DRL_PROJ/classifier/notebooks/06_model_families.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 06 - Phase 3 Model-Family Analysis\n",
    "\n",
    "Phase 3 starts after the preprocessing decision was already made. Phase 2 selected the strongest supported setting available at that point: 224x224 input, facecrop enabled, ImageNet/default normalization, and no augmentation. Phase 3 keeps that data pipeline fixed and changes the pretrained backbone.\n",
    "\n",
    "The question is simple: once the input is already focused on faces, does a stronger or more modern pretrained architecture improve the detector beyond ResNet18?\n",
    "\n",
    "This notebook is evidence-only. It reads existing configs and logs, uses the saved log schema as the source of truth, and does not train, reevaluate, or invent missing results.\n",
    "\n",
    "Roadmap link: `01_eda` -> `02_preprocessing` -> `03_baselines` -> `04_ablation_questions` -> `05_gradcam` -> this model-family comparison -> `07_data_scaling`.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "7d59dfc4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Project root: c:\\Users\\diogo\\Documents\\MIA_UP\\2_Semestre\\DRL\\DRL_2\\DRL_PROJ\n"
     ]
    }
   ],
   "source": [
    "from __future__ import annotations\n",
    "\n",
    "import json\n",
    "import sys\n",
    "from pathlib import Path\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "\n",
    "def find_project_root(start: Path | None = None) -> Path:\n",
    "    \"\"\"Find DRL_PROJ whether the notebook runs from repo root or classifier/notebooks.\"\"\"\n",
    "    start = Path.cwd() if start is None else Path(start)\n",
    "    for candidate in [start, *start.parents]:\n",
    "        if (candidate / \"classifier\").is_dir() and (candidate / \"docs\" / \"DRL_Project.md\").exists():\n",
    "            return candidate\n",
    "    raise RuntimeError(\"Could not find DRL_PROJ root. Run this notebook from inside the repository.\")\n",
    "\n",
    "\n",
    "PROJECT_ROOT = find_project_root()\n",
    "CLASSIFIER_ROOT = PROJECT_ROOT / \"classifier\"\n",
    "if str(CLASSIFIER_ROOT) not in sys.path:\n",
    "    sys.path.insert(0, str(CLASSIFIER_ROOT))\n",
    "\n",
    "CONFIGS_DIR = CLASSIFIER_ROOT / \"configs\"\n",
    "LOGS_DIR = CLASSIFIER_ROOT / \"outputs\" / \"logs\"\n",
    "MODELS_DIR = CLASSIFIER_ROOT / \"outputs\" / \"models\"\n",
    "FIGURES_DIR = CLASSIFIER_ROOT / \"outputs\" / \"figures\"\n",
    "ANALYSIS_DIR = CLASSIFIER_ROOT / \"outputs\" / \"analysis\"\n",
    "FIGURES_DIR.mkdir(parents=True, exist_ok=True)\n",
    "ANALYSIS_DIR.mkdir(parents=True, exist_ok=True)\n",
    "\n",
    "print(f\"Project root: {PROJECT_ROOT}\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "14e87bcd",
   "metadata": {},
   "source": [
    "## 1. Load Phase 3 evidence\n",
    "\n",
    "The comparison uses the selected Phase 2 model as the reference point and then adds the five Phase 3 model-family runs. Every row below is loaded from existing logs and resolved configs.\n",
    "\n",
    "All Phase 3 runs use the same high-level setting: pretrained backbone, 224x224 input, facecropped classifier data, 20% subsample, no augmentation, and 5-fold evaluation. That keeps the story focused on architecture choice.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "e6b0417f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style type=\"text/css\">\n",
       "</style>\n",
       "<table id=\"T_28652\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th class=\"blank level0\" >&nbsp;</th>\n",
       "      <th id=\"T_28652_level0_col0\" class=\"col_heading level0 col0\" >label</th>\n",
       "      <th id=\"T_28652_level0_col1\" class=\"col_heading level0 col1\" >run</th>\n",
       "      <th id=\"T_28652_level0_col2\" class=\"col_heading level0 col2\" >backbone</th>\n",
       "      <th id=\"T_28652_level0_col3\" class=\"col_heading level0 col3\" >image_size</th>\n",
       "      <th id=\"T_28652_level0_col4\" class=\"col_heading level0 col4\" >data_dir</th>\n",
       "      <th id=\"T_28652_level0_col5\" class=\"col_heading level0 col5\" >subsample</th>\n",
       "      <th id=\"T_28652_level0_col6\" class=\"col_heading level0 col6\" >auc</th>\n",
       "      <th id=\"T_28652_level0_col7\" class=\"col_heading level0 col7\" >auc_std</th>\n",
       "      <th id=\"T_28652_level0_col8\" class=\"col_heading level0 col8\" >accuracy</th>\n",
       "      <th id=\"T_28652_level0_col9\" class=\"col_heading level0 col9\" >f1</th>\n",
       "      <th id=\"T_28652_level0_col10\" class=\"col_heading level0 col10\" >checkpoint_mb</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th id=\"T_28652_level0_row0\" class=\"row_heading level0 row0\" >5</th>\n",
       "      <td id=\"T_28652_row0_col0\" class=\"data row0 col0\" >ConvNeXt-Tiny</td>\n",
       "      <td id=\"T_28652_row0_col1\" class=\"data row0 col1\" >p3_convnext_tiny</td>\n",
       "      <td id=\"T_28652_row0_col2\" class=\"data row0 col2\" >convnext_tiny</td>\n",
       "      <td id=\"T_28652_row0_col3\" class=\"data row0 col3\" >224</td>\n",
       "      <td id=\"T_28652_row0_col4\" class=\"data row0 col4\" >cropped/classifier</td>\n",
       "      <td id=\"T_28652_row0_col5\" class=\"data row0 col5\" >0.200000</td>\n",
       "      <td id=\"T_28652_row0_col6\" class=\"data row0 col6\" >0.9868</td>\n",
       "      <td id=\"T_28652_row0_col7\" class=\"data row0 col7\" >0.0013</td>\n",
       "      <td id=\"T_28652_row0_col8\" class=\"data row0 col8\" >0.9473</td>\n",
       "      <td id=\"T_28652_row0_col9\" class=\"data row0 col9\" >0.9650</td>\n",
       "      <td id=\"T_28652_row0_col10\" class=\"data row0 col10\" >106.2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_28652_level0_row1\" class=\"row_heading level0 row1\" >2</th>\n",
       "      <td id=\"T_28652_row1_col0\" class=\"data row1 col0\" >ResNet50</td>\n",
       "      <td id=\"T_28652_row1_col1\" class=\"data row1 col1\" >p3_resnet50</td>\n",
       "      <td id=\"T_28652_row1_col2\" class=\"data row1 col2\" >resnet50</td>\n",
       "      <td id=\"T_28652_row1_col3\" class=\"data row1 col3\" >224</td>\n",
       "      <td id=\"T_28652_row1_col4\" class=\"data row1 col4\" >cropped/classifier</td>\n",
       "      <td id=\"T_28652_row1_col5\" class=\"data row1 col5\" >0.200000</td>\n",
       "      <td id=\"T_28652_row1_col6\" class=\"data row1 col6\" >0.9857</td>\n",
       "      <td id=\"T_28652_row1_col7\" class=\"data row1 col7\" >0.0011</td>\n",
       "      <td id=\"T_28652_row1_col8\" class=\"data row1 col8\" >0.9532</td>\n",
       "      <td id=\"T_28652_row1_col9\" class=\"data row1 col9\" >0.9688</td>\n",
       "      <td id=\"T_28652_row1_col10\" class=\"data row1 col10\" >90.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_28652_level0_row2\" class=\"row_heading level0 row2\" >3</th>\n",
       "      <td id=\"T_28652_row2_col0\" class=\"data row2 col0\" >EfficientNet-B0</td>\n",
       "      <td id=\"T_28652_row2_col1\" class=\"data row2 col1\" >p3_efficientnet_b0</td>\n",
       "      <td id=\"T_28652_row2_col2\" class=\"data row2 col2\" >efficientnet_b0</td>\n",
       "      <td id=\"T_28652_row2_col3\" class=\"data row2 col3\" >224</td>\n",
       "      <td id=\"T_28652_row2_col4\" class=\"data row2 col4\" >cropped/classifier</td>\n",
       "      <td id=\"T_28652_row2_col5\" class=\"data row2 col5\" >0.200000</td>\n",
       "      <td id=\"T_28652_row2_col6\" class=\"data row2 col6\" >0.9845</td>\n",
       "      <td id=\"T_28652_row2_col7\" class=\"data row2 col7\" >0.0025</td>\n",
       "      <td id=\"T_28652_row2_col8\" class=\"data row2 col8\" >0.9450</td>\n",
       "      <td id=\"T_28652_row2_col9\" class=\"data row2 col9\" >0.9628</td>\n",
       "      <td id=\"T_28652_row2_col10\" class=\"data row2 col10\" >15.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_28652_level0_row3\" class=\"row_heading level0 row3\" >1</th>\n",
       "      <td id=\"T_28652_row3_col0\" class=\"data row3 col0\" >ResNet34</td>\n",
       "      <td id=\"T_28652_row3_col1\" class=\"data row3 col1\" >p3_resnet34</td>\n",
       "      <td id=\"T_28652_row3_col2\" class=\"data row3 col2\" >resnet34</td>\n",
       "      <td id=\"T_28652_row3_col3\" class=\"data row3 col3\" >224</td>\n",
       "      <td id=\"T_28652_row3_col4\" class=\"data row3 col4\" >cropped/classifier</td>\n",
       "      <td id=\"T_28652_row3_col5\" class=\"data row3 col5\" >0.200000</td>\n",
       "      <td id=\"T_28652_row3_col6\" class=\"data row3 col6\" >0.9779</td>\n",
       "      <td id=\"T_28652_row3_col7\" class=\"data row3 col7\" >0.0034</td>\n",
       "      <td id=\"T_28652_row3_col8\" class=\"data row3 col8\" >0.9305</td>\n",
       "      <td id=\"T_28652_row3_col9\" class=\"data row3 col9\" >0.9529</td>\n",
       "      <td id=\"T_28652_row3_col10\" class=\"data row3 col10\" >81.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_28652_level0_row4\" class=\"row_heading level0 row4\" >0</th>\n",
       "      <td id=\"T_28652_row4_col0\" class=\"data row4 col0\" >ResNet18 facecrop reference</td>\n",
       "      <td id=\"T_28652_row4_col1\" class=\"data row4 col1\" >p2c_resnet18_facecrop</td>\n",
       "      <td id=\"T_28652_row4_col2\" class=\"data row4 col2\" >resnet18</td>\n",
       "      <td id=\"T_28652_row4_col3\" class=\"data row4 col3\" >224</td>\n",
       "      <td id=\"T_28652_row4_col4\" class=\"data row4 col4\" >data_cropped</td>\n",
       "      <td id=\"T_28652_row4_col5\" class=\"data row4 col5\" >0.200000</td>\n",
       "      <td id=\"T_28652_row4_col6\" class=\"data row4 col6\" >0.9755</td>\n",
       "      <td id=\"T_28652_row4_col7\" class=\"data row4 col7\" >0.0048</td>\n",
       "      <td id=\"T_28652_row4_col8\" class=\"data row4 col8\" >0.9223</td>\n",
       "      <td id=\"T_28652_row4_col9\" class=\"data row4 col9\" >0.9468</td>\n",
       "      <td id=\"T_28652_row4_col10\" class=\"data row4 col10\" >42.7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_28652_level0_row5\" class=\"row_heading level0 row5\" >4</th>\n",
       "      <td id=\"T_28652_row5_col0\" class=\"data row5 col0\" >MobileNetV3-Small</td>\n",
       "      <td id=\"T_28652_row5_col1\" class=\"data row5 col1\" >p3_mobilenetv3_small</td>\n",
       "      <td id=\"T_28652_row5_col2\" class=\"data row5 col2\" >mobilenet_v3_small</td>\n",
       "      <td id=\"T_28652_row5_col3\" class=\"data row5 col3\" >224</td>\n",
       "      <td id=\"T_28652_row5_col4\" class=\"data row5 col4\" >cropped/classifier</td>\n",
       "      <td id=\"T_28652_row5_col5\" class=\"data row5 col5\" >0.200000</td>\n",
       "      <td id=\"T_28652_row5_col6\" class=\"data row5 col6\" >0.9680</td>\n",
       "      <td id=\"T_28652_row5_col7\" class=\"data row5 col7\" >0.0032</td>\n",
       "      <td id=\"T_28652_row5_col8\" class=\"data row5 col8\" >0.9175</td>\n",
       "      <td id=\"T_28652_row5_col9\" class=\"data row5 col9\" >0.9443</td>\n",
       "      <td id=\"T_28652_row5_col10\" class=\"data row5 col10\" >5.9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x1a619bccec0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 900x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "PHASE2_REFERENCE = {\"p2c_resnet18_facecrop\": \"ResNet18 facecrop reference\"}\n",
    "PHASE3_RUNS = {\n",
    "    \"p3_resnet34\": \"ResNet34\",\n",
    "    \"p3_resnet50\": \"ResNet50\",\n",
    "    \"p3_efficientnet_b0\": \"EfficientNet-B0\",\n",
    "    \"p3_mobilenetv3_small\": \"MobileNetV3-Small\",\n",
    "    \"p3_convnext_tiny\": \"ConvNeXt-Tiny\",\n",
    "}\n",
    "RUN_LABELS = {**PHASE2_REFERENCE, **PHASE3_RUNS}\n",
    "\n",
    "\n",
    "def load_json(path: Path) -> dict:\n",
    "    return json.loads(path.read_text(encoding=\"utf-8\"))\n",
    "\n",
    "\n",
    "def resolve_config(path: Path) -> dict:\n",
    "    cfg = load_json(path)\n",
    "    parent = cfg.pop(\"extends\", None)\n",
    "    if parent:\n",
    "        base = resolve_config(path.parent / parent)\n",
    "        base.update(cfg)\n",
    "        cfg = base\n",
    "    return cfg\n",
    "\n",
    "\n",
    "def config_path_for_run(run_name: str) -> Path:\n",
    "    for phase in [\"phase2\", \"phase3\"]:\n",
    "        path = CONFIGS_DIR / phase / f\"{run_name}.json\"\n",
    "        if path.exists():\n",
    "            return path\n",
    "    raise FileNotFoundError(f\"No config found for {run_name}\")\n",
    "\n",
    "\n",
    "def load_run(run_name: str) -> dict:\n",
    "    path = LOGS_DIR / f\"{run_name}.json\"\n",
    "    if not path.exists():\n",
    "        raise FileNotFoundError(f\"Missing log for {run_name}: {path}\")\n",
    "    return load_json(path)\n",
    "\n",
    "\n",
    "def agg_metric(results: dict, metric: str, field: str = \"mean\"):\n",
    "    return results.get(\"aggregated_metrics\", {}).get(metric, {}).get(field, np.nan)\n",
    "\n",
    "\n",
    "def checkpoint_mb(run_name: str) -> float:\n",
    "    path = MODELS_DIR / f\"{run_name}_fold0_best.pt\"\n",
    "    return path.stat().st_size / (1024 * 1024) if path.exists() else np.nan\n",
    "\n",
    "\n",
    "def summarize_run(run_name: str, label: str) -> dict:\n",
    "    results = load_run(run_name)\n",
    "    cfg = {**resolve_config(config_path_for_run(run_name)), **results.get(\"config\", {})}\n",
    "    return {\n",
    "        \"label\": label,\n",
    "        \"run\": run_name,\n",
    "        \"backbone\": cfg.get(\"backbone\"),\n",
    "        \"image_size\": cfg.get(\"image_size\"),\n",
    "        \"data_dir\": cfg.get(\"data_dir\"),\n",
    "        \"subsample\": cfg.get(\"subsample\"),\n",
    "        \"augment\": cfg.get(\"augment\", False),\n",
    "        \"pretrained\": cfg.get(\"pretrained\", True),\n",
    "        \"auc\": agg_metric(results, \"auc_roc\"),\n",
    "        \"auc_std\": agg_metric(results, \"auc_roc\", \"std\"),\n",
    "        \"accuracy\": agg_metric(results, \"accuracy\"),\n",
    "        \"f1\": agg_metric(results, \"f1\"),\n",
    "        \"checkpoint_mb\": checkpoint_mb(run_name),\n",
    "    }\n",
    "\n",
    "\n",
    "summary_df = pd.DataFrame([summarize_run(run, label) for run, label in RUN_LABELS.items()])\n",
    "summary_df = summary_df.sort_values(\"auc\", ascending=False)\n",
    "display(\n",
    "    summary_df[[\"label\", \"run\", \"backbone\", \"image_size\", \"data_dir\", \"subsample\", \"auc\", \"auc_std\", \"accuracy\", \"f1\", \"checkpoint_mb\"]]\n",
    "    .style.format({\"auc\": \"{:.4f}\", \"auc_std\": \"{:.4f}\", \"accuracy\": \"{:.4f}\", \"f1\": \"{:.4f}\", \"checkpoint_mb\": \"{:.1f}\"})\n",
    ")\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(9, 4.8))\n",
    "plot_df = summary_df.sort_values(\"auc\")\n",
    "colors = [\"#4C78A8\" if run.startswith(\"p3_\") else \"#9A9A9A\" for run in plot_df[\"run\"]]\n",
    "ax.barh(plot_df[\"label\"], plot_df[\"auc\"], xerr=plot_df[\"auc_std\"], color=colors, alpha=0.9)\n",
    "ax.set_xlim(0.95, 1.0)\n",
    "ax.set_xlabel(\"Mean held-out AUC across folds\")\n",
    "ax.set_title(\"Phase 3 model families compared with the Phase 2 reference\")\n",
    "for y, (_, row) in enumerate(plot_df.iterrows()):\n",
    "    ax.text(row[\"auc\"] + 0.0007, y, f\"{row['auc']:.4f}\", va=\"center\", fontsize=8)\n",
    "fig.tight_layout()\n",
    "fig.savefig(FIGURES_DIR / \"06_phase3_ranked_auc.png\", dpi=180, bbox_inches=\"tight\")\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "95983d44",
   "metadata": {},
   "source": [
    "The architecture sweep improves on the Phase 2 reference. The strongest AUC is `p3_convnext_tiny` at `0.9868`, followed closely by `p3_resnet50` at `0.9857` and `p3_efficientnet_b0` at `0.9845`. ResNet50 has the best accuracy and F1 among these runs, while ConvNeXt-Tiny has the best mean AUC.\n",
    "\n",
    "MobileNetV3-Small is useful as the lightweight boundary case. It is much smaller, but it does not beat the selected ResNet18 facecrop reference on AUC. That suggests Phase 3 is not just about using any modern pretrained network; capacity and architecture still matter.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "24da3c90",
   "metadata": {},
   "source": [
    "## 2. Fold stability and training dynamics\n",
    "\n",
    "The next view checks whether the improvements are stable across folds and whether the training curves look plausible. This is where the decision should move beyond one headline AUC number. ConvNeXt-Tiny has the highest mean AUC, but the margin over ResNet50 is very small (`0.9868` vs `0.9857`). If the goal is to choose one detector, the better question is which model gives the strongest overall operating behavior across AUC, accuracy, F1, fold stability, and the error balance shown later.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "91d40ea5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 900x450 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x700 with 6 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def fold_metric_rows(run_name: str, label: str) -> list[dict]:\n",
    "    rows = []\n",
    "    results = load_run(run_name)\n",
    "    for fold in results.get(\"fold_results\", []):\n",
    "        tm = fold.get(\"test_metrics\", {})\n",
    "        rows.append({\n",
    "            \"run\": run_name,\n",
    "            \"label\": label,\n",
    "            \"fold\": fold.get(\"fold\"),\n",
    "            \"auc\": tm.get(\"auc_roc\", np.nan),\n",
    "            \"accuracy\": tm.get(\"accuracy\", np.nan),\n",
    "            \"f1\": tm.get(\"f1\", np.nan),\n",
    "        })\n",
    "    return rows\n",
    "\n",
    "\n",
    "fold_df = pd.DataFrame([row for run, label in RUN_LABELS.items() for row in fold_metric_rows(run, label)])\n",
    "order = summary_df.sort_values(\"auc\", ascending=False)[\"label\"].tolist()\n",
    "fig, ax = plt.subplots(figsize=(9, 4.5))\n",
    "for pos, label in enumerate(order):\n",
    "    vals = fold_df.loc[fold_df[\"label\"] == label, \"auc\"].to_numpy()\n",
    "    ax.scatter(np.full_like(vals, pos, dtype=float), vals, s=35, alpha=0.75)\n",
    "    ax.plot([pos - 0.18, pos + 0.18], [np.nanmean(vals), np.nanmean(vals)], color=\"black\", linewidth=2)\n",
    "ax.set_xticks(range(len(order)))\n",
    "ax.set_xticklabels(order, rotation=25, ha=\"right\")\n",
    "ax.set_ylabel(\"Fold held-out AUC\")\n",
    "ax.set_title(\"Fold-level AUC stability\")\n",
    "fig.tight_layout()\n",
    "fig.savefig(FIGURES_DIR / \"06_phase3_fold_auc_stability.png\", dpi=180, bbox_inches=\"tight\")\n",
    "plt.show()\n",
    "\n",
    "\n",
    "def history_frame(run_name: str, label: str) -> pd.DataFrame:\n",
    "    rows = []\n",
    "    results = load_run(run_name)\n",
    "    for fold in results.get(\"fold_results\", []):\n",
    "        hist = fold.get(\"history\", {})\n",
    "        n_epochs = max(len(hist.get(\"train_auc\", [])), len(hist.get(\"val_auc\", [])))\n",
    "        for epoch in range(n_epochs):\n",
    "            rows.append({\n",
    "                \"run\": run_name,\n",
    "                \"label\": label,\n",
    "                \"fold\": fold.get(\"fold\"),\n",
    "                \"epoch\": epoch + 1,\n",
    "                \"train_auc\": hist.get(\"train_auc\", [np.nan] * n_epochs)[epoch] if epoch < len(hist.get(\"train_auc\", [])) else np.nan,\n",
    "                \"val_auc\": hist.get(\"val_auc\", [np.nan] * n_epochs)[epoch] if epoch < len(hist.get(\"val_auc\", [])) else np.nan,\n",
    "            })\n",
    "    return pd.DataFrame(rows)\n",
    "\n",
    "\n",
    "history_df = pd.concat([history_frame(run, label) for run, label in RUN_LABELS.items()], ignore_index=True)\n",
    "fig, axes = plt.subplots(2, 3, figsize=(12, 7), sharey=True)\n",
    "for ax, label in zip(axes.ravel(), order):\n",
    "    sub = history_df[history_df[\"label\"] == label]\n",
    "    curve = sub.groupby(\"epoch\")[[\"train_auc\", \"val_auc\"]].mean()\n",
    "    ax.plot(curve.index, curve[\"train_auc\"], label=\"train AUC\")\n",
    "    ax.plot(curve.index, curve[\"val_auc\"], label=\"val AUC\")\n",
    "    ax.set_title(label, fontsize=10)\n",
    "    ax.set_xlabel(\"epoch\")\n",
    "    ax.set_ylim(0.85, 1.01)\n",
    "axes[0, 0].set_ylabel(\"AUC\")\n",
    "axes[1, 0].set_ylabel(\"AUC\")\n",
    "axes[0, 0].legend(fontsize=8)\n",
    "fig.suptitle(\"Mean training and validation AUC across folds\", fontsize=13)\n",
    "fig.tight_layout()\n",
    "fig.savefig(FIGURES_DIR / \"06_phase3_training_curves.png\", dpi=180, bbox_inches=\"tight\")\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9d401be6",
   "metadata": {},
   "source": [
    "The fold-level AUC stability plot shows three clear groups, ConvNeXt-Tiny is the highest and very stable, with fold AUCs around `0.985` to `0.988` and mean AUC `0.9868`. ResNet50 is almost tied, with fold AUCs around `0.984` to `0.987` and mean AUC `0.9857`. EfficientNet-B0 is also close at mean AUC `0.9845`, but its fold spread is a little wider.\n",
    "\n",
    "The next group is ResNet34 and the Phase 2 ResNet18 facecrop reference. ResNet34 improves the mean AUC to `0.9779`, while the ResNet18 reference is `0.9755`. MobileNetV3-Small is the weakest Phase 3 model at `0.9680`. So the fold plot supports the main architecture conclusion: the strongest Phase 3 backbones move the detector from the mid/high `0.97` AUC range into the mid/high `0.98` range, with ConvNeXt-Tiny and ResNet50 clearly ahead.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e033564f",
   "metadata": {},
   "source": [
    "## 3. Source-wise behavior\n",
    "\n",
    "Overall AUC can hide source-specific weaknesses, so Phase 3 keeps the same source diagnostics as earlier notebooks. For fake sources, `detection_rate` measures how often that fake source is classified as fake, and `pairwise_auc` measures wiki-vs-that-source discrimination. For wiki, the important value is false alarm rate.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "e79c74c7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style type=\"text/css\">\n",
       "</style>\n",
       "<table id=\"T_8207c\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th class=\"blank level0\" >&nbsp;</th>\n",
       "      <th id=\"T_8207c_level0_col0\" class=\"col_heading level0 col0\" >run</th>\n",
       "      <th id=\"T_8207c_level0_col1\" class=\"col_heading level0 col1\" >label</th>\n",
       "      <th id=\"T_8207c_level0_col2\" class=\"col_heading level0 col2\" >source</th>\n",
       "      <th id=\"T_8207c_level0_col3\" class=\"col_heading level0 col3\" >accuracy</th>\n",
       "      <th id=\"T_8207c_level0_col4\" class=\"col_heading level0 col4\" >detection_rate</th>\n",
       "      <th id=\"T_8207c_level0_col5\" class=\"col_heading level0 col5\" >false_alarm_rate</th>\n",
       "      <th id=\"T_8207c_level0_col6\" class=\"col_heading level0 col6\" >pairwise_auc</th>\n",
       "      <th id=\"T_8207c_level0_col7\" class=\"col_heading level0 col7\" >pairwise_f1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row0\" class=\"row_heading level0 row0\" >23</th>\n",
       "      <td id=\"T_8207c_row0_col0\" class=\"data row0 col0\" >p3_convnext_tiny</td>\n",
       "      <td id=\"T_8207c_row0_col1\" class=\"data row0 col1\" >ConvNeXt-Tiny</td>\n",
       "      <td id=\"T_8207c_row0_col2\" class=\"data row0 col2\" >wiki</td>\n",
       "      <td id=\"T_8207c_row0_col3\" class=\"data row0 col3\" >0.8838</td>\n",
       "      <td id=\"T_8207c_row0_col4\" class=\"data row0 col4\" ></td>\n",
       "      <td id=\"T_8207c_row0_col5\" class=\"data row0 col5\" >0.1162</td>\n",
       "      <td id=\"T_8207c_row0_col6\" class=\"data row0 col6\" ></td>\n",
       "      <td id=\"T_8207c_row0_col7\" class=\"data row0 col7\" ></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row1\" class=\"row_heading level0 row1\" >20</th>\n",
       "      <td id=\"T_8207c_row1_col0\" class=\"data row1 col0\" >p3_convnext_tiny</td>\n",
       "      <td id=\"T_8207c_row1_col1\" class=\"data row1 col1\" >ConvNeXt-Tiny</td>\n",
       "      <td id=\"T_8207c_row1_col2\" class=\"data row1 col2\" >inpainting</td>\n",
       "      <td id=\"T_8207c_row1_col3\" class=\"data row1 col3\" >0.9753</td>\n",
       "      <td id=\"T_8207c_row1_col4\" class=\"data row1 col4\" >0.9753</td>\n",
       "      <td id=\"T_8207c_row1_col5\" class=\"data row1 col5\" ></td>\n",
       "      <td id=\"T_8207c_row1_col6\" class=\"data row1 col6\" >0.9884</td>\n",
       "      <td id=\"T_8207c_row1_col7\" class=\"data row1 col7\" >0.9329</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row2\" class=\"row_heading level0 row2\" >21</th>\n",
       "      <td id=\"T_8207c_row2_col0\" class=\"data row2 col0\" >p3_convnext_tiny</td>\n",
       "      <td id=\"T_8207c_row2_col1\" class=\"data row2 col1\" >ConvNeXt-Tiny</td>\n",
       "      <td id=\"T_8207c_row2_col2\" class=\"data row2 col2\" >insight</td>\n",
       "      <td id=\"T_8207c_row2_col3\" class=\"data row2 col3\" >0.9452</td>\n",
       "      <td id=\"T_8207c_row2_col4\" class=\"data row2 col4\" >0.9452</td>\n",
       "      <td id=\"T_8207c_row2_col5\" class=\"data row2 col5\" ></td>\n",
       "      <td id=\"T_8207c_row2_col6\" class=\"data row2 col6\" >0.9791</td>\n",
       "      <td id=\"T_8207c_row2_col7\" class=\"data row2 col7\" >0.9172</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row3\" class=\"row_heading level0 row3\" >22</th>\n",
       "      <td id=\"T_8207c_row3_col0\" class=\"data row3 col0\" >p3_convnext_tiny</td>\n",
       "      <td id=\"T_8207c_row3_col1\" class=\"data row3 col1\" >ConvNeXt-Tiny</td>\n",
       "      <td id=\"T_8207c_row3_col2\" class=\"data row3 col2\" >text2img</td>\n",
       "      <td id=\"T_8207c_row3_col3\" class=\"data row3 col3\" >0.9850</td>\n",
       "      <td id=\"T_8207c_row3_col4\" class=\"data row3 col4\" >0.9850</td>\n",
       "      <td id=\"T_8207c_row3_col5\" class=\"data row3 col5\" ></td>\n",
       "      <td id=\"T_8207c_row3_col6\" class=\"data row3 col6\" >0.9927</td>\n",
       "      <td id=\"T_8207c_row3_col7\" class=\"data row3 col7\" >0.9378</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row4\" class=\"row_heading level0 row4\" >11</th>\n",
       "      <td id=\"T_8207c_row4_col0\" class=\"data row4 col0\" >p3_resnet50</td>\n",
       "      <td id=\"T_8207c_row4_col1\" class=\"data row4 col1\" >ResNet50</td>\n",
       "      <td id=\"T_8207c_row4_col2\" class=\"data row4 col2\" >wiki</td>\n",
       "      <td id=\"T_8207c_row4_col3\" class=\"data row4 col3\" >0.9092</td>\n",
       "      <td id=\"T_8207c_row4_col4\" class=\"data row4 col4\" ></td>\n",
       "      <td id=\"T_8207c_row4_col5\" class=\"data row4 col5\" >0.0908</td>\n",
       "      <td id=\"T_8207c_row4_col6\" class=\"data row4 col6\" ></td>\n",
       "      <td id=\"T_8207c_row4_col7\" class=\"data row4 col7\" ></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row5\" class=\"row_heading level0 row5\" >8</th>\n",
       "      <td id=\"T_8207c_row5_col0\" class=\"data row5 col0\" >p3_resnet50</td>\n",
       "      <td id=\"T_8207c_row5_col1\" class=\"data row5 col1\" >ResNet50</td>\n",
       "      <td id=\"T_8207c_row5_col2\" class=\"data row5 col2\" >inpainting</td>\n",
       "      <td id=\"T_8207c_row5_col3\" class=\"data row5 col3\" >0.9733</td>\n",
       "      <td id=\"T_8207c_row5_col4\" class=\"data row5 col4\" >0.9733</td>\n",
       "      <td id=\"T_8207c_row5_col5\" class=\"data row5 col5\" ></td>\n",
       "      <td id=\"T_8207c_row5_col6\" class=\"data row5 col6\" >0.9879</td>\n",
       "      <td id=\"T_8207c_row5_col7\" class=\"data row5 col7\" >0.9431</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row6\" class=\"row_heading level0 row6\" >9</th>\n",
       "      <td id=\"T_8207c_row6_col0\" class=\"data row6 col0\" >p3_resnet50</td>\n",
       "      <td id=\"T_8207c_row6_col1\" class=\"data row6 col1\" >ResNet50</td>\n",
       "      <td id=\"T_8207c_row6_col2\" class=\"data row6 col2\" >insight</td>\n",
       "      <td id=\"T_8207c_row6_col3\" class=\"data row6 col3\" >0.9443</td>\n",
       "      <td id=\"T_8207c_row6_col4\" class=\"data row6 col4\" >0.9443</td>\n",
       "      <td id=\"T_8207c_row6_col5\" class=\"data row6 col5\" ></td>\n",
       "      <td id=\"T_8207c_row6_col6\" class=\"data row6 col6\" >0.9792</td>\n",
       "      <td id=\"T_8207c_row6_col7\" class=\"data row6 col7\" >0.9280</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row7\" class=\"row_heading level0 row7\" >10</th>\n",
       "      <td id=\"T_8207c_row7_col0\" class=\"data row7 col0\" >p3_resnet50</td>\n",
       "      <td id=\"T_8207c_row7_col1\" class=\"data row7 col1\" >ResNet50</td>\n",
       "      <td id=\"T_8207c_row7_col2\" class=\"data row7 col2\" >text2img</td>\n",
       "      <td id=\"T_8207c_row7_col3\" class=\"data row7 col3\" >0.9860</td>\n",
       "      <td id=\"T_8207c_row7_col4\" class=\"data row7 col4\" >0.9860</td>\n",
       "      <td id=\"T_8207c_row7_col5\" class=\"data row7 col5\" ></td>\n",
       "      <td id=\"T_8207c_row7_col6\" class=\"data row7 col6\" >0.9900</td>\n",
       "      <td id=\"T_8207c_row7_col7\" class=\"data row7 col7\" >0.9495</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row8\" class=\"row_heading level0 row8\" >15</th>\n",
       "      <td id=\"T_8207c_row8_col0\" class=\"data row8 col0\" >p3_efficientnet_b0</td>\n",
       "      <td id=\"T_8207c_row8_col1\" class=\"data row8 col1\" >EfficientNet-B0</td>\n",
       "      <td id=\"T_8207c_row8_col2\" class=\"data row8 col2\" >wiki</td>\n",
       "      <td id=\"T_8207c_row8_col3\" class=\"data row8 col3\" >0.9292</td>\n",
       "      <td id=\"T_8207c_row8_col4\" class=\"data row8 col4\" ></td>\n",
       "      <td id=\"T_8207c_row8_col5\" class=\"data row8 col5\" >0.0708</td>\n",
       "      <td id=\"T_8207c_row8_col6\" class=\"data row8 col6\" ></td>\n",
       "      <td id=\"T_8207c_row8_col7\" class=\"data row8 col7\" ></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row9\" class=\"row_heading level0 row9\" >12</th>\n",
       "      <td id=\"T_8207c_row9_col0\" class=\"data row9 col0\" >p3_efficientnet_b0</td>\n",
       "      <td id=\"T_8207c_row9_col1\" class=\"data row9 col1\" >EfficientNet-B0</td>\n",
       "      <td id=\"T_8207c_row9_col2\" class=\"data row9 col2\" >inpainting</td>\n",
       "      <td id=\"T_8207c_row9_col3\" class=\"data row9 col3\" >0.9470</td>\n",
       "      <td id=\"T_8207c_row9_col4\" class=\"data row9 col4\" >0.9470</td>\n",
       "      <td id=\"T_8207c_row9_col5\" class=\"data row9 col5\" ></td>\n",
       "      <td id=\"T_8207c_row9_col6\" class=\"data row9 col6\" >0.9838</td>\n",
       "      <td id=\"T_8207c_row9_col7\" class=\"data row9 col7\" >0.9386</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row10\" class=\"row_heading level0 row10\" >13</th>\n",
       "      <td id=\"T_8207c_row10_col0\" class=\"data row10 col0\" >p3_efficientnet_b0</td>\n",
       "      <td id=\"T_8207c_row10_col1\" class=\"data row10 col1\" >EfficientNet-B0</td>\n",
       "      <td id=\"T_8207c_row10_col2\" class=\"data row10 col2\" >insight</td>\n",
       "      <td id=\"T_8207c_row10_col3\" class=\"data row10 col3\" >0.9257</td>\n",
       "      <td id=\"T_8207c_row10_col4\" class=\"data row10 col4\" >0.9257</td>\n",
       "      <td id=\"T_8207c_row10_col5\" class=\"data row10 col5\" ></td>\n",
       "      <td id=\"T_8207c_row10_col6\" class=\"data row10 col6\" >0.9776</td>\n",
       "      <td id=\"T_8207c_row10_col7\" class=\"data row10 col7\" >0.9273</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row11\" class=\"row_heading level0 row11\" >14</th>\n",
       "      <td id=\"T_8207c_row11_col0\" class=\"data row11 col0\" >p3_efficientnet_b0</td>\n",
       "      <td id=\"T_8207c_row11_col1\" class=\"data row11 col1\" >EfficientNet-B0</td>\n",
       "      <td id=\"T_8207c_row11_col2\" class=\"data row11 col2\" >text2img</td>\n",
       "      <td id=\"T_8207c_row11_col3\" class=\"data row11 col3\" >0.9782</td>\n",
       "      <td id=\"T_8207c_row11_col4\" class=\"data row11 col4\" >0.9782</td>\n",
       "      <td id=\"T_8207c_row11_col5\" class=\"data row11 col5\" ></td>\n",
       "      <td id=\"T_8207c_row11_col6\" class=\"data row11 col6\" >0.9921</td>\n",
       "      <td id=\"T_8207c_row11_col7\" class=\"data row11 col7\" >0.9548</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row12\" class=\"row_heading level0 row12\" >7</th>\n",
       "      <td id=\"T_8207c_row12_col0\" class=\"data row12 col0\" >p3_resnet34</td>\n",
       "      <td id=\"T_8207c_row12_col1\" class=\"data row12 col1\" >ResNet34</td>\n",
       "      <td id=\"T_8207c_row12_col2\" class=\"data row12 col2\" >wiki</td>\n",
       "      <td id=\"T_8207c_row12_col3\" class=\"data row12 col3\" >0.9067</td>\n",
       "      <td id=\"T_8207c_row12_col4\" class=\"data row12 col4\" ></td>\n",
       "      <td id=\"T_8207c_row12_col5\" class=\"data row12 col5\" >0.0933</td>\n",
       "      <td id=\"T_8207c_row12_col6\" class=\"data row12 col6\" ></td>\n",
       "      <td id=\"T_8207c_row12_col7\" class=\"data row12 col7\" ></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row13\" class=\"row_heading level0 row13\" >4</th>\n",
       "      <td id=\"T_8207c_row13_col0\" class=\"data row13 col0\" >p3_resnet34</td>\n",
       "      <td id=\"T_8207c_row13_col1\" class=\"data row13 col1\" >ResNet34</td>\n",
       "      <td id=\"T_8207c_row13_col2\" class=\"data row13 col2\" >inpainting</td>\n",
       "      <td id=\"T_8207c_row13_col3\" class=\"data row13 col3\" >0.9375</td>\n",
       "      <td id=\"T_8207c_row13_col4\" class=\"data row13 col4\" >0.9375</td>\n",
       "      <td id=\"T_8207c_row13_col5\" class=\"data row13 col5\" ></td>\n",
       "      <td id=\"T_8207c_row13_col6\" class=\"data row13 col6\" >0.9780</td>\n",
       "      <td id=\"T_8207c_row13_col7\" class=\"data row13 col7\" >0.9233</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row14\" class=\"row_heading level0 row14\" >5</th>\n",
       "      <td id=\"T_8207c_row14_col0\" class=\"data row14 col0\" >p3_resnet34</td>\n",
       "      <td id=\"T_8207c_row14_col1\" class=\"data row14 col1\" >ResNet34</td>\n",
       "      <td id=\"T_8207c_row14_col2\" class=\"data row14 col2\" >insight</td>\n",
       "      <td id=\"T_8207c_row14_col3\" class=\"data row14 col3\" >0.9105</td>\n",
       "      <td id=\"T_8207c_row14_col4\" class=\"data row14 col4\" >0.9105</td>\n",
       "      <td id=\"T_8207c_row14_col5\" class=\"data row14 col5\" ></td>\n",
       "      <td id=\"T_8207c_row14_col6\" class=\"data row14 col6\" >0.9688</td>\n",
       "      <td id=\"T_8207c_row14_col7\" class=\"data row14 col7\" >0.9087</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row15\" class=\"row_heading level0 row15\" >6</th>\n",
       "      <td id=\"T_8207c_row15_col0\" class=\"data row15 col0\" >p3_resnet34</td>\n",
       "      <td id=\"T_8207c_row15_col1\" class=\"data row15 col1\" >ResNet34</td>\n",
       "      <td id=\"T_8207c_row15_col2\" class=\"data row15 col2\" >text2img</td>\n",
       "      <td id=\"T_8207c_row15_col3\" class=\"data row15 col3\" >0.9672</td>\n",
       "      <td id=\"T_8207c_row15_col4\" class=\"data row15 col4\" >0.9672</td>\n",
       "      <td id=\"T_8207c_row15_col5\" class=\"data row15 col5\" ></td>\n",
       "      <td id=\"T_8207c_row15_col6\" class=\"data row15 col6\" >0.9870</td>\n",
       "      <td id=\"T_8207c_row15_col7\" class=\"data row15 col7\" >0.9389</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row16\" class=\"row_heading level0 row16\" >3</th>\n",
       "      <td id=\"T_8207c_row16_col0\" class=\"data row16 col0\" >p2c_resnet18_facecrop</td>\n",
       "      <td id=\"T_8207c_row16_col1\" class=\"data row16 col1\" >ResNet18 facecrop reference</td>\n",
       "      <td id=\"T_8207c_row16_col2\" class=\"data row16 col2\" >wiki</td>\n",
       "      <td id=\"T_8207c_row16_col3\" class=\"data row16 col3\" >0.9027</td>\n",
       "      <td id=\"T_8207c_row16_col4\" class=\"data row16 col4\" ></td>\n",
       "      <td id=\"T_8207c_row16_col5\" class=\"data row16 col5\" >0.0973</td>\n",
       "      <td id=\"T_8207c_row16_col6\" class=\"data row16 col6\" ></td>\n",
       "      <td id=\"T_8207c_row16_col7\" class=\"data row16 col7\" ></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row17\" class=\"row_heading level0 row17\" >0</th>\n",
       "      <td id=\"T_8207c_row17_col0\" class=\"data row17 col0\" >p2c_resnet18_facecrop</td>\n",
       "      <td id=\"T_8207c_row17_col1\" class=\"data row17 col1\" >ResNet18 facecrop reference</td>\n",
       "      <td id=\"T_8207c_row17_col2\" class=\"data row17 col2\" >inpainting</td>\n",
       "      <td id=\"T_8207c_row17_col3\" class=\"data row17 col3\" >0.9297</td>\n",
       "      <td id=\"T_8207c_row17_col4\" class=\"data row17 col4\" >0.9297</td>\n",
       "      <td id=\"T_8207c_row17_col5\" class=\"data row17 col5\" ></td>\n",
       "      <td id=\"T_8207c_row17_col6\" class=\"data row17 col6\" >0.9759</td>\n",
       "      <td id=\"T_8207c_row17_col7\" class=\"data row17 col7\" >0.9171</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row18\" class=\"row_heading level0 row18\" >1</th>\n",
       "      <td id=\"T_8207c_row18_col0\" class=\"data row18 col0\" >p2c_resnet18_facecrop</td>\n",
       "      <td id=\"T_8207c_row18_col1\" class=\"data row18 col1\" >ResNet18 facecrop reference</td>\n",
       "      <td id=\"T_8207c_row18_col2\" class=\"data row18 col2\" >insight</td>\n",
       "      <td id=\"T_8207c_row18_col3\" class=\"data row18 col3\" >0.8955</td>\n",
       "      <td id=\"T_8207c_row18_col4\" class=\"data row18 col4\" >0.8955</td>\n",
       "      <td id=\"T_8207c_row18_col5\" class=\"data row18 col5\" ></td>\n",
       "      <td id=\"T_8207c_row18_col6\" class=\"data row18 col6\" >0.9654</td>\n",
       "      <td id=\"T_8207c_row18_col7\" class=\"data row18 col7\" >0.8982</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row19\" class=\"row_heading level0 row19\" >2</th>\n",
       "      <td id=\"T_8207c_row19_col0\" class=\"data row19 col0\" >p2c_resnet18_facecrop</td>\n",
       "      <td id=\"T_8207c_row19_col1\" class=\"data row19 col1\" >ResNet18 facecrop reference</td>\n",
       "      <td id=\"T_8207c_row19_col2\" class=\"data row19 col2\" >text2img</td>\n",
       "      <td id=\"T_8207c_row19_col3\" class=\"data row19 col3\" >0.9613</td>\n",
       "      <td id=\"T_8207c_row19_col4\" class=\"data row19 col4\" >0.9613</td>\n",
       "      <td id=\"T_8207c_row19_col5\" class=\"data row19 col5\" ></td>\n",
       "      <td id=\"T_8207c_row19_col6\" class=\"data row19 col6\" >0.9851</td>\n",
       "      <td id=\"T_8207c_row19_col7\" class=\"data row19 col7\" >0.9338</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row20\" class=\"row_heading level0 row20\" >19</th>\n",
       "      <td id=\"T_8207c_row20_col0\" class=\"data row20 col0\" >p3_mobilenetv3_small</td>\n",
       "      <td id=\"T_8207c_row20_col1\" class=\"data row20 col1\" >MobileNetV3-Small</td>\n",
       "      <td id=\"T_8207c_row20_col2\" class=\"data row20 col2\" >wiki</td>\n",
       "      <td id=\"T_8207c_row20_col3\" class=\"data row20 col3\" >0.8692</td>\n",
       "      <td id=\"T_8207c_row20_col4\" class=\"data row20 col4\" ></td>\n",
       "      <td id=\"T_8207c_row20_col5\" class=\"data row20 col5\" >0.1308</td>\n",
       "      <td id=\"T_8207c_row20_col6\" class=\"data row20 col6\" ></td>\n",
       "      <td id=\"T_8207c_row20_col7\" class=\"data row20 col7\" ></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row21\" class=\"row_heading level0 row21\" >16</th>\n",
       "      <td id=\"T_8207c_row21_col0\" class=\"data row21 col0\" >p3_mobilenetv3_small</td>\n",
       "      <td id=\"T_8207c_row21_col1\" class=\"data row21 col1\" >MobileNetV3-Small</td>\n",
       "      <td id=\"T_8207c_row21_col2\" class=\"data row21 col2\" >inpainting</td>\n",
       "      <td id=\"T_8207c_row21_col3\" class=\"data row21 col3\" >0.9125</td>\n",
       "      <td id=\"T_8207c_row21_col4\" class=\"data row21 col4\" >0.9125</td>\n",
       "      <td id=\"T_8207c_row21_col5\" class=\"data row21 col5\" ></td>\n",
       "      <td id=\"T_8207c_row21_col6\" class=\"data row21 col6\" >0.9594</td>\n",
       "      <td id=\"T_8207c_row21_col7\" class=\"data row21 col7\" >0.8931</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row22\" class=\"row_heading level0 row22\" >17</th>\n",
       "      <td id=\"T_8207c_row22_col0\" class=\"data row22 col0\" >p3_mobilenetv3_small</td>\n",
       "      <td id=\"T_8207c_row22_col1\" class=\"data row22 col1\" >MobileNetV3-Small</td>\n",
       "      <td id=\"T_8207c_row22_col2\" class=\"data row22 col2\" >insight</td>\n",
       "      <td id=\"T_8207c_row22_col3\" class=\"data row22 col3\" >0.9137</td>\n",
       "      <td id=\"T_8207c_row22_col4\" class=\"data row22 col4\" >0.9137</td>\n",
       "      <td id=\"T_8207c_row22_col5\" class=\"data row22 col5\" ></td>\n",
       "      <td id=\"T_8207c_row22_col6\" class=\"data row22 col6\" >0.9601</td>\n",
       "      <td id=\"T_8207c_row22_col7\" class=\"data row22 col7\" >0.8937</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_8207c_level0_row23\" class=\"row_heading level0 row23\" >18</th>\n",
       "      <td id=\"T_8207c_row23_col0\" class=\"data row23 col0\" >p3_mobilenetv3_small</td>\n",
       "      <td id=\"T_8207c_row23_col1\" class=\"data row23 col1\" >MobileNetV3-Small</td>\n",
       "      <td id=\"T_8207c_row23_col2\" class=\"data row23 col2\" >text2img</td>\n",
       "      <td id=\"T_8207c_row23_col3\" class=\"data row23 col3\" >0.9747</td>\n",
       "      <td id=\"T_8207c_row23_col4\" class=\"data row23 col4\" >0.9747</td>\n",
       "      <td id=\"T_8207c_row23_col5\" class=\"data row23 col5\" ></td>\n",
       "      <td id=\"T_8207c_row23_col6\" class=\"data row23 col6\" >0.9845</td>\n",
       "      <td id=\"T_8207c_row23_col7\" class=\"data row23 col7\" >0.9258</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x1a61cef0320>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 650x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 650x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x380 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def source_rows(run_name: str, label: str) -> list[dict]:\n",
    "    rows = []\n",
    "    per_source = load_run(run_name).get(\"aggregated_per_source\", {})\n",
    "    for source, metrics in per_source.items():\n",
    "        rows.append({\n",
    "            \"run\": run_name,\n",
    "            \"label\": label,\n",
    "            \"source\": source,\n",
    "            \"accuracy\": metrics.get(\"accuracy\", {}).get(\"mean\", np.nan),\n",
    "            \"detection_rate\": metrics.get(\"detection_rate\", {}).get(\"mean\", np.nan),\n",
    "            \"false_alarm_rate\": metrics.get(\"false_alarm_rate\", {}).get(\"mean\", np.nan),\n",
    "            \"pairwise_auc\": metrics.get(\"pairwise_auc\", {}).get(\"mean\", np.nan),\n",
    "            \"pairwise_f1\": metrics.get(\"pairwise_f1\", {}).get(\"mean\", np.nan),\n",
    "        })\n",
    "    return rows\n",
    "\n",
    "\n",
    "def ordered_source_table(source_df: pd.DataFrame) -> pd.DataFrame:\n",
    "    source_order = {source: pos for pos, source in enumerate([\"wiki\", \"inpainting\", \"insight\", \"text2img\"])}\n",
    "    label_order = {label: pos for pos, label in enumerate(order)}\n",
    "    return (\n",
    "        source_df.assign(\n",
    "            label_rank=source_df[\"label\"].map(label_order),\n",
    "            source_rank=source_df[\"source\"].map(source_order),\n",
    "        )\n",
    "        .sort_values([\"label_rank\", \"source_rank\"])\n",
    "        .drop(columns=[\"label_rank\", \"source_rank\"])\n",
    "    )\n",
    "\n",
    "\n",
    "def plot_fake_source_heatmap(metric: str, title: str, colorbar_label: str, out_name: str, *, vmin=None, vmax=None):\n",
    "    pivot = (\n",
    "        source_df[source_df[\"source\"].isin(fake_sources)]\n",
    "        .pivot(index=\"label\", columns=\"source\", values=metric)\n",
    "        .reindex(index=order, columns=fake_sources)\n",
    "    )\n",
    "    fig, ax = plt.subplots(figsize=(6.5, 4.8))\n",
    "    im = ax.imshow(pivot.values, cmap=\"viridis\", vmin=vmin, vmax=vmax)\n",
    "    ax.set_xticks(range(len(pivot.columns)))\n",
    "    ax.set_xticklabels(pivot.columns)\n",
    "    ax.set_yticks(range(len(pivot.index)))\n",
    "    ax.set_yticklabels(pivot.index)\n",
    "    for row_idx in range(pivot.shape[0]):\n",
    "        for col_idx in range(pivot.shape[1]):\n",
    "            value = pivot.iloc[row_idx, col_idx]\n",
    "            if pd.isna(value):\n",
    "                text = \"\"\n",
    "                color = \"black\"\n",
    "            else:\n",
    "                text = f\"{value:.3f}\"\n",
    "                threshold = (vmin + vmax) / 2 if vmin is not None and vmax is not None else np.nanmean(pivot.values)\n",
    "                color = \"white\" if value < threshold else \"black\"\n",
    "            ax.text(col_idx, row_idx, text, ha=\"center\", va=\"center\", color=color, fontsize=8)\n",
    "    ax.set_title(title)\n",
    "    fig.colorbar(im, ax=ax, label=colorbar_label)\n",
    "    fig.tight_layout()\n",
    "    fig.savefig(FIGURES_DIR / out_name, dpi=180, bbox_inches=\"tight\")\n",
    "    plt.show()\n",
    "\n",
    "\n",
    "source_df = pd.DataFrame([row for run, label in RUN_LABELS.items() for row in source_rows(run, label)])\n",
    "source_table = ordered_source_table(source_df)\n",
    "display(\n",
    "    source_table.style.format(\n",
    "        {\n",
    "            \"accuracy\": \"{:.4f}\",\n",
    "            \"detection_rate\": \"{:.4f}\",\n",
    "            \"false_alarm_rate\": \"{:.4f}\",\n",
    "            \"pairwise_auc\": \"{:.4f}\",\n",
    "            \"pairwise_f1\": \"{:.4f}\",\n",
    "        },\n",
    "        na_rep=\"\",\n",
    "    )\n",
    ")\n",
    "\n",
    "fake_sources = [\"inpainting\", \"insight\", \"text2img\"]\n",
    "plot_fake_source_heatmap(\n",
    "    \"pairwise_auc\",\n",
    "    \"Pairwise AUC by fake source\",\n",
    "    \"wiki-vs-source AUC\",\n",
    "    \"06_phase3_pairwise_source_heatmap.png\",\n",
    "    vmin=0.94,\n",
    "    vmax=1.0,\n",
    ")\n",
    "plot_fake_source_heatmap(\n",
    "    \"pairwise_f1\",\n",
    "    \"Pairwise F1 by fake source\",\n",
    "    \"wiki-vs-source F1\",\n",
    "    \"06_phase3_pairwise_f1_heatmap.png\",\n",
    "    vmin=0.88,\n",
    "    vmax=0.97,\n",
    ")\n",
    "\n",
    "wiki_df = source_df[source_df[\"source\"] == \"wiki\"].set_index(\"label\").reindex(order)\n",
    "fig, ax = plt.subplots(figsize=(8, 3.8))\n",
    "ax.bar(wiki_df.index, wiki_df[\"false_alarm_rate\"], color=\"#E45756\", alpha=0.85)\n",
    "ax.set_ylabel(\"False alarm rate on wiki real images\")\n",
    "ax.set_title(\"Real-image false alarms by model family\")\n",
    "ax.tick_params(axis=\"x\", rotation=25)\n",
    "fig.tight_layout()\n",
    "fig.savefig(FIGURES_DIR / \"06_phase3_wiki_false_alarm.png\", dpi=180, bbox_inches=\"tight\")\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b6e87ba0",
   "metadata": {},
   "source": [
    "The top Phase 3 models improve pairwise source discrimination, but they do not remove the need for source diagnostics. `insight` remains the most delicate fake source for several models, and wiki false alarms still matter because a useful detector must avoid calling too many real images fake.\n",
    "\n",
    "The heatmaps below are the right place to compare source behavior. Relative to the Phase 2 ResNet18 facecrop reference, the strongest Phase 3 backbones lift the weaker fake-source pairs and make the three fake-source columns more similar. This matters because the improvement is not only a higher global AUC; the best backbones also reduce the gap between fake sources. ResNet50 is essentially tied with ConvNeXt-Tiny in this source-wise view, while later confusion and balanced-metric analysis explain why ResNet50 is selected as the best practical detector.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cbeffa38",
   "metadata": {},
   "source": [
    "## 4. Confusion and error balance\n",
    "\n",
    "AUC ranks examples across thresholds, while the confusion matrix shows what happens at the operating threshold used in the saved evaluation. This view separates the two practical errors: false positives on real wiki images and false negatives on fake images.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "ac46bf91",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style type=\"text/css\">\n",
       "</style>\n",
       "<table id=\"T_b6de1\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th class=\"blank level0\" >&nbsp;</th>\n",
       "      <th id=\"T_b6de1_level0_col0\" class=\"col_heading level0 col0\" >label</th>\n",
       "      <th id=\"T_b6de1_level0_col1\" class=\"col_heading level0 col1\" >run</th>\n",
       "      <th id=\"T_b6de1_level0_col2\" class=\"col_heading level0 col2\" >true_real</th>\n",
       "      <th id=\"T_b6de1_level0_col3\" class=\"col_heading level0 col3\" >true_fake</th>\n",
       "      <th id=\"T_b6de1_level0_col4\" class=\"col_heading level0 col4\" >false_positive_rate</th>\n",
       "      <th id=\"T_b6de1_level0_col5\" class=\"col_heading level0 col5\" >false_negative_rate</th>\n",
       "      <th id=\"T_b6de1_level0_col6\" class=\"col_heading level0 col6\" >fake_detection_rate</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th id=\"T_b6de1_level0_row0\" class=\"row_heading level0 row0\" >5</th>\n",
       "      <td id=\"T_b6de1_row0_col0\" class=\"data row0 col0\" >ConvNeXt-Tiny</td>\n",
       "      <td id=\"T_b6de1_row0_col1\" class=\"data row0 col1\" >p3_convnext_tiny</td>\n",
       "      <td id=\"T_b6de1_row0_col2\" class=\"data row0 col2\" >6000</td>\n",
       "      <td id=\"T_b6de1_row0_col3\" class=\"data row0 col3\" >18000</td>\n",
       "      <td id=\"T_b6de1_row0_col4\" class=\"data row0 col4\" >0.1162</td>\n",
       "      <td id=\"T_b6de1_row0_col5\" class=\"data row0 col5\" >0.0315</td>\n",
       "      <td id=\"T_b6de1_row0_col6\" class=\"data row0 col6\" >0.9685</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_b6de1_level0_row1\" class=\"row_heading level0 row1\" >2</th>\n",
       "      <td id=\"T_b6de1_row1_col0\" class=\"data row1 col0\" >ResNet50</td>\n",
       "      <td id=\"T_b6de1_row1_col1\" class=\"data row1 col1\" >p3_resnet50</td>\n",
       "      <td id=\"T_b6de1_row1_col2\" class=\"data row1 col2\" >6000</td>\n",
       "      <td id=\"T_b6de1_row1_col3\" class=\"data row1 col3\" >18000</td>\n",
       "      <td id=\"T_b6de1_row1_col4\" class=\"data row1 col4\" >0.0908</td>\n",
       "      <td id=\"T_b6de1_row1_col5\" class=\"data row1 col5\" >0.0321</td>\n",
       "      <td id=\"T_b6de1_row1_col6\" class=\"data row1 col6\" >0.9679</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_b6de1_level0_row2\" class=\"row_heading level0 row2\" >3</th>\n",
       "      <td id=\"T_b6de1_row2_col0\" class=\"data row2 col0\" >EfficientNet-B0</td>\n",
       "      <td id=\"T_b6de1_row2_col1\" class=\"data row2 col1\" >p3_efficientnet_b0</td>\n",
       "      <td id=\"T_b6de1_row2_col2\" class=\"data row2 col2\" >6000</td>\n",
       "      <td id=\"T_b6de1_row2_col3\" class=\"data row2 col3\" >18000</td>\n",
       "      <td id=\"T_b6de1_row2_col4\" class=\"data row2 col4\" >0.0708</td>\n",
       "      <td id=\"T_b6de1_row2_col5\" class=\"data row2 col5\" >0.0497</td>\n",
       "      <td id=\"T_b6de1_row2_col6\" class=\"data row2 col6\" >0.9503</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_b6de1_level0_row3\" class=\"row_heading level0 row3\" >1</th>\n",
       "      <td id=\"T_b6de1_row3_col0\" class=\"data row3 col0\" >ResNet34</td>\n",
       "      <td id=\"T_b6de1_row3_col1\" class=\"data row3 col1\" >p3_resnet34</td>\n",
       "      <td id=\"T_b6de1_row3_col2\" class=\"data row3 col2\" >6000</td>\n",
       "      <td id=\"T_b6de1_row3_col3\" class=\"data row3 col3\" >18000</td>\n",
       "      <td id=\"T_b6de1_row3_col4\" class=\"data row3 col4\" >0.0933</td>\n",
       "      <td id=\"T_b6de1_row3_col5\" class=\"data row3 col5\" >0.0616</td>\n",
       "      <td id=\"T_b6de1_row3_col6\" class=\"data row3 col6\" >0.9384</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_b6de1_level0_row4\" class=\"row_heading level0 row4\" >4</th>\n",
       "      <td id=\"T_b6de1_row4_col0\" class=\"data row4 col0\" >MobileNetV3-Small</td>\n",
       "      <td id=\"T_b6de1_row4_col1\" class=\"data row4 col1\" >p3_mobilenetv3_small</td>\n",
       "      <td id=\"T_b6de1_row4_col2\" class=\"data row4 col2\" >6000</td>\n",
       "      <td id=\"T_b6de1_row4_col3\" class=\"data row4 col3\" >18000</td>\n",
       "      <td id=\"T_b6de1_row4_col4\" class=\"data row4 col4\" >0.1308</td>\n",
       "      <td id=\"T_b6de1_row4_col5\" class=\"data row4 col5\" >0.0664</td>\n",
       "      <td id=\"T_b6de1_row4_col6\" class=\"data row4 col6\" >0.9336</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_b6de1_level0_row5\" class=\"row_heading level0 row5\" >0</th>\n",
       "      <td id=\"T_b6de1_row5_col0\" class=\"data row5 col0\" >ResNet18 facecrop reference</td>\n",
       "      <td id=\"T_b6de1_row5_col1\" class=\"data row5 col1\" >p2c_resnet18_facecrop</td>\n",
       "      <td id=\"T_b6de1_row5_col2\" class=\"data row5 col2\" >6000</td>\n",
       "      <td id=\"T_b6de1_row5_col3\" class=\"data row5 col3\" >18000</td>\n",
       "      <td id=\"T_b6de1_row5_col4\" class=\"data row5 col4\" >0.0973</td>\n",
       "      <td id=\"T_b6de1_row5_col5\" class=\"data row5 col5\" >0.0712</td>\n",
       "      <td id=\"T_b6de1_row5_col6\" class=\"data row5 col6\" >0.9288</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x1a61cca2090>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABKUAAAFGCAYAAAC7VSzAAAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjEwLjksIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvJkbTWQAAAAlwSFlzAAAPYQAAD2EBqD+naQAAnDtJREFUeJzt3Qd4U1UbB/A3SQez7E2h7L33kiHIniKoKMhUBEFwAMpGQRQQBAQBmYIgAoLAB8oG2XvvvfeGjuR+z/+EpEmbtgm0adL+f9+TT5rc5N7ccXLOe895j07TNE2IiIiIiIiIiIjcSO/OlREREREREREREQGDUkRERERERERE5HYMShERERERERERkdsxKEVERERERERERG7HoBQREREREREREbkdg1JEREREREREROR2DEoREREREREREZHbMShFRERERERERERux6AUERERERERERG5HYNSRERECcj+/fvl9ddflzRp0ohOp5PBgwfHyXpmzpypPn/Dhg1x8vkJEfbXBx98IJ7u559/loIFC4q/v7/a5vPnz4s3qlGjhgQFBYm3w/6Pi2sZ5yI+lzzHqx4TnCPefM0SUeLEoBQREVEsePr0qYwdO1aqVasmadOmFV9fX8mUKZM0aNBABXDCwsLifD9jHW+++aacOnVKhg0bJnPmzJEWLVrE+XoTk7/++ivOAn2eYP369dKtWzcVlJo8ebI6hzJkyBDfm+XRcH3j2vcElmCx5aHX6yVVqlRStWpVmT17tiQ0o0ePVsHHLFmyqCAq/luzZk1ZsmRJfG8aERE5ycfZBYmIiMix06dPS8OGDeXkyZNSu3Zt6devn6RPn15u3rwpa9askfbt28vRo0fl+++/j9NdePbsWfVAQ6179+5xuq73339f3n77bfHz85PEFpSaNWvWSwWmnj17JgaDQTzZv//+q/47ffp0FVwl5wJB6Jny6aefeszu6tGjh5QrV05MJpPatqlTp0q7du3k8uXL8tVXX0lCsXPnTtUbDsF/lLl3796VhQsXqmD80KFDZcCAAfG9iUREFAMGpYiIiF4BAg2NGjVSwaBFixZF6pnUp08f2bVrl3rEtevXr6v/uiOYgOCKpwdYPOX8QK85Hx8fSZIkiXi6uDqHbPcDxT302GzZsqX1bwTGCxQoICNHjpQvv/wywRyHBQsWRHoOwcEyZcqomwAIwLGcIiLybBy+R0RE9AqmTZsmJ06ckM8++yzKoXLosfDxxx9H6nFTpUoVSZ48uaRIkUL9e+nSpZHei14AGJ5y/Phx1RsrZcqUajgOGpyWAAJgmerVq1sboJbhO+glEV3+J0d5d7Zu3Sr169eXzJkzq0BKtmzZVE+E7du3W5eJ6jNv376thn8FBgaqXlT4L/6+c+eO3XKW969bt05GjRolefLkUcNv8ufPr3oiOQPrxmfgs5AHCY1ubG+xYsVk+fLlaplDhw5JvXr1JCAgQNKlS6d6kISGhkbqbYFcLlh3smTJ1D7G8Yg4BAj7yrJttkOksH7bfDC3bt2SDh06qOGbOL7onRJdTikMmcOxxfZh+3Pnzi0dO3ZU+zJiAxzDsLB92M4KFSrIn3/+GenzVqxYoc4F9BxJmjSp5MiRQ52b6MkXU96iGTNm2H0/fGeLgwcPSvPmza3bWbhwYdXwNxqNdp8V035w5Pnz56r3GY4hvlvq1KnVcfziiy8iLYveh2+88YZaBttRvHhxNdTQWRjeip5+GOqFcxTnP9bz5MmTSMviGsM5g2OC8zNjxoxSp04da48yvHfjxo1y4cIFu3PC9rpwZX1btmxR5x6OG/Ybejw+fvxYXhWuQxyvhw8fquNi68GDB9K1a1f13bA/sf4dO3bYLYMeV99++6289tprqlzA98B5hfdFvLYBQwXLly+vjhGOPfZfmzZtIq3blX3jLATcUGbhMyJe645Yzld8D/wb1w2usWbNmlnL2ClTpkihQoXU/sHQVkdlNYZPI+iH/YzlcJ3gekEZ5Oh8x/fMmjWrOtbYV//880+U2/gq+wm9x3r16qXKWMt2IWj3ww8/xPheIiJ3SBi3SYiIiOKJJSjQpUsXp9+DAIolb8/AgQPVcwhsoBH0yy+/RPqsK1euqOAAGjhoSBw4cEAthwampSHz9ddfq8bk8OHD1fvRUwJczQeEABsa3Wh49uzZUzWMb9y4oRrLWG/FihWjfC8at5UrV1bDGRGMKF26tOzbt08mTZqkgk8I/qCxZws9GdCL5sMPP1SNfiyLhmHevHnV93HGxIkT5d69e9KpUyfV6Prpp5/UvsIwns6dO8s777yj9i321fjx41Xju3///tb3I/iEoF+rVq0kZ86cqnGK4BMCOXPnzpV3333Xuo/RON+8ebPKtWSB72zLsv8wdAiNRgQdo4LjiIY9GtH4L9Z/8eJF+fvvv1UQBw1kwPYiKIAAG/KFIVcQtvutt96SCRMmqPMJECBp0qSJFC1aVA0jRVDg6tWrKpCD44LAmyM4T/Cd0Pi2/X44/rB7924V6EJvJ6wL3w/biJ6AOC+wnyJyZT/gMzFksG3bttK7d2/VwEdDHOeNLWzfRx99pM5DHA8EPBAgwr47c+ZMjA3tPXv2SK1atdR+wTmH/Y7txznz33//qf2H72gJ1OEcxPmP7Spbtqz6HgjOYn/i+yGXFPYzAog//vijdT0IYLi6PgSCMPwX1wj2K94zf/58te5XFRwcrM4rBGzwubbq1q2rjj/KIpz7Y8aMUUHSc+fOWa/XkJAQtW+Rs65p06Zqv6P356+//qrKBnxPy1BenDsYKogyCEPoEHS5dOmSrFy5Ug1ptpRJruwbZwIvCI7iOOC6X7Vqlcot5UrvRFxb2bNnV9uMa8VSjqAcwHmHQLGlfMFNAQR5c+XKZX0/gm5//PGHOi9wPiKghbKpUqVK6poqVaqUdVmUSbgx0bhxY7X/ce5iPbafZ/Gq+wllxKZNm9R1gwAuyttjx46pwKmjoC8RkdtpRERE9NLSpk2rBQQEOL383bt3teTJk2t58uTRHjx4YH0e/86dO7eWIkUK7d69e9bnc+bMqeHnesGCBXaf8/HHH6vnjx8/bn1u/fr16rkZM2bYLYu/8Txej6h69epqHRbjxo1Ty+7YsSPa7+HoM7/66iv13MSJE+2WnTBhgnq+f//+kd5fsmRJLTg42Pr85cuXNT8/P+3tt9+Odv223zdr1qza/fv3rc8fOHBAPa/T6bRFixbZvad06dJa5syZ7Z57/PhxpM9+8uSJlj9/fq1QoUJ2z7dr1059tiOW19q0aePwdbyGZSwuXbqkvivWYXvMLYxGo/rvnj171Hv79esXaZmmTZtqKVOm1B4+fKj+7tWrl1r2xo0b2suI6vtVrlxZMxgMat9amEwm7a233lLLr1mzxun94EiaNGm0+vXrR7vM1atXNX9/f+2dd96J9FqPHj00vV6vnTlzJspzG4oXL64VKFDAur8sFi9eHOnawfbguVWrVkV5bKJaz8usr1KlSpqvr6924sQJ63O4NsqVK6eWHTRokBYTy3U1ffp07datW+o82LlzpzpP8LztdWU5Tl27drX7jD/++EM9P3nyZLtj/fTp00jrmzZtWqTyqXnz5uqcDA0NjXZbXdk3MUmXLp16Dx4+Pj5ay5YttZs3bzr1Xst+QJlqy3ItBQYG2pXVlvKlb9++1uf++ecf9VyrVq3UvrLYv3+/um6qVq1qfW716tWRygJYsmSJ9Tu87H7COYLnzp07p/5GuejoGBMReRIO3yMiInoF6K0UsfdPdNCrA70tMCQIQ8os8G88h6E66IVhC0M80IvHFu6cA3qTxCYMDQQMT8EQE1eg5w56QUTs6YW7+3je0YxYGNZomywdvQDQm8eV74WeVZbtBvQGwP7Efos4pBLD39CDwXZIFHp92M6iiN4i+C/2MXoU4Bi74vPPP3dqOfToQA+UQYMGReq9AugNBeiFhOFF6H2CniC2D/SKevTokWzbtk0ta9kPyG8WWzM+oncLhnRiXdi3Ftgm9FYCR8fW2f1g2e4jR47I4cOHo+2ViB4/lqGNtg/0OEEvtojXji0Mo8IQRPR8w+fYvh/nBc4DS89D9LxBbxv0nkFPlqiOTXRcWR/2MY4heiHZ9mbDtYGhV65CT0Vcc+jphqFh6KWE8wcJzyOK+PmOyhYca/R4AvRIun//vvoelmVth/vhWOL6wTBScyz21faNMxYvXiyrV69Wve3QUwm9gXBduCJionpLb1P0VLMtqy3li+3+sZz/uB6wryxKlCihzk30JrMMXUQPKYjYSwm9OTF8NTb3E44ZeqDi+KDnHxGRJ2JQioiI6BWgceJK4wdDYqBIkSKRXrM8h6TptpCPJSLkBQFH+VxeBWbUwxAiDANEsms0OpEnBTlznPluaFRFTKKMv9HQjvi9ovturnwvR5+RJk0ah0Nh8DzYfj4CAgikWXIfYcgcGvSWPEVogLsiqiFyEVkatbbDehxBYAyNewz3xHbZPhCgAQwxA+Qgwuch2Ifjh1xgGOYTMZePK6I7ZzFMDQEaR8fW2f0AGAaHIZjII4XcNxiKicAoAk22+wFwfkbcDwhE2O4HRyzvRxAw4vsxpBPBYsv7MXwL+zymYxMdV9Zn2X84xhEhR5GrMBQPAfC1a9eqYbM43zFE2NEQyojXT1RlC4amIY8ZAh24jvA9LO/FsbMdkothqAiyYBkM+UPuPdty0pV94wzkukKeMeTTQwDOkhfOdrtiEnE/WMqKqMoR2/2DawTXgWXYpi3LdWO5jnCssayj6yPi+191PyGoiWsLwV58D2zLJ598os4LIiJPwZxSRERErwC5e5CvAw0NR8GR2BDd7FFR9USwZXvnPqKIvWlwVx2NWTRk0fMA3w0NXCShnjdvnsqx4o7v5sz3iukznNlv+C8as2j8IYcW8gahpwfei6Tf+M62gRFnIFF3bMI24hj+73//i/I7WRq+CCgg1w9y2OA44vihJwwatWisI7+Nu7iyH9BDCD05sI3IkYMeT8hXhN4q+Dca15ZjhiTaSPjsSHTXoOX9mJQAPaAcsQQiYoO712cLwT0E72LrGkRPpNatW6teV+PGjVOJ05FfCb2m8N1sr5F8+fLJ0aNHVeADDxxP5HbDOYjzEUHHuN436BWGfFzYbkvgNrbLEVfKqJcVG/sJuaRwfaHnGo4FehwiDx2OJ/YREVF8Y1CKiIjoFaAXABpa6AmA3kUxsTSaMVTp9ddft3sNDTnbZWILesxYhiRFhLv3jpLkovGJByBJMXqMINl2dEEpbDcSpSPQZdtbCn8jKXBcBe1eBYbGIGkwAm9Dhgyxew3H1JUAn6ssPSX2798fba8iNPIxlAyznTnqieGoEY3E+JaZ8/AdMdvWN998oxqmrrL0FME5GxESxCMgERvHFufpe++9px5ojPft21fN7oceU0jWjP0A6MnmbMDFluX92D8xvR+J9nGscWxiEtU54cr6LPsY+zMiS7kQn5C8HEEozBRpG2x0tL2W4DZ66eEBCDYieTqSqCP5tyv75mVg+F5UZV5cwPmP6wDBbdshrrbHz3KMLcuiTIzY+9DSM8oitvYTgrjofYgHAomYye/3339XwS7MDktEFJ84fI+IiOgVoJKPIWujRo1yOE24ZfYkzLgHGGaEIWKYBc52OAv+jecwvMYyFCm2WAIeEfPtoFGCmdlsIVdJRJiRCsNFYmrgYbgOholFDOYgjw2ej+1eVrHB0gsiYq8HDHdxlCfJMvwpNhq7mMELPYAQDHOUt8qyTWhAWoZFoUEZke3wHUfHD0PCMOTqZbcZw4QwwyBm27PN+YTtGzFihPr3qxxbS46iiIEey9A5y3YjrxqCHehxYwk6RJz9EXl3ooLPQ89GDMt0NNwQwVPLuhAgq1+/vuqd5ihPle35gnMCw8QinkOurA9DRzGjIMoQBCsskHPMdla/+LxOcExse0Th+yLQGZGjcxAzcYLl+7qyb6KCoWu2ueFszycEviC62UJjE8o+wPVgex7gelm2bJnK/2SZdRC9liDiTJHINYWgvq1X3U/I7YVHxGNpCZy5K2hHRBQd9pQiIiJ6Beg1sHz5ctULAA0TDAVDUAnDqBCIQc8CDIP78ssv1fJIaI3eH926dVP5WZCkG5DvBXlsfvnlF7uk3bEBQTPcZcdno8FUsmRJ1QMEQRf0CAkNDbUui0YmEuc2atRI3dnH8ghGoEeE5TtEBa8jeTe+2969e1WDat++fWoYFrYhpvfHB/Q8Qm8FHBM03rCdCApgX2EIFAKKttDIxdAX5GzCMUcvMxxHR3lnYoJgH/K9YH9hXUiojFw8V65cUcEJJG3GsUJPBgyfxAN/o9cQkrhfu3ZNbR96oSB4ARgmdfnyZXUe4rMQvFmwYIEKeuLzXxaGbFWvXl0Np8P2Zs6cWZ33OLeRhDlirz9XYNvQkwOJ1HHOIAiGHnyTJk1SQ5OQKNqyv/AcAsE4bgjW4TviOkNCaDTq0SslKCjI4XoQVEGPH+RJQ6McycBx7HHcce1hqBeCCpZrEscZwTgEpzAcDL3NsD+RNBrrQK41yzmBfYF8XlgejX6sA9/DlfWhFxF6tyEXEvYxygoMr4qthPWvGkBF8nx8F5xHKDOwvyMGPADnHrYd5wqG+SHgiPIN+98SYHX1WESVkw3nJLYN1y0Cibh2EGxHcAfHzJKsPK6hzEfQFMcLAUqUn5hQAcEx9DBDXjcLJM7HOT1r1iwVFMKwvDNnzqgyBwEo28Dvq+4nlGXYRwga47NxPaE3Fq4jlFnu2j9ERNGK7+n/iIiIEoInT55oY8aM0apUqaKlTp1aTUueMWNGrUGDBtrs2bO1sLCwSNN5Ywr4ZMmSqQf+jSnBI8JU85hyPqL169dHmg7c0XMW165dU9OkY6r25MmTa/Xq1dOOHj0aaTp7fAamNcdzSZIk0dKkSaOVL19emzp1qt1U55ap57G8LUzDjunHs2XLpvYB/oup1jE9va2o3g8Rtykq0X3fqPZbxCnT4fz582rfpE+fXkuaNKlWrlw5dXwcLWs0GrXPPvtMfS+9Xm+3fsvU8lFxNA28ZYr42rVrawEBAZq/v7+WK1curVOnTtrt27ftllu+fLn2xhtvqGPi5+enZc+eXR3HSZMmWZdZtGiR1rhxY7V9WAbf6bXXXtP+/PPPGPdnTN8B09s3bdrUuv6CBQtqI0eOjHRux7QfIgoODtb69u2r9nvatGnVZ+P4tW/fXjt58mSk5bds2aI1a9ZMy5Ahg+br66tlyZJFq1GjhjZq1Cjt2bNnMZ5HON4ffviheg3vxzpLly6ttuHixYt2y16+fFktGxgYqJbFNV2nTh1tzZo1dtd+hw4d1GuWc8L2vHZlfRs3blRlAc4DfB6unUOHDqnPxPkYE8t1tXDhwhiXje44OTpXp0yZohUqVEhtW+bMmbXOnTtrd+7cibQslsP5nClTJvV9sWz9+vW1devWRVqPK/smIpQp3bp104oXL67OSZQ36dKlU+v+7bff7Mqrl9kPrpYvoaGh2nfffaeuC5zD2CZcLwcPHoz0/qdPn2q9e/dW+wjlLM59lANRbYuz+ylimYUy5NNPP9VKlCihpUqVSq0rT548Ws+ePbWrV686tX+IiOKaDv8XfdiKiIiIiIiIiIgodjGnFBERERERERERuR2DUkRERERERERE5HYMShERERERERERkdsxKEVERERERERERG7HoBQREREREREREbkdg1JEREREREREROR2DEoREREREREREZHbMShFRERERERERERux6AUERERERERERG5HYNSRERERERERETkdgxKERERERERERGR2zEoRUREREREREREbsegFBERERERERERuR2DUkRERERERERE5HYMShERERERERERkdsxKEVERERERERERG7HoBQREREREREREbkdg1JEREREREREROR2DEoREREREREREZHbMShFRERERERERERux6AUERERERERERG5HYNSRERERERERETkdgxKERERERERERGR2zEoRUREREREREREbsegFBERERERERERuR2DUkRERERERERE5HYMShERERERERERkdsxKEVERERERERERG7HoBQREREREREREbkdg1JEREREREREROR2DEoREREREREREZHbMShFRERERERERERux6AUeYTBgwdLpkyZRKfTyV9//RXfm+M1rl+/LnXq1JHkyZNL6tSp43tziMgFQUFBMnbsWO4zIiJKcDZs2KDq9ffv349ymZkzZ9rVX9EeKFmypJu2MHGrUaOGfPrpp9a/WSeh+MSglJt88MEHqmDGw9fXV3LlyiVffvmlPH/+PNbWgc9OkiSJXLhwwe75Zs2aqfW/6o/Ipk2bpHHjxpI1a9Yog0ePHz+W7t27S/bs2SVp0qRSuHBhmTx5crTrO3bsmAwZMkR++eUXuXbtmtSvX9/pbU3sfvzxR7XP9u/fLydPnozvzaFEGhj95JNPJHfu3OLv7y+BgYGqnFi7dq3btgGVWpRJ9erVs3seZRieR5nmrNatW0v58uXFaDRanwsNDZUyZcpImzZt1N/nz59Xn4vrLroyNLoHltm1a5d06dLlpb83UUIS1/WkuK4joTEd8TovWLCg3TL4Lt26dZN06dJJihQp5M0335QbN2684jcjiv3r8KOPPor0Gs5dvObK9eLMb+6r1l/j8tpGfSZi3cJi8+bNavmDBw/KnTt31HJoI1nqQmgPPXz4MNp1Pn36VPr16yd58uRR3yFDhgxSvXp1Wbp0qdPbTZQQMCjlRiisEEA4e/asCiYgCDNo0KBYXQcKx4EDB0pcePLkiZQoUUImTpwY5TK9e/eWVatWyW+//aaCTYjAo1BetmxZlO85c+aM+m/Tpk0lc+bMqjD3NJqmSVhYmNvWh0awM7Dv0FjOly+fZMyY8aXWFRIS8lLvI0JwBuffunXr5IcffpBDhw6p679mzZqq8upOPj4+smbNGlm/fv0rfc7PP/8sFy9elO+++8763LBhw1TZPWHCBKc+o3Llymp5y6NVq1bW8t/ywDKofCