0m 443ms/step - accuracy: 0.7271 - loss: 0.5327 - val_accuracy: 0.7059 - val_loss: 0.5523 - learning_rate: 2.0000e-05\n",487            "Epoch 19/50\n",488            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 441ms/step - accuracy: 0.7327 - loss: 0.5149\n",489            "Epoch 19: ReduceLROnPlateau reducing learning rate to 3.999999898951501e-06.\n",490            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m83s\u001b[0m 454ms/step - accuracy: 0.7327 - loss: 0.5149 - val_accuracy: 0.7122 - val_loss: 0.5542 - learning_rate: 2.0000e-05\n",491            "Epoch 20/50\n",492            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m46s\u001b[0m 443ms/step - accuracy: 0.7554 - loss: 0.5063 - val_accuracy: 0.7132 - val_loss: 0.5514 - learning_rate: 4.0000e-06\n",493            "Epoch 21/50\n",494            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m47s\u001b[0m 459ms/step - accuracy: 0.7209 - loss: 0.5302 - val_accuracy: 0.7101 - val_loss: 0.5507 - learning_rate: 4.0000e-06\n",495            "Epoch 22/50\n",496            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m46s\u001b[0m 442ms/step - accuracy: 0.7277 - loss: 0.5231 - val_accuracy: 0.7122 - val_loss: 0.5499 - learning_rate: 4.0000e-06\n",497            "Epoch 23/50\n",498            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m46s\u001b[0m 445ms/step - accuracy: 0.7335 - loss: 0.5188 - val_accuracy: 0.7080 - val_loss: 0.5496 - learning_rate: 4.0000e-06\n",499            "Epoch 24/50\n",500            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m83s\u001b[0m 456ms/step - accuracy: 0.7289 - loss: 0.5177 - val_accuracy: 0.7111 - val_loss: 0.5504 - learning_rate: 4.0000e-06\n",501            "Epoch 25/50\n",502            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m46s\u001b[0m 444ms/step - accuracy: 0.7310 - loss: 0.5193 - val_accuracy: 0.7143 - val_loss: 0.5510 - learning_rate: 4.0000e-06\n",503            "Epoch 26/50\n",504            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 447ms/step - accuracy: 0.7183 - loss: 0.5311\n",505            "Epoch 26: ReduceLROnPlateau reducing learning rate to 7.999999979801942e-07.\n",506            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m47s\u001b[0m 460ms/step - accuracy: 0.7184 - loss: 0.5310 - val_accuracy: 0.7143 - val_loss: 0.5499 - learning_rate: 4.0000e-06\n",507            "Epoch 27/50\n",508            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m80s\u001b[0m 444ms/step - accuracy: 0.7278 - loss: 0.5191 - val_accuracy: 0.7143 - val_loss: 0.5499 - learning_rate: 8.0000e-07\n",509            "Epoch 28/50\n",510            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m81s\u001b[0m 439ms/step - accuracy: 0.7228 - loss: 0.5238 - val_accuracy: 0.7153 - val_loss: 0.5498 - learning_rate: 8.0000e-07\n",511            "Epoch 28: early stopping\n",512            "Restoring model weights from the end of the best epoch: 23.\n"513          ]514        }515      ],516      "source": [517        "# 8. Training tahap 1 - transfer learning\n",518        "# -------------------------------------------------------------------\n",519        "print(\"\\n--- Training tahap 1 (transfer learning) ---\")\n",520        "\n",521        "history_1 = model.fit(\n",522        "    train_ds,\n",523        "    validation_data=val_ds,\n",524        "    epochs=50,\n",525        "    callbacks=[early_stop, reduce_lr]\n",526        ")"527      ]528    },529    {530      "cell_type": "code",531      "execution_count": null,532      "metadata": {533        "colab": {534          "base_uri": "https://localhost:8080/"535        },536        "id": "k4qUmeeCYkt6",537        "outputId": "2d99c6bf-7fba-4c41-841a-4d5046ceec82"538      },539      "outputs": [540        {541          "output_type": "stream",542          "name": "stdout",543          "text": [544            "\n",545            "--- Fine tuning EfficientNet ---\n",546            "Epoch 1/50\n",547            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m124s\u001b[0m 801ms/step - accuracy: 0.5802 - loss: 0.7179 - val_accuracy: 0.6891 - val_loss: 0.5782 - learning_rate: 5.0000e-06\n",548            "Epoch 2/50\n",549            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m48s\u001b[0m 470ms/step - accuracy: 0.6337 - loss: 0.6552 - val_accuracy: 0.6775 - val_loss: 0.5823 - learning_rate: 5.0000e-06\n",550            "Epoch 3/50\n",551            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m50s\u001b[0m 481ms/step - accuracy: 0.6444 - loss: 0.6290 - val_accuracy: 0.6775 - val_loss: 0.5861 - learning_rate: 5.0000e-06\n",552            "Epoch 4/50\n",553            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 457ms/step - accuracy: 0.6516 - loss: 0.6127\n",554            "Epoch 4: ReduceLROnPlateau reducing learning rate to 9.999999747378752e-07.\n",555            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m48s\u001b[0m 470ms/step - accuracy: 0.6517 - loss: 0.6126 - val_accuracy: 0.6859 - val_loss: 0.5806 - learning_rate: 5.0000e-06\n",556            "Epoch 5/50\n",557            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m83s\u001b[0m 481ms/step - accuracy: 0.6692 - loss: 0.5988 - val_accuracy: 0.6796 - val_loss: 0.5804 - learning_rate: 1.0000e-06\n",558            "Epoch 6/50\n",559            "\u001b[1m103/103\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m49s\u001b[0m 471ms/step - accuracy: 0.6720 - loss: 0.5947 - val_accuracy: 0.6849 - val_loss: 0.5800 - learning_rate: 1.0000e-06\n",560            "Epoch 6: early stopping\n",561            "Restoring model weights from the end of the best epoch: 1.\n"562          ]563        }564      ],565      "source": [566        "# 9. Fine tuning EfficientNet\n",567        "# -------------------------------------------------------------------\n",568        "print(\"\\n--- Fine tuning EfficientNet ---\")\n",569        "\n",570        "base_model.trainable = True\n",571        "\n",572        "# Bekukan layer awal (lebih umum)\n",573        "fine_tune_at = 150\n",574        "for layer in base_model.layers[:fine_tune_at]:\n",575        "    layer.trainable = False\n",576        "\n",577        "model.compile(\n",578        "    optimizer=tf.keras.optimizers.Adam(learning_rate=5e-6),\n",579        "    loss=\"binary_crossentropy\",\n",580        "    metrics=[\"accuracy\"]\n",581        ")\n",582        "\n",583        "history_2 = model.fit(\n",584        "    train_ds,\n",585        "    validation_data=val_ds,\n",586        "    epochs=50,\n",587        "    # class_weight=class_weight_dict,\n",588        "    callbacks=[early_stop, reduce_lr]\n",589        ")"590      ]591    },592    {593      "cell_type": "code",594      "execution_count": null,595      "metadata": {596        "id": "ktRijw1sYmSN",597        "colab": {598          "base_uri": "https://localhost:8080/"599        },600        "outputId": "6b35fad2-86f7-4ffd-b92e-78b386687db4"601      },602      "outputs": [603        {604          "output_type": "stream",605          "name": "stdout",606          "text": [607            "\n",608            "--- Evaluasi di test set ---\n",609            "\u001b[1m32/32\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m10s\u001b[0m 331ms/step - accuracy: 0.6732 - loss: 0.5686\n",610            "Test loss: 0.5519\n",611            "Test accuracy: 0.6894\n"612          ]613        }614      ],615      "source": [616        "# 10. Evaluasi di test set\n",617        "# -------------------------------------------------------------------\n",618        "print(\"\\n--- Evaluasi di test set ---\")\n",619        "test_loss, test_acc = model.evaluate(test_ds)\n",620        "print(f\"Test loss: {test_loss:.4f}\")\n",621        "print(f\"Test accuracy: {test_acc:.4f}\")"622      ]623    },624    {625      "cell_type": "code",626      "source": [627        "model.save(\"model.h5\")"628      ],629      "metadata": {630        "colab": {631          "base_uri": "https://localhost:8080/"632        },633        "id": "j9JWPx5b82YI",634        "outputId": "78a1d83f-f658-4208-971c-abe4ccf508bf"635      },636      "execution_count": null,637      "outputs": [638        {639          "output_type": "stream",640          "name": "stderr",641          "text": [642            "WARNING:absl:You are saving your model as an HDF5 file via `model.save()` or `keras.saving.save_model(model)`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')` or `keras.saving.save_model(model, 'my_model.keras')`. \n"643          ]644        }645      ]646    },647    {648      "cell_type": "code",649      "execution_count": null,650      "metadata": {651        "id": "vqKor3J7Ynrv",652        "colab": {653          "base_uri": "https://localhost:8080/"654        },655        "outputId": "b0757963-0e75-494b-bb64-43824e46588e"656      },657      "outputs": [658        {659          "output_type": "display_data",660          "data": {661            "text/plain": [662              "<Figure size 1200x400 with 2 Axes>"663            ],664            "image/png": "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\n"665          },666          "metadata": {}667        }668      ],669      "source": [670        "# 11. Plot kurva training\n",671        "# -------------------------------------------------------------------\n",672        "acc = history_1.history[\"accuracy\"] + history_2.history[\"accuracy\"]\n",673        "val_acc = history_1.history[\"val_accuracy\"] + history_2.history[\"val_accuracy\"]\n",674        "loss = history_1.history[\"loss\"] + history_2.history[\"loss\"]\n",675        "val_loss = history_1.history[\"val_loss\"] + history_2.history[\"val_loss\"]\n",676        "\n",677        "epochs_range = range(len(acc))\n",678        "\n",679        "plt.figure(figsize=(12, 4))\n",680        "\n",681        "plt.subplot(1, 2, 1)\n",682        "plt.plot(epochs_range, acc, label=\"Train acc\")\n",683        "plt.plot(epochs_range, val_acc, label=\"Val acc\")\n",684        "plt.xlabel(\"Epoch\")\n",685        "plt.ylabel(\"Accuracy\")\n",686        "plt.legend()\n",687        "\n",688        "plt.subplot(1, 2, 2)\n",689        "plt.plot(epochs_range, loss, label=\"Train loss\")\n",690        "plt.plot(epochs_range, val_loss, label=\"Val loss\")\n",691        "plt.xlabel(\"Epoch\")\n",692        "plt.ylabel(\"Loss\")\n",693        "plt.legend()\n",694        "\n",695        "plt.tight_layout()\n",696        "plt.show()"697      ]698    },699    {700      "cell_type": "code",701      "execution_count": null,702      "metadata": {703        "id": "t7wy4JkSYbog",704        "colab": {705          "base_uri": "https://localhost:8080/"706        },707        "outputId": "214f2910-0f7a-4d92-9998-066c155b4475"708      },709      "outputs": [710        {711          "output_type": "stream",712          "name": "stdout",713          "text": [714            "\u001b[1m32/32\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m15s\u001b[0m 250ms/step\n",715            "=== Hasil dengan threshold akurasi (0.56) ===\n",716            "              precision    recall  f1-score   support\n",717            "\n",718            "   depressed       0.72      0.84      0.78       642\n",719            "     neutral       0.58      0.40      0.47       353\n",720            "\n",721            "    accuracy                           0.69       995\n",722            "   macro avg       0.65      0.62      0.62       995\n",723            "weighted avg       0.67      0.69      0.67       995\n",724            "\n"725          ]726        },727        {728          "output_type": "display_data",729          "data": {730            "text/plain": [731              "<Figure size 400x400 with 2 Axes>"732            ],733            "image/png": "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\n"734          },735          "metadata": {}736        },737        {738          "output_type": "stream",739          "name": "stdout",740          "text": [741            "\n",742            "=== Hasil dengan threshold recall depressed (0.70) ===\n",743            "              precision    recall  f1-score   support\n",744            "\n",745            "   depressed       0.68      0.95      0.79       642\n",746            "     neutral       0.67      0.20      0.31       353\n",747            "\n",748            "    accuracy                           0.68       995\n",749            "   macro avg       0.68      0.57      0.55       995\n",750            "weighted avg       0.68      0.68      0.62       995\n",751            "\n"752          ]753        },754        {755          "output_type": "display_data",756          "data": {757            "text/plain": [758              "<Figure size 400x400 with 2 Axes>"759            ],760            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAGJCAYAAACEpvgnAAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjEwLjAsIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvlHJYcgAAAAlwSFlzAAAPYQAAD2EBqD+naQAATrRJREFUeJzt3XdYFNf6wPHvgrB0EKVasYslxhJF7BKJLfZe0KApFyuihGgSJSq5mliS3IjmJmq8eo1RYxI79qiE2LDFHhULYEFALIDs/P7wx96soLK0hd334zPP4545M/POLrycPXNmjkpRFAUhhBBGy8zQAQghhChakuiFEMLISaIXQggjJ4leCCGMnCR6IYQwcpLohRDCyEmiF0IIIyeJXgghjJwkeiGEMHKS6PPgwoULdOrUCUdHR1QqFRs2bCjU/V+5cgWVSsWyZcsKdb+lWbt27WjXrp2hw9Axffp0VCoVd+7cMXQoQNHEk9f3fc+ePahUKvbs2VNoxxZFp9Qk+kuXLvHOO+9QrVo1rKyscHBwwNfXl4ULF/Lo0aMiPXZAQAAnT55k1qxZrFixgqZNmxbp8YrTiBEjUKlUODg45Po+XrhwAZVKhUql4rPPPtN7/zdv3mT69OnExsYWQrTFY/bs2YX+x9yUpaenExoaiqenJ9bW1jRv3pyoqKg8bVu1alXtz9+zS82aNXPU//bbb6lbty5WVlbUrFmTL7/8srBPp1QqY+gA8mLTpk3069cPtVrN8OHDqV+/PhkZGezfv5/Jkydz+vRplixZUiTHfvToEdHR0UydOpUxY8YUyTGqVKnCo0ePsLCwKJL9v0yZMmV4+PAhv/76K/3799dZt3LlSqysrHj8+HG+9n3z5k1mzJhB1apVadSoUZ632759e76OVxhmz55N37596dmzp8FiMCYjRoxg7dq1TJgwgZo1a7Js2TK6dOnC7t27adWq1Qu3XbBgAWlpaTplV69eZdq0aXTq1EmnfPHixbz77rv06dOH4OBgfvvtN8aNG8fDhw8JDQ0t9PMqTUp8or98+TIDBw6kSpUq7Nq1Cw8PD+26oKAgLl68yKZNm4rs+Ldv3wbAycmpyI6hUqmwsrIqsv2/jFqtxtfXl//+9785Ev2qVavo2rUr69atK5ZYHj58iI2NDZaWlsVyvOKi0WjIyMgw6OdsCH/88QerV69m7ty5hISEAGgba1OmTOHgwYMv3D63P7YzZ84EYMiQIdqyR48eMXXqVLp27cratWsBGD16NBqNhk8++YS3336bsmXLFtJZlT4lvutmzpw5pKWl8e233+ok+Ww1atRg/Pjx2tdPnjzhk08+oXr16qjVaqpWrcoHH3xAenq6znZVq1alW7du7N+/n9deew0rKyuqVavG999/r60zffp0qlSpAsDkyZNRqVRUrVoVeNpKyf7/32X3m/5dVFQUrVq1wsnJCTs7O2rXrs0HH3ygXf+8Pvpdu3bRunVrbG1tcXJyokePHpw5cybX4128eJERI0bg5OSEo6MjI0eO5OHDh89/Y58xePBgtmzZQnJysrbs0KFDXLhwgcGDB+eon5SUREhICA0aNMDOzg4HBwc6d+7M8ePHtXX27NlDs2bNABg5cqT2K3f2ebZr14769etz5MgR2rRpg42NjfZ9ebavOCAgACsrqxzn7+/vT9myZbl582aez/VFVCoVDx48YPny5dp4R4wYoVMnOTn5pe+1SqVizJgxrFy5knr16qFWq9m6dSsAN27c4K233sLNzQ21Wk29evX47rvvcsTy5ZdfUq9ePWxsbChbtixNmzZl1apVOerlJZ68/l7k5vr16/Ts2RNbW1tcXV2ZOHFinrYDWLt2Lebm5rz99tvaMisrKwIDA4mOjubatWt52s/frVq1Ci8vL1q2bKkt2717N3fv3uUf//iHTt2goCAePHhQpI3B0qDEt+h//fVXqlWrpvOhvsioUaNYvnw5ffv2ZdKkScTExBAREcGZM2f46aefdOpevHiRvn37EhgYSEBAAN999x0jRoygSZMm1KtXj969e+Pk5MTEiRMZNGgQXbp0wc7OTq/4T58+Tbdu3WjYsCHh4eGo1WouXrzIgQMHXrjdjh076Ny5M9WqVWP69Ok8evSIL7/8El9fX44ePZrjj0z//v3x8vIiIiKCo0eP8u9//xtXV1f++c9/5inO3r178+6777J+/Xreeust4OkvVJ06dWjcuHGO+n/99RcbNmygX79+eHl5kZiYyOLFi2nbti1//vknnp6e1K1bl/DwcD766CPefvttWrduDaDzWd69e5fOnTszcOBAhg4dipubW67xLVy4kF27dhEQEEB0dDTm5uYsXryY7du3s2LFCjw9PfN0ni+zYsUKRo0axWuvvaZNTtWrV9epk9f3eteuXaxZs4YxY8ZQvnx5qlatSmJiIi1atND+IXBxcWHLli0EBgaSmprKhAkTAPjmm28YN24cffv2Zfz48Tx+/JgTJ04QExOT4w9vXuLR5/fi7x49ekTHjh2Ji4tj3LhxeHp6smLFCnbt2pWn9/PYsWPUqlULBwcHnfLXXnsNgNjYWCpVqpSnfWXv78yZM0ydOjVHOZDj+lmTJk0wMzPj2LFjDB06NM/HMTpKCZaSkqIASo8ePfJUPzY2VgGUUaNG6ZSHhIQogLJr1y5tWZUqVRRA2bdvn7bs1q1bilqtViZNmqQtu3z5sgIoc+fO1dlnQECAUqVKlRwxfPzxx8rf39b58+crgHL79u3nxp19jKVLl2rLGjVqpLi6uip3797Vlh0/flwxMzNThg8fnuN4b731ls4+e/XqpZQrV+65x/z7edja2iqKoih9+/ZVOnbsqCiKomRlZSnu7u7KjBkzcn0PHj9+rGRlZeU4D7VarYSHh2vLDh06lOPcsrVt21YBlMjIyFzXtW3bVqds27ZtCqDMnDlT+euvvxQ7OzulZ8+eLz1Hfdna2ioBAQE5yvV5rwHFzMxMOX36tE55YGCg4uHhody5c0enfODAgYqjo6Py8OFDRVEUpUePHkq9evVeGGde49Hn9+LZ933BggUKoKxZs0Zb9uDBA6VGjRoKoOzevfuFMdarV0/p0KFDjvLTp08/97N/kUmTJimA8ueff+qUBwUFKebm5rlu4+LiogwcOFCv4xibEt11k5qaCoC9vX2e6m/evBmA4OBgnfJJkyYB5Pj65u3trW1lAri4uFC7dm3++uuvfMf8rOy+/Z9//hmNRpOnbeLj44mNjWXEiBE4Oztryxs2bMjrr7+uPc+/e/fdd3Vet27dmrt372rfw7wYPHgwe/bsISEhgV27dpGQkJBrtw087dc3M3v645OVlcXdu3e13VJHjx7N8zHVajUjR47MU91OnTrxzjvvEB4eTu/evbGysmLx4sV5PlZhyet73bZtW7y9vbWvFUVh3bp1dO/eHUVRuHPnjnbx9/cnJSVF+945OTlx/fp1Dh06VOB49P29+LvNmzfj4eFB3759tWU2NjY6XTEv8ujRI9RqdY7y7GsV+oyY02g0rF69mldffZW6devmOM7zrutYWVkV+ci8kq5EJ/rsr3v379/PU/2rV69iZmZGjRo1dMrd3d1xcnLi6tWrOuWVK1fOsY+yZcty7969fEac04ABA/D19WXUqFG4ubkxcOBA1qxZ88Kknx1n7dq1c6yrW7cud+7c4cGDBzrlz55L9oUnfc6lS5cu2Nvb88MPP7By5UqaNWuW473MptFomD9/PjVr1kStVlO+fHlcXFw4ceIEKSkpeT5mhQoV9Lrw+tlnn+Hs7ExsbCxffPEFrq6uL93m9u3bJCQkaJdnR3HoK6/vtZeXV444kpOTWbJkCS4uLjpL9h+7W7duARAaGoqdnR2vvfYaNWvWJCgo6LndfS+LR9/fi7+7evUqNWrUyHHdKbefzdxYW1vn2p+fPYrL2to6T/sB2Lt3Lzdu3NC5CPv342RkZOS63ePHj/U6jjEq8Yne09OTU6dO6bXdsz+Uz2Nubp5ruZKH2RWfd4ysrCyd19bW1uzbt48dO3YwbNgwTpw4wYABA3j99ddz1C2IgpxLNrVaTe/evVm+fDk//fTTc1vz8HQIYnBwMG3atOE///kP27ZtIyoqinr16uX5mwvo94sOT/tis5PhyZMn87RNs2bN8PDw0C75uR/g7/L6Xj97btnvy9ChQ4mKisp18fX1BZ7+QT937hyrV6+mVatWrFu3jlatWvHxxx/nO568/l4UJg8PD+Lj43OUZ5fpc21l5cqVmJmZMWjQoFyPk5WVpf3ZyJaRkcHdu3cL7RpOaVXiL8Z269aNJUuWEB0djY+PzwvrVqlSBY1Gw4ULF3S+2iUmJpKcnKwdQVMYypYtqzNCJVturSMzMzM6duxIx44dmTdvHrNnz2bq1Kns3r0bPz+/XM8D4Ny5cznWnT17lvLly2Nra1vwk8jF4MGD+e677zAzM2PgwIHPrbd27Vrat2/Pt99+q1OenJxM+fLlta8LM7k8ePCAkSNH4u3tTcuWLZkzZw69evXSjux5npUrV+p8da9WrdoL6xdVQnRxccHe3p6srKxcP/dn2draMmDAAAYMGEBGRga9e/dm1qxZhIWF6TVMsyC/F1WqVOHUqVMoiqLzvuT2s5mbRo0asXv3blJTU3UuyMbExGjX50V6ejrr1q2jXbt2uSbt7P0cPnyYLl26aMsPHz6MRqPR6x4OY1SiW/QAU6ZMwdbWllGjRpGYmJhj/aVLl1i4cCGA9gNesGCBTp158+YB0LVr10KLq3r16qSkpHDixAltWXx8fI4RDElJSTm2zf6he94QNQ8PDxo1asTy5ct1/picOnWK7du36/wgF7b27dvzySef8NVXX+Hu7v7ceubm5jlajD/++CM3btzQKcv+g5TbH0V9hYaGEhcXx/Lly5k3bx5Vq1YlICDgpUP9fH198fPz0y4vS/S2traFEu+zzM3N6dOnD+vWrcv1W2r2PRvwdDTS31laWuLt7Y2iKGRmZup13IL8XnTp0oWbN29qx6bD03sd8nqDYt++fcnKytKpn56eztKlS2nevLnOiJu4uDjOnj2b6342b95McnJyrt02AB06dMDZ2ZlFixbplC9atAgbG5tC/d0vjUp8i7569eqsWrWKAQMGULduXZ07Yw8ePMiPP/6oHef8yiuvEBAQwJIlS0hOTqZt27b88ccfLF++nJ49e9K+fftCi2vgwIGEhobSq1cv7d13ixYtolatWjoXI8PDw9m3bx9du3alSpUq3Lp1i6+//pqKFSu+8K7AuXPn0rlzZ3x8fAgMDNQOr3R0dGT69OmFdh7PMjMzY9q0aS+t161bN8LDwxk5ciQtW7bk5MmTrFy5MkcSrV69Ok5OTkRGRmJvb4+trS3NmzfP0X/9Mrt27eLrr7/m448/1g73XLp0Ke3atePDDz9kzpw5eu3vRZo0acKOHTuYN28enp6eeHl50bx580LZ96effsru3btp3rw5o0ePxtvbm6SkJI4ePcqOHTu0DYNOnTrh7u6Or68vbm5unDlzhq+++oquXbvmeXBCtoL8XowePZqvvvqK4cOHc+TIETw8PFixYgU2NjZ5Onbz5s3p168fYWFh3Lp1ixo1arB8+XKuXLmS49vg8OHD2bt3b67djStXrkStVtOnT59cj2Ntbc0nn3xCUFAQ/fr1w9/fn99++43//Oc/zJo1S2dQg0ky1HAffZ0/f14ZPXq0UrVqVcXS0lKxt7dXfH19lS+//FJ5/Pixtl5mZqYyY8YMxcvLS7GwsFAqVaqkhIWF6dRRlKfDK7t27ZrjOM8OL3ve8EpFUZTt27cr9evXVywtLZXatWsr//nPf3IMr9y5c6fSo0cPxdPTU7G0tFQ8PT2VQYMGKefPn89xjGeHIO7YsUPx9fVVrK2tFQcHB6V79+45hpVlH+/Z4ZtLly5VAOXy5cvPfU8VRXd45fM8b3jlpEmTFA8PD8Xa2lrx9fVVoqOjcx0W+fPPPyve3t5KmTJldM6zbdu2zx1C+Pf9pKamKlWqVFEaN26sZGZm6tSbOHGiYmZmpkRHR7/wHPRx9uxZpU2bNoq1tbUCaIda6vNeA0pQUFCu+09MTFSCgoKUSpUqKRYWFoq7u7vSsWNHZcmSJdo6ixcvVtq0aaOUK1dOUavVSvXq1ZXJkycrKSkp2jr6xJPX34vcPr+rV68qb775pmJjY6OUL19eGT9+vLJ169Y8Da9UFEV59OiREhISori7uytqtVpp1qyZsnXr1hz1sofbPislJUWxsrJSevfu/dJjLVmyRKldu7ZiaWmpVK9eXZk/f76i0Wheup2xUymKHlfrhBBClDolvo9eCCFEwUiiF0IIIyeJXgghjJwkeiGEMHKS6IUQwshJohdCCCMniV4IIYxcib8zNj8y7xTeY4ZFyWft2frllYTReJJx4+WVXqAg+cGi/Isfn/GsGzduEBoaypYtW3j48CE1atRg6dKl2glSFEXh448/5ptvviE5ORlfX18WLVqkM/F5UlISY8eO5ddff8XMzIw+ffqwcOFCvSZBkha9EMK0aLLyv+jh3r17+Pr6YmFhwZYtW/jzzz/5/PPPdeaunTNnDl988QWRkZHExMRga2uLv7+/9jHO8HRu3NOnTxMVFcXGjRvZt29fnucDyGaUd8ZKi960SIvetBS4RZ+Ytydv5kbjVDXHQ/TUanWuk6u8//77HDhwgN9++y3XfSmKgqenJ5MmTdJOnJ6SkoKbmxvLli1j4MCBnDlzBm9vbw4dOqT9FrB161a6dOnC9evX8/z4ZWnRCyFMi0aT7yUiIgJHR0edJSIiItfD/PLLLzRt2pR+/frh6urKq6++yjfffKNdf/nyZRISEnQeWe3o6Ejz5s2Jjo4GIDo6GicnJ525cP38/DAzM9M+6jkvJNELIUQehYWFkZKSorOEhYXlWvevv/7S9rdv27aN9957j3HjxrF8+XIAEhISAHBzc9PZzs3NTbsuISEhxyxqZcqUwdnZWVsnL4zyYqwQQjyPouR9BrRnPa+bJjcajYamTZsye/ZsAF599VVOnTpFZGQkAQEB+Y4hP6RFL4QwLQXoutGHh4eHzuTw8HSKyLi4OADtxD7PTqiUmJioXefu7p5jesQnT56QlJT0womBniWJXghhWhRN/hc9+Pr65phy8fz589qpG728vHB3d2fnzp3a9ampqcTExGinTfXx8SE5OZkjR45o6+zatQuNRqPXZDjSdSOEMC16DpPMr4kTJ9KyZUtmz55N//79+eOPP1iyZIl2WkWVSsWECROYOXMmNWvWxMvLiw8//BBPT0969uwJPP0G8MYbbzB69GgiIyPJzMxkzJgxDBw4UK8JzyXRCyFMSwH66PXRrFkzfvrpJ8LCwggPD8fLy4sFCxbozHs7ZcoUHjx4wNtvv01ycjKtWrVi69atOpO/r1y5kjFjxtCxY0ftDVNffPGFXrHIOHpR6sk4etNS0HH0GVcO53tby6pNX16pBJIWvRDCtOh5UdUYSKIXQpiUggyvLK0k0QshTIu06IUQwshJi14IIYxcMQ2vLEnkhikhhDBy0qIXQpgW6boRQggjJxdjhRDCyEmLXgghjJy06IUQwrgpioy6EUIIYWSkRS+EMC3SRy+EEEZO+uiFEMLISYteCCGMnAk+AkESvRDCtJhgi15G3QghhJGTFr0QwrTIxVghhDByJth1I4leCGFapEUvhBBGThK9EEIYN3nWjRBCCKMjLXohhGmRrhshhDByMupGCCGMnLTohRDCyEmLXgghjJwJtuhl1I0QQhg5adELIUyLdN0IIYSRM8GuG0n0QgjTIom+eAQHB+e57rx584owEiGEyZGum+Jx7NgxnddHjx7lyZMn1K5dG4Dz589jbm5OkyZNDBGeEMKYSYu+eOzevVv7/3nz5mFvb8/y5cspW7YsAPfu3WPkyJG0bt3aEOEJIYRRUSmKohgygAoVKrB9+3bq1aunU37q1Ck6derEzZs39d5n5p2/Cis8UQpYe0qDwJQ8ybhRoO0f/Twn39ta95hSoGMbisEvxqampnL79u0c5bdv3+b+/fsGiEgIYdRMsOvG4DdM9erVi5EjR7J+/XquX7/O9evXWbduHYGBgfTu3dvQ4QkhjI2iyf9SShm8RR8ZGUlISAiDBw8mMzMTgDJlyhAYGMjcuXMNHJ0QwuiYYIve4H302R48eMClS5cAqF69Ora2tvnel/TRmxbpozctBe6jXxOe722t+39UoGMbisG7brLFx8cTHx9PzZo1sbW1pYT8/RFCiFLP4In+7t27dOzYkVq1atGlSxfi4+MBCAwMZNKkSQaOTghhdBQl/4sepk+fjkql0lnq1KmjXf/48WOCgoIoV64cdnZ29OnTh8TERJ19xMXF0bVrV2xsbHB1dWXy5Mk8efJE71M2eKKfOHEiFhYWxMXFYWNjoy0fMGAAW7duNWBkQgijpNHkf9FTvXr1tL0V8fHx7N+/X7tu4sSJ/Prrr/z444/s3buXmzdv6gxAycrKomvXrmRkZHDw4EGWL1/OsmXL+Ogj/buPDH4xdvv27Wzbto2KFSvqlNesWZOrV68aKCohhNEqxouxZcqUwd3dPUd5SkoK3377LatWraJDhw4ALF26lLp16/L777/TokULtm/fzp9//smOHTtwc3OjUaNGfPLJJ4SGhjJ9+nQsLS3zHIfBW/QPHjzQaclnS0pKQq1WGyAiIYRRK8DwyvT0dFJTU3WW9PT05x7qwoULeHp6Uq1aNYYMGUJcXBwAR44cITMzEz8/P23dOnXqULlyZaKjowGIjo6mQYMGuLm5aev4+/uTmprK6dOn9Tplgyf61q1b8/3332tfq1QqNBoNc+bMoX379gaMTAhhlArQdRMREYGjo6POEhERkethmjdvzrJly9i6dSuLFi3i8uXLtG7dmvv375OQkIClpSVOTk4627i5uZGQkABAQkKCTpLPXp+9Th8G77qZM2cOHTt25PDhw2RkZDBlyhROnz5NUlISBw4cMHR4QgihFRYWluPpu8/reejcubP2/w0bNqR58+ZUqVKFNWvWYG1tXaRxPsvgLfr69etz/vx5WrVqRY8ePXjw4AG9e/fm2LFjVK9e3dDhCSGMTQFG3ajVahwcHHSWvHYxOzk5UatWLS5evIi7uzsZGRkkJyfr1ElMTNT26bu7u+cYhZP9Ord+/xcxeIsewNHRkalTpxo6DCGEKTDQnbFpaWlcunSJYcOG0aRJEywsLNi5cyd9+vQB4Ny5c8TFxeHj4wOAj48Ps2bN4tatW7i6ugIQFRWFg4MD3t7eeh3b4C36rVu36gw5+te//kWjRo0YPHgw9+7dM2BkQgijVEzDK0NCQti7dy9Xrlzh4MGD9OrVC3NzcwYNGoSjoyOBgYEEBweze/dujhw5wsiRI/Hx8aFFixYAdOrUCW9vb4YNG8bx48fZtm0b06ZNIygoSO+BKgZP9JMnTyY1NRWAkydPEhwcTJcuXbh8+bJeM1EJIUSeFNNDza5fv86gQYOoXbs2/fv3p1y5cvz++++4uLgAMH/+fLp160afPn1o06YN7u7urF+/Xru9ubk5GzduxNzcHB8fH4YOHcrw4cMJD9f/EQ4Gf9aNnZ0dp06domrVqkyfPp1Tp06xdu1ajh49SpcuXfS+ugzyrBtTI8+6MS0FfdbNwyUT872tzdvzC3RsQzF4i97S0pKHDx8CsGPHDjp16gSAs7OztqUvhBAi/wx+MbZVq1YEBwfj6+vLH3/8wQ8//AA8nTf22btlhRCiwEzwMcUGb9F/9dVXlClThrVr17Jo0SIqVKgAwJYtW3jjjTcMHJ0QwuiY4MQjBu+jLwrSR29apI/etBS4j/5fY/K9rU3QVwU6tqEYvEV/9OhRTp48qX39888/07NnTz744AMyMjIMGJkQwigV49MrSwqDJ/p33nmH8+fPA/DXX38xcOBAbGxs+PHHH5kypXTOuC6EECWJwRP9+fPnadSoEQA//vgjbdq0YdWqVSxbtox169YZNjghhPGRFn3xUxQFzf+/gTt27KBLly4AVKpUiTt37hgytBIh8fYdQmfMwbdzf5q070GvYe9x6sx57XpFUfjqm+9p9+ZgmrTvwajxYVy99r8+zD+OnqC+b+dcl5NnzhnilIQe3nl7OEePRJF05yxJd86yf98vvOH/v6e67oz6kScZN3SWf331qQEjLgWKaYapksTgwyubNm3KzJkz8fPzY+/evSxatAiAy5cv53hEp6lJSb3PsHcn8VrjV4j8/BPKOjly9doNHOzttHW+W/kjK9f+wqxpk6jg4c5X33zPO8HT+Pk/i1GrLXm1QV32/LJSZ79ffrOCmCOx1K9Tq7hPSejpxo14pk6N4MLFy6hUKoYP68f6dd/R9DV//vzz6R/8b/79H6bP+Ey7zcOHjwwVbulQilvm+WXwRL9gwQKGDBnChg0bmDp1KjVq1ABg7dq1tGzZ0sDRGdZ3K3/E3dWFmVP/9yiIip7/e2qdoiisWLOBtwMG0qH10wchzf4whLbdB7Hzt4N08WuHhYUF5cs5a7fJfPKE3b9FM7jvm6hUquI7GZEvGzdF6bz+8KN/8s7bw2j+WmNton/48DGJibcNEV7ppCm9LfP8Mniib9iwoc6om2xz587F3NzcABGVHLv3/47va00InjaLw8dO4upSjoG9u9H3zafPub5+M4E7d+/h0/RV7Tb2drY09K7N8VNn6eLXLsc+9/z2O8mp9+nZ9fXiOg1RSMzMzOjbtxu2tjb8HnNEWz54UC+GDO5NQuItNm2KYuasBTx69NiAkZZwpXg8fH4ZPNEDJCcns3btWi5dusTkyZNxdnbmzz//xM3NTXsDlSm6fjOBHzZsYviA3owePoBTZ84TMT8SizJl6NHlde4kPX26ZznnsjrblXMuy527uT/5c/3Gbfi+1hh3V5cij18Ujvr167B/3y9YWalJS3tA336jOHPmAgD/Xb2BuLjr3IxPpEGDukTMmkqtWtXp13+0gaMWJYnBE/2JEyfo2LEjTk5OXLlyhdGjR+Ps7Mz69euJi4vTmWYwN+np6TnmbDRLTzeK+WY1GoV6dWoy4d0RANStVYMLf11lzYbN9Oiif4s84dZtDvxxlM/Dwwo5UlGUzp27RJNmnXB0sKdPn6589+0COvj14cyZC/z72/9dfzl16iwJ8beI2r6GatWq8NdfVw0YdQlmgl03Bh91ExwczMiRI7lw4QJWVlba8i5durBv376Xbp/bHI7/XBhZlCEXG5dyzlSvWlmnrFrVSsT/f39s+f9vyd9N0m293026R/lyuq18gA2bonBysKdd6xZFFLEoCpmZmVy6dIWjx04yddqnnDjxJ2PHjMq1bswfRwGoUb1qMUZYuigaTb6X0srgif7QoUO88847OcorVKiQp0cUh4WFkZKSorOEjn+3KEItdq829OZK3HWdsqtxN/BwfzrbTEVPd8qXK8vvR2K169MePODEn+d4pX4dne0URWHD5ii6d+6IRRmDf5ETBWBmZoZabZnrukav1AMgPuFWcYZUumiU/C+llMF/49Vqda6PIz5//rz2Af0v2/7ZbprMDOMYfz9sQE+GvTOJJctX80bHNpz88xxrf9nCx1PGAaBSqRjWvydLlq+mSsUKVPB046tvVuBavhwdW+uOWIo5Esv1mwn06S4PiitNZs18n61bdxN37Qb29nYMGtiTtm196NJ1MNWqVWHQwF5s2bKTu0n3aNCgLp/Pnc6+fdGcPHnG0KGXXHIxtvi9+eabhIeHs2bNGuBp8oqLiyM0NFQ7l6KpalC3NgsiPmRh5DIil62igoc7oePfoZt/B22dt4b049Gjx0yf8wX309Jo3LAekZ9/kqPFt37jdho18KZalUrFfRqiAFxcyrP0u4V4eLiSknKfkyfP0KXrYHbs/I2KFT3p2KEV48aOwtbWmmvX4vlpw2ZmzV5o6LBLtlLcMs8vgz+9MiUlhb59+3L48GHu37+Pp6cnCQkJ+Pj4sHnzZmxtbfXepzy90rTI0ytNS0GfXvkgfEi+t7X9aOXLK5VABm/ROzo6EhUVxYEDBzh+/DhpaWk0btwYPz8/Q4cmhDBGpfiian4ZNNFnZmZibW1NbGwsvr6++Pr6GjIcIYQpMMGuG4MmegsLCypXrkxWVpYhwxBCmBITvBhr8OGVU6dO5YMPPiApKcnQoQghTIEMryx+X331FRcvXsTT05MqVarkuPh69OhRA0UmhDBGpfnGp/wyeKLv2bOnoUMQQgijZvBE//HHHxs6BCGEKSnFXTD5ZfBEn+3w4cOcOfP0bj5vb2+aNGli4IiEEEZJEn3xu379OoMGDeLAgQM4OTkBTx9b3LJlS1avXk3FihUNG6AQwrjIqJviN2rUKDIzMzlz5gxJSUkkJSVx5swZNBoNo0bl/oQ+IYTINxl1U/z27t3LwYMHqV27trasdu3afPnll7RuLbe2CyEKl1KKE3Z+GbxFX6lSJTIzM3OUZ2Vl4enpaYCIhBDCuBg80c+dO5exY8dy+PBhbdnhw4cZP348n3322Qu2FEKIfDDBrhuDP72ybNmyPHz4kCdPnlDm/yfEyP7/szdP5fXuWXl6pWmRp1ealoI+vfL+mC753tb+q80FOrahGLyPfsGCBYYOQQhhSkpxyzy/DJ7oAwICDB2CEMKUmGCiN3gfPcClS5eYNm0agwYN4tatp3NdbtmyhdOnTxs4MiGEsVEUJd9LaWXwRL93714aNGhATEwM69evJy0tDYDjx4/L4xGEEKIQGDzRv//++8ycOZOoqCgsLf83z2mHDh34/fffDRiZEMIomeCoG4P30Z88eZJVq1blKHd1deXOnTsGiEgIYdRKccLOL4O36J2cnIiPj89RfuzYMSpUqGCAiIQQxkzRKPleSiuDJ/qBAwcSGhpKQkICKpUKjUbDgQMHCAkJYfjw4YYOTwhhbEyw68bgiX727NnUqVOHSpUqkZaWhre3N61bt6Zly5ZMmzbN0OEJIYyNpgBLKWXwO2OzXbt2jZMnT5KWlsarr75KzZo1870vuTPWtMidsaaloHfGpgzrmO9tHVfsLNCxDcUgF2ODg4NfuP7vo23mzZtX1OEIIUxIae5rzy+DdN0cO3ZMZ/n2229ZvHgxe/bsYc+ePSxZsoRvv/2W2NhYQ4QnhDBmBuqj//TTT1GpVEyYMEFb9vjxY4KCgihXrhx2dnb06dOHxMREne3i4uLo2rUrNjY2uLq6MnnyZJ48eaLXsQ3Sot+9e7f2//PmzcPe3p7ly5dTtmxZAO7du8fIkSPlefRCiMJngL72Q4cOsXjxYho2bKhTPnHiRDZt2sSPP/6Io6MjY8aMoXfv3hw4cAB4+rj2rl274u7uzsGDB4mPj2f48OFYWFgwe/bsPB/f4H30FSpUYPv27dSrV0+n/NSpU3Tq1ImbN2/qvU/pozct0kdvWgraR3+vX7t8b1v2xz16b5OWlkbjxo35+uuvmTlzJo0aNWLBggWkpKTg4uLCqlWr6Nu3LwBnz56lbt26REdH06JFC7Zs2UK3bt24efMmbm5uAERGRhIaGsrt27d1bjJ9EYOPuklNTeX27ds5ym/fvs39+/cNEJEQwqgVYNRNeno6qampOkt6evoLDxcUFETXrl3x8/PTKT9y5AiZmZk65XXq1KFy5cpER0cDEB0dTYMGDbRJHsDf35/U1FS9ngVm8ETfq1cvRo4cyfr167l+/TrXr19n3bp1BAYG0rt3b0OHJ4QQWhERETg6OuosERERz62/evVqjh49mmudhIQELC0tcXJy0il3c3MjISFBW+fvST57ffa6vDL4IxAiIyMJCQlh8ODB2ikFy5QpQ2BgIHPnzjVwdEIIY1OQUTdhYWE5Rg2q1epc6167do3x48cTFRWFlZVVvo9ZGAye6G1sbPj666+ZO3culy5dAqB69eo5ZpcSQohCUYCLsWq1+rmJ/VlHjhzh1q1bNG7cWFuWlZXFvn37+Oqrr9i2bRsZGRkkJyfrtOoTExNxd3cHwN3dnT/++ENnv9mjcrLr5IXBu26y2dra0rBhQxo2bChJXghRZBRN/hd9dOzYkZMnTxIbG6tdmjZtypAhQ7T/t7CwYOfO/92Ede7cOeLi4vDx8QHAx8eHkydPaufpAIiKisLBwQFvb+88x2LwFr0QQhSrYhpeaW9vT/369XXKbG1tKVeunLY8MDCQ4OBgnJ2dcXBwYOzYsfj4+NCiRQsAOnXqhLe3N8OGDWPOnDkkJCQwbdo0goKC8vzNAiTRCyFMjL4t86I0f/58zMzM6NOnD+np6fj7+/P1119r15ubm7Nx40bee+89fHx8sLW1JSAggPDwcL2OY/Bx9EVBxtGbFhlHb1oKOo7+Tue2+d62/Ja9BTq2oUiLXghhWkpQi764SKIXQpiUktR1U1wk0QshTIokeiGEMHKS6IUQwtgpKkNHUOxKzA1TQgghioa06IUQJkW6boQQwsgpGtPrupFEL4QwKdKiF0III6eY4MVYSfRCCJNiii16GXUjhBBGTlr0QgiTIhdjhRDCyBnf83pfThK9EMKkSIteCCGMnCR6IYQwcqbYdSOjboQQwshJi14IYVKk60YIIYyc3BkrhBBGzhTvjJVEL4QwKRpp0QshhHEzxa6bfI26+e233xg6dCg+Pj7cuHEDgBUrVrB///5CDU4IIUTB6Z3o161bh7+/P9bW1hw7doz09HQAUlJSmD17dqEHKIQQhUnRqPK9lFZ6J/qZM2cSGRnJN998g4WFhbbc19eXo0ePFmpwQghR2BQl/0tppXcf/blz52jTpk2OckdHR5KTkwsjJiGEKDKluWWeX3q36N3d3bl48WKO8v3791OtWrVCCUoIIYqKRlHleymt9E70o0ePZvz48cTExKBSqbh58yYrV64kJCSE9957ryhiFEIIUQB6d928//77aDQaOnbsyMOHD2nTpg1qtZqQkBDGjh1bFDEKIUShMcXhlSpFyd8lhoyMDC5evEhaWhre3t7Y2dkVdmz5lnnnL0OHIIqRtWdrQ4cgitGTjBsF2v5E1e753rbhlV8LdGxDyfcNU5aWlnh7exdmLEIIUeRKc197fumd6Nu3b49K9fw3ateuXQUKSAghipIpdt3onegbNWqk8zozM5PY2FhOnTpFQEBAYcUlhBBFojSPh88vvRP9/Pnzcy2fPn06aWlpBQ5ICCFE4Sq0GaaGDh3Kd999V1i7E0KIImGK4+gL7emV0dHRWFlZFdbuCmRT/WmGDkEUIxcbR0OHIEoR6aPPg969e+u8VhSF+Ph4Dh8+zIcfflhogQkhRFEozS3z/NI70Ts66raezMzMqF27NuHh4XTq1KnQAhNCiKJggtdi9Uv0WVlZjBw5kgYNGlC2bNmiikkIIYqMKbbo9boYa25uTqdOneQplUIIUYroPeqmfv36/PWXPGJACFE6KYoq30tpla+JR0JCQti4cSPx8fGkpqbqLEIIUZJpCrCUVnlO9OHh4Tx48IAuXbpw/Phx3nzzTSpWrEjZsmUpW7YsTk5O0m8vhCjxFFT5XvSxaNEiGjZsiIODAw4ODvj4+LBlyxbt+sePHxMUFES5cuWws7OjT58+JCYm6uwjLi6Orl27YmNjg6urK5MnT+bJkyd6n3OeL8bOmDGDd999l927d+t9ECGEKCk0xTTspmLFinz66afUrFkTRVFYvnw5PXr04NixY9SrV4+JEyeyadMmfvzxRxwdHRkzZgy9e/fmwIEDwNPBL127dsXd3Z2DBw8SHx/P8OHDsbCw0Ht+7jw/ptjMzIyEhARcXV31P+NitsF9sKFDEMXovcexhg5BFKP45D8LtP0ut/753rZD4poCHdvZ2Zm5c+fSt29fXFxcWLVqFX379gXg7Nmz1K1bl+joaFq0aMGWLVvo1q0bN2/exM3NDYDIyEhCQ0O5ffs2lpaWeT6uXn30L3pqpRBCGLv09PQc1yXT09Nful1WVharV6/mwYMH+Pj4cOTIETIzM/Hz89PWqVOnDpUrVyY6Ohp4+rSBBg0aaJM8gL+/P6mpqZw+fVqvuPVK9LVq1cLZ2fmFixBClGQF6aOPiIjA0dFRZ4mIiHjusU6ePImdnR1qtZp3332Xn376CW9vbxISErC0tMTJyUmnvpubGwkJCQAkJCToJPns9dnr9KHXDVMzZszIcWesEEKUJgUZPRMWFkZwcLBOmVqtfm792rVrExsbS0pKCmvXriUgIIC9e/cWIIL80SvRDxw4sFT00QshxPPoO3rm79Rq9QsT+7MsLS2pUaMGAE2aNOHQoUMsXLiQAQMGkJGRQXJysk6rPjExEXd3dwDc3d35448/dPaXPSonu05e5bnrRvrnhRDGwJDj6DUaDenp6TRp0gQLCwt27typXXfu3Dni4uLw8fEBwMfHh5MnT3Lr1i1tnaioKBwcHPSexjXPLfp8ziEuhBAlSnHd+BQWFkbnzp2pXLky9+/fZ9WqVezZs4dt27bh6OhIYGAgwcHBODs74+DgwNixY/Hx8aFFixYAdOrUCW9vb4YNG8acOXNISEhg2rRpBAUF6fWtAvRI9BpNab4vTAghitetW7cYPnw48fHxODo60rBhQ7Zt28brr78OPJ2tz8zMjD59+pCeno6/vz9ff/21dntzc3M2btzIe++9h4+PD7a2tgQEBBAeHq53LHkeR1+ayDh60yLj6E1LQcfRb3IblO9tuyb+t0DHNpRCm2FKCCFKA40JXm6URC+EMCmaAoy6Ka0k0QshTIrR9VXngSR6IYRJMcVhJXo/j14IIUTpIi16IYRJ0ZjgzZ+S6IUQJkX66IUQwsiZYh+9JHohhEmRcfRCCGHkTHEcvYy6EUIIIycteiGESZGLsUIIYeSkj14IIYycjLoRQggjJ103Qghh5Eyx60ZG3QghhJGTFr0QwqRIH70QQhg5SfRCCGHkFBPso5dEL4QwKdKiF0III2eKiV5G3QghhJGTFr0QwqTIDVNCCGHkTPGGKUn0QgiTYop99JLohRAmRRK9EEIYOVPso5dRN0IIYeSkRS+EMClyMVYIIYyc9NELIYSRM8U+ekn0QgiTojHBVC+JXghhUkyx60ZG3QghhJGTFr0QwqSYXseNJHohhIkxxa4bSfRCCJMi4+iL0RdffJHnuuPGjSvCSIQQpkRG3RSj+fPn56meSqWSRC+EKDSml+YNmOgvX75sqEMLIYRJkT56IYRJkYuxBnT9+nV++eUX4uLiyMjI0Fk3b948A0UlhDA2pthHXyJumNq5cye1a9dm0aJFfP755+zevZulS5fy3XffERsba+jwhBBGRCnAoo+IiAiaNWuGvb09rq6u9OzZk3PnzunUefz4MUFBQZQrVw47Ozv69OlDYmKiTp24uDi6du2KjY0Nrq6uTJ48mSdPnugVS4lI9GFhYYSEhHDy5EmsrKxYt24d165do23btvTr18/Q4QkhjIimAIs+9u7dS1BQEL///jtRUVFkZmbSqVMnHjx4oK0zceJEfv31V3788Uf27t3LzZs36d27t3Z9VlYWXbt2JSMjg4MHD7J8+XKWLVvGRx99pFcsKkVRDP49xt7entjYWKpXr07ZsmXZv38/9erV4/jx4/To0YMrV67otb8N7oOLJlBRIr33ONbQIYhiFJ/8Z4G2D646MN/bzruyOt/b3r59G1dXV/bu3UubNm1ISUnBxcWFVatW0bdvXwDOnj1L3bp1iY6OpkWLFmzZsoVu3bpx8+ZN3NzcAIiMjCQ0NJTbt29jaWmZp2OXiBa9ra2ttl/ew8ODS5cuadfduXPHUGEJIYSO9PR0UlNTdZb09PQ8bZuSkgKAs7MzAEeOHCEzMxM/Pz9tnTp16lC5cmWio6MBiI6OpkGDBtokD+Dv709qaiqnT5/Oc9wlItG3aNGC/fv3A9ClSxcmTZrErFmzeOutt2jRooWBoxNCGJOC9NFHRETg6Oios0RERLz0mBqNhgkTJuDr60v9+vUBSEhIwNLSEicnJ526bm5uJCQkaOv8Pclnr89el1clYtTNvHnzSEtLA2DGjBmkpaXxww8/ULNmTRlxI4QoVAUZXhkWFkZwcLBOmVqtful2QUFBnDp1StugLW4GT/RZWVlcv36dhg0bAk+7cSIjIw0clRDCWCkFGF6pVqvzlNj/bsyYMWzcuJF9+/ZRsWJFbbm7uzsZGRkkJyfrtOoTExNxd3fX1vnjjz909pc9Kie7Tl4YvOvG3NycTp06ce/ePUOHIoQwAcU16kZRFMaMGcNPP/3Erl278PLy0lnfpEkTLCws2Llzp7bs3LlzxMXF4ePjA4CPjw8nT57k1q1b2jpRUVE4ODjg7e2d51gM3qIHqF+/Pn/99VeON0IIIUqroKAgVq1axc8//4y9vb22T93R0RFra2scHR0JDAwkODgYZ2dnHBwcGDt2LD4+Ptprk506dcLb25thw4YxZ84cEhISmDZtGkFBQXp9szB4ix5g5syZhISEsHHjRuLj43Nc1RZCiMKiQcn3oo9FixaRkpJCu3bt8PDw0C4//PCDts78+fPp1q0bffr0oU2bNri7u7N+/XrtenNzczZu3Ii5uTk+Pj4MHTqU4cOHEx4erlcsJWIcvZnZ//7eqFT/e1i0oiioVCqysrL02p+xjKOvOfZNPLs2w66GJ5rHGSQdusDpmf8l7VK8to7axZH6Hw3GpW0DythZkXYxnvMLN3Bz0yEAbCqVp/bEXpRvVQ8rFyceJ97j2rr9nFuwASVTv/e1pDLmcfR/nIiiUuUKOcqXfrOKDybPZGhAP3r160qDht7YO9hRu0pzUlPuGyDS4lPQcfTvVe2f720XXVlToGMbSonoutm9e7ehQyiRyvvU5fLSKO7FXkJlbo73BwNo+cP77GwzhayHT8fuNvnyPSwcbfk94HMy7t6nYu+WNFsynj3+U0k5dRW7Gp5gZkbs5G95cDkRhzoVafT5aMxt1JyescrAZyhepnP7/piZm2tf16lbkzU/f8uvP28DwNrGit079rN7x36mTg9+3m7E35jis25KRKL38vKiUqVKOq15eNqiv3btmoGiMrzowf/UeX10fCRdTi/GqaEXd38/C4Bzs1ocD/2O5GNPbzI7v2ADNd7ujFNDL1JOXeXW7hPc2n1Cu4+Hcbe4uGgTXgF+kuhLgbt3dQcpjJ04ist/xRG9/+k3tm8WrQDAp1WzYo+ttDLFp1eWiD56Ly8vbt++naM8KSlJLtD+jYW9DQAZyWnasqRD56nQowUWTragUlGhhw9mVhbcOXjmBfux1tmHKB0sLCzo0787q/+z/uWVxXMpBfhXWpWIFn12X/yz0tLSsLKyMkBEJZBKRYNPhnE35hz3z17XFh96+wuaLh5H17PfoMl8QtajDGJGzufBlcRcd2Nb1Y1qgf6cmrGyuCIXheSNrh1xcLTnh1U/GToUUcoYNNFn32GmUqn48MMPsbGx0a7LysoiJiaGRo0avXAf6enpOZ41kalkYaEyf84WpdMrn47EoU4l9r05Q6e8bmg/LBxt2N93FhlJ9/Ho3JTXlozjtx7hpJ7V7fayci+Lz39DufFrDFdXynWR0mbwsN7s2vEbiQk5v/2KvDPFrhuDJvpjx44BT1v0J0+e1HkSm6WlJa+88gohISEv3EdERAQzZugmvwG29Rlo16DwAzaQhrNH4Ob3Kvt7hfM4PklbblPFlWqB/uxsO5n7524AkPpnHOWa18Fr5OscD/1OW9fKzYlW66aRdPgCsSH/LvZzEAVTsZInrdv5EDhsvKFDKfVKcxdMfhk00WePthk5ciQLFy7EwcFB733k9uyJbTVHF0p8JUHD2SPw6NyU/b1n8jBOtyVXxvrpDROKRvcHV8nSoPrbkFUr97K0WjeN5BOXOTo+Egw/olboacCQXty5ncSObXsNHUqpJy16A1m6dGm+t83t2RPG0m3T8NORVOrVkt9HfM6TtEeoXRwByLz/EM3jTO5fvEnaXwk0mhPIqfBVZCTdx7NzU1zb1uf3YZ8B/5/k13/Iw+t3ODVjJepy//tjmn47xSDnJfSjUqkYOKQXa/67Icc9JS6u5XF1K4+XV2UA6nrXIi3tATeuxZOcLJ9vbjQm2NApEYm+Q4cOL1y/a9euYoqkZKk24nUAWv+kO5vM0fGRxP2wD+VJFtFD5lBv6kBafB9CGVs1Dy4ncnRcJIk7YwFwbdsAu2ru2FVz543Yf+nsx1huLDN2bdr5ULGSZ66jbYa/NYCQ94O0rzdseTrccvw/PmDNqg3FFWKpYnppvoTcGTtx4kSd15mZmcTGxnLq1CkCAgJYuHChXvuTBGZajPnOWJFTQe+MHVql98srPcd/rpbOoa0lokU/f/78XMunT5+ufU69EEIUBlO8M7ZE3DD1PEOHDuW77757eUUhhMgjuWGqhImOjpYbpoQQhUpG3RhI7966fWaKohAfH8/hw4f58MMPDRSVEMIYmWLXTYlI9I6OjjqvzczMqF27NuHh4XTq1MlAUQkhjFFp7oLJrxKR6Asyjl4IIcSLlZiLscnJyfz73/8mLCyMpKSnt/kfPXqUGzduGDgyIYQxKa45Y0uSEtGiP3HiBB07dsTJyYkrV64wevRonJ2dWb9+PXFxcXz//feGDlEIYSRKwK1Dxa5EtOiDg4MZOXIkFy5c0Bll06VLF/bt22fAyIQQxqa45owtSUpEi/7QoUMsXrw4R3mFChW0M6cLIURhKM1dMPlVIhK9Wq0mNTU1R/n58+dxcXExQERCCGNliqNuSkTXzZtvvkl4eDiZmZnA06f1xcXFERoaSp8+fQwcnRBClG4lItF//vnnpKWl4erqyqNHj2jbti01atTAzs6OWbNmGTo8IYQRkT56A3F0dCQqKooDBw5w/Phx0tLSaNy4MX5+foYOTQhhZExx1E2JSPQAO3fuZOfOndy6dQuNRsPZs2dZtWoVgDzYTAhRaORirIHMmDGD8PBwmjZtioeHByqVytAhCSGMlClejC0RiT4yMpJly5YxbNgwQ4cihDBypbmvPb9KxMXYjIwMWrZsaegwhBDCKJWIRD9q1Chtf7wQQhQlRVHyvZRWJaLr5vHjxyxZsoQdO3bQsGFDLCwsdNbPmzfPQJEJIYyNKXbdlIhEf+LECRo1agTAqVOndNbJhVkhRGGSi7EGsnv3bkOHIIQwEZpS3AWTXyUi0QshRHExvTRfQi7GCiGEKDrSohdCmBS5GCuEEEZOEr0QQhi50jwePr8k0QshTIq06IUQwsiZ4jh6GXUjhBBGTlr0QgiTIn30Qghh5KSPXgghjJwptuilj14IYVKKa3Lwffv20b17dzw9PVGpVGzYsEFnvaIofPTRR3h4eGBtbY2fnx8XLlzQqZOUlMSQIUNwcHDAycmJwMBA0tLS9D5nSfRCCJOiFOCfPh48eMArr7zCv/71r1zXz5kzhy+++ILIyEhiYmKwtbXF39+fx48fa+sMGTKE06dPExUVxcaNG9m3bx9vv/223uesUozwe8wG98GGDkEUo/cexxo6BFGM4pP/LND2Dd198r3tiYTofG2nUqn46aef6NmzJ/C0Ne/p6cmkSZMICQkBICUlBTc3N5YtW8bAgQM5c+YM3t7eHDp0iKZNmwKwdetWunTpwvXr1/H09Mzz8aVFL4QwKRpFyfeSnp5OamqqzpKenq53DJcvXyYhIQE/Pz9tmaOjI82bNyc6+ukfk+joaJycnLRJHsDPzw8zMzNiYmL0Op4keiGESSlI101ERASOjo46S0REhN4xJCQkAODm5qZT7ubmpl2XkJCAq6urzvoyZcrg7OysrZNXMupGCGFSCjLxSFhYGMHBwTplarW6oCEVOUn0QgiTUpBHIKjV6kJJ7O7u7gAkJibi4eGhLU9MTNROq+ru7s6tW7d0tnvy5AlJSUna7fNKum6EECalIH30hcXLywt3d3d27typLUtNTSUmJgYfn6cXi318fEhOTubIkSPaOrt27UKj0dC8eXO9jicteiGEKAJpaWlcvHhR+/ry5cvExsbi7OxM5cqVmTBhAjNnzqRmzZp4eXnx4Ycf4unpqR2ZU7duXd544w1Gjx5NZGQkmZmZjBkzhoEDB+o14gYk0QshTExxPb3y8OHDtG/fXvs6u28/ICCAZcuWMWXKFB48eMDbb79NcnIyrVq1YuvWrVhZWWm3WblyJWPGjKFjx46YmZnRp08fvvjiC71jkXH0otSTcfSmpaDj6KuXb5zvbS/dOVqgYxuKtOiFECbFFJ9HL4leCGFSFEVj6BCKnSR6IYRJMcXHFMvwSiGEMHLSohdCmBQjHH/yUpLohRAmxRS7biTRCyFMirTohRDCyBXmowxKC0n0QgiTYorj6GXUjRBCGDlp0QshTIr00QshhJGTUTdCCGHkpEUvhBBGTkbdCCGEkTPFFr2MuhFCCCMnLXohhEmRi7FCCGHkTLHrRhK9EMKkyMVYIYQwcvIIBCGEEEZHWvRCCJMiXTdCCGHk5GKsEEIYOVPso5dEL4QwKdKiF0III2eKiV5G3QghhJGTFr0QwqSYXnseVIopfo8xQunp6URERBAWFoZarTZ0OKKIyect9CGJ3kikpqbi6OhISkoKDg4Ohg5HFDH5vIU+pI9eCCGMnCR6IYQwcpLohRDCyEmiNxJqtZqPP/5YLsyZCPm8hT7kYqwQQhg5adELIYSRk0QvhBBGThK9EEIYOUn0RaRdu3ZMmDDB0GEUO1M9b2NStWpVFixYYOgwRCGSRC9EKSd/XMXLSKIvpTIyMgwdgihFFEXhyZMnhg5DGIgk+kLw4MEDhg8fjp2dHR4eHnz++ec669PT0wkJCaFChQrY2trSvHlz9uzZo12/bNkynJyc2LBhAzVr1sTKygp/f3+uXbumrTN9+nQaNWrEv//9b7y8vLCysgIgOTmZUaNG4eLigoODAx06dOD48ePa7Y4fP0779u2xt7fHwcGBJk2acPjwYQCuXr1K9+7dKVu2LLa2ttSrV4/Nmzdrtz116hSdO3fGzs4ONzc3hg0bxp07d/J83uJpa3vcuHFMmTIFZ2dn3N3dmT59unb9yz6/ESNG0LNnT519TpgwgXbt2mnX7927l4ULF6JSqVCpVFy5coU9e/agUqnYsmULTZo0Qa1Ws3//fi5dukSPHj1wc3PDzs6OZs2asWPHjmJ4J4QhSaIvBJMnT2bv3r38/PPPbN++nT179nD06FHt+jFjxhAdHc3q1as5ceIE/fr144033uDChQvaOg8fPmTWrFl8//33HDhwgOTkZAYOHKhznIsXL7Ju3TrWr19PbGwsAP369ePWrVts2bKFI0eO0LhxYzp27EhSUhIAQ4YMoWLFihw6dIgjR47w/vvvY2FhAUBQUBDp6ens27ePkydP8s9//hM7OzvgaQLq0KEDr776KocPH2br1q0kJibSv3//PJ+3eGr58uXY2toSExPDnDlzCA8PJyoqCnj55/cyCxcuxMfHh9GjRxMfH098fDyVKlXSrn///ff59NNPOXPmDA0bNiQtLY0uXbqwc+dOjh07xhtvvEH37t2Ji4srknMXJYQiCuT+/fuKpaWlsmbNGm3Z3bt3FWtra2X8+PHK1atXFXNzc+XGjRs623Xs2FEJCwtTFEVRli5dqgDK77//rl1/5swZBVBiYmIURVGUjz/+WLGwsFBu3bqlrfPbb78pDg4OyuPHj3X2Xb16dWXx4sWKoiiKvb29smzZslxjb9CggTJ9+vRc133yySdKp06ddMquXbumAMq5c+deet7iqbZt2yqtWrXSKWvWrJkSGhqap88vICBA6dGjh8768ePHK23bttU5xrPv+e7duxVA2bBhw0tjrFevnvLll19qX1epUkWZP3/+y09OlBoy8UgBXbp0iYyMDJo3b64tc3Z2pnbt2gCcPHmSrKwsatWqpbNdeno65cqV074uU6YMzZo1076uU6cOTk5OnDlzhtdeew2AKlWq4OLioq1z/Phx0tLSdPYD8OjRIy5dugRAcHAwo0aNYsWKFfj5+dGvXz+qV68OwLhx43jvvffYvn07fn5+9OnTh4YNG2r3vXv3bm0L/9lzfvTo0QvPW/xP9nuazcPDg1u3buXp8yuopk2b6rxOS0tj+vTpbNq0ifj4eJ48ecKjR4+kRW/kJNEXsbS0NMzNzTly5Ajm5uY663JLoi9ia2ubY98eHh46/f3ZnJycgKd9+4MHD2bTpk1s2bKFjz/+mNWrV9OrVy9GjRqFv78/mzZtYvv27URERPD5558zduxY0tLS6N69O//85z9z7NvDw4OLFy/qFbspy+4qy6ZSqdBoNHn6/MzMzHLMcZqZmZnnYz/7MxMSEkJUVBSfffYZNWrUwNramr59+8rFfSMnib6AqlevjoWFBTExMVSuXBmAe/fucf78edq2bcurr75KVlYWt27donXr1s/dz5MnTzh8+LC29X7u3DmSk5OpW7fuc7dp3LgxCQkJlClThqpVqz63Xq1atahVqxYTJ05k0KBBLF26lF69egFQqVIl3n33Xd59913CwsL45ptvGDt2LI0bN2bdunVUrVqVMmVy/pi87LzFy+Xl83NxceHUqVM6ZbGxsTp/PCwtLcnKysrTMQ8cOMCIESO0n39aWhpXrlzJV/yi9JCLsQVkZ2dHYGAgkydPZteuXZw6dYoRI0ZgZvb0ra1VqxZDhgxh+PDhrF+/nsuXL/PHH38QERHBpk2btPuxsLBg7NixxMTEcOTIEUaMGEGLFi20iT83fn5++Pj40LNnT7Zv386VK1c4ePAgU6dO5fDhwzx69IgxY8awZ88erl69yoEDBzh06JD2j8eECRPYtm0bly9f5ujRo+zevVu7LigoiKSkJAYNGsShQ4e4dOkS27ZtY+TIkWRlZb30vMXLvezzA+jQoQOHDx/m+++/58KFC3z88cc5En/VqlWJiYnhypUr3LlzB41G89xj1qxZU3sx//jx4wwePPiF9YVxkN/KQjB37lxat25N9+7d8fPzo1WrVjRp0kS7funSpQwfPpxJkyZRu3ZtevbsyaFDh7QtYQAbGxtCQ0MZPHgwvr6+2NnZ8cMPP7zwuCqVis2bN9OmTRtGjhxJrVq1GDhwIFevXsXNzQ1zc3Pu3r3L8OHDqVWrFv3796dz587MmDEDgKysLIKCgqhbty5vvPEGtWrV4uuvvwbA09OTAwcOkJWVRadOnWjQoAETJkzAyclJm8xfdt7ixV72+QH4+/vz4YcfMmXKFJo1a8b9+/cZPny4zn5CQkIwNzfH29sbFxeXF/a3z5s3j7Jly9KyZUu6d++Ov78/jRs3LtLzFIYnjykuAZYtW8aECRNITk42dChCCCMkLXohhDBykuiFEMLISdeNEEIYOWnRCyGEkZNEL4QQRk4SvRBCGDlJ9EIIYeQk0QshhJGTRC9KlWcn4jDUNHrZE3vITW6iNJBELwrFiBEjtDMcWVpaUqNGDcLDw4t8+rr169fzySef5KmuJGdhquTplaLQvPHGGyxdupT09HQ2b95MUFAQFhYWhIWF6dTLyMjA0tKyUI7p7OxcKPsRwphJi14UGrVajbu7O1WqVOG9997Dz8+PX375RdvdMmvWLDw9PbWTk1y7do3+/fvj5OSEs7MzPXr00HlkblZWFsHBwTg5OVGuXDmmTJmS49nsz3bdpKenExoaSqVKlVCr1dSoUYNvv/2WK1eu0L59ewDKli2LSqVixIgRAGg0GiIiIvDy8sLa2ppXXnmFtWvX6hxn8+bN1KpVC2tra9q3by+P9hWliiR6UWSsra21E1rs3LmTc+fOERUVxcaNG8nMzMTf3x97e3t+++03Dhw4gJ2dHW+88YZ2m88//5xly5bx3XffsX//fpKSkvjpp59eeMzhw4fz3//+ly+++IIzZ86wePFi7OzsqFSpEuvWrQOePus/Pj6ehQsXAhAREcH3339PZGQkp0+fZuLEiQwdOpS9e/cCT/8g9e7dm+7duxMbG8uoUaN4//33i+ptE6LwGXIeQ2E8/j63qUajUaKiohS1Wq2EhIQoAQEBipubm5Kenq6tv2LFCqV27dqKRqPRlqWnpyvW1tbKtm3bFEVRFA8PD2XOnDna9ZmZmUrFihV15lD9+3yp586dUwAlKioq1xiz51G9d++etuzx48eKjY2NcvDgQZ26gYGByqBBgxRFUZSwsDDF29tbZ31oaGiOfQlRUkkfvSg0GzduxM7OjszMTDQaDYMHD2b69OkEBQXRoEEDnX7548ePc/HiRezt7XX28fjxYy5dukRKSgrx8fE6c9KWKVOGpk2b5ui+yRYbG4u5ubleM1xdvHiRhw8f8vrrr+uUZ2Rk8OqrrwJw5swZnTgAfHx88nwMIQxNEr0oNO3bt2fRokVYWlri6empMwVhbvPdNmnShJUrV+bYz98nQNeHtbW13tukpaUBsGnTJipUqKCzTq1W5ysOIUoaSfSi0Nja2lKjRo081W3cuDE//PADrq6uODg45FrHw8ODmJgY2rRpAzydV/fIkSPPnRGpQYMGaDQa9u7di5+fX4712d8o/j6/qre3N2q1mri4uOd+E6hbty6//PKLTtnvv//+8pMUooSQi7HCIIYMGUL58uXp0aMHv/32G5cvX2bPnj2MGzeO69evAzB+/Hg+/fRTNmzYwNmzZ/nHP/7xwjHwVatWJSAggLfeeosNGzZo97lmzRoAqlSpgkqlYuPGjdy+fZu0tDTs7e0JCQlh4sSJLF++nEuXLnH06FG+/PJLli9fDsC7777LhQsXmDx5MufOnWPVqlUsW7asqN8iIQqNJHphEDY2Nuzbt4/KlSvTu3dv6tatS2BgII8fP9a28CdNmsSwYcMICAjAx8cHe3t7evXq9cL9Llq0iL59+/KPf/yDOnXqMHr0aB48eABAhQoVmDFjBu+//z5ubm6MGTMGgE8++YQPP/yQiIgI7fy5mzZtwsvLC4DKlSuzbt06NmzYwCuvvEJkZCSzZ88uwndHiMIlE48IIYSRkxa9EEIYOUn0Qghh5CTRCyGEkZNEL4QQRk4SvRBCGDlJ9EIIYeQk0QshhJGTRC+EEEZOEr0QQhg5SfRCCGHkJNELIYSR+z/haQYtA1wXxgAAAABJRU5ErkJggg==\n"761          },762          "metadata": {}763        }764      ],765      "source": [766        "from sklearn.metrics import classification_report, confusion_matrix\n",767        "import numpy as np\n",768        "import seaborn as sns\n",769        "import matplotlib.pyplot as plt\n",770        "\n",771        "# probabilitas di test set\n",772        "y_test_true = np.array(test_labels)\n",773        "y_test_prob = model.predict(test_ds).ravel()\n",774        "\n",775        "t_acc = 0.56   # threshold terbaik untuk akurasi (dari valid)\n",776        "t_dep = 0.70   # threshold terbaik untuk recall depressed\n",777        "\n",778        "# 1. Threshold untuk akurasi\n",779        "y_test_pred_acc = (y_test_prob >= t_acc).astype(int)\n",780        "print(\"=== Hasil dengan threshold akurasi (0.56) ===\")\n",781        "print(classification_report(y_test_true, y_test_pred_acc, target_names=class_names))\n",782        "\n",783        "cm_acc = confusion_matrix(y_test_true, y_test_pred_acc)\n",784        "plt.figure(figsize=(4,4))\n",785        "sns.heatmap(cm_acc, annot=True, fmt=\"d\", xticklabels=class_names, yticklabels=class_names)\n",786        "plt.xlabel(\"Predicted\")\n",787        "plt.ylabel(\"True\")\n",788        "plt.title(\"Confusion Matrix - threshold 0.56\")\n",789        "plt.show()\n",790        "\n",791        "# 2. Threshold untuk recall depressed\n",792        "y_test_pred_dep = (y_test_prob >= t_dep).astype(int)\n",793        "print(\"\\n=== Hasil dengan threshold recall depressed (0.70) ===\")\n",794        "print(classification_report(y_test_true, y_test_pred_dep, target_names=class_names))\n",795        "\n",796        "cm_dep = confusion_matrix(y_test_true, y_test_pred_dep)\n",797        "plt.figure(figsize=(4,4))\n",798        "sns.heatmap(cm_dep, annot=True, fmt=\"d\", xticklabels=class_names, yticklabels=class_names)\n",799        "plt.xlabel(\"Predicted\")\n",800        "plt.ylabel(\"True\")\n",801        "plt.title(\"Confusion Matrix - threshold 0.70\")\n",802        "plt.show()\n"803      ]804    },805    {806      "cell_type": "code",807      "execution_count": null,808      "metadata": {809        "id": "2W2QBqVvTAYr",810        "colab": {811          "base_uri": "https://localhost:8080/"812        },813        "outputId": "0aa849b1-c0e2-448f-d3c6-eb1c5818d74f"814      },815      "outputs": [816        {817          "output_type": "stream",818          "name": "stdout",819          "text": [820            "\u001b[1m32/32\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 40ms/step\n"821          ]822        },823        {824          "output_type": "display_data",825          "data": {826            "text/plain": [827              "<Figure size 700x500 with 2 Axes>"828            ],829            "image/png": "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\n"830          },831          "metadata": {}832        }833      ],834      "source": [835        "y_true = np.array(test_labels)\n",836        "\n",837        "y_prob = model.predict(test_ds).ravel()\n",838        "y_pred = (y_prob >= 0.5).astype(int)\n",839        "\n",840        "cm = confusion_matrix(y_true, y_pred)\n",841        "plt.figure(figsize=(7, 5))\n",842        "sns.heatmap(cm, annot=True, fmt='d', xticklabels=class_names, yticklabels=class_names, cmap='Blues')\n",843        "plt.xlabel(\"Predicted\")\n",844        "plt.ylabel(\"True\")\n",845        "plt.title(\"Confusion Matrix\")\n",846        "plt.show()"847      ]848    }849  ],850  "metadata": {851    "colab": {852      "provenance": [],853      "gpuType": "T4"854    },855    "kernelspec": {856      "display_name": "Python 3",857      "name": "python3"858    },859    "language_info": {860      "name": "python"861    },862    "accelerator": "GPU"863  },864  "nbformat": 4,865  "nbformat_minor": 0866}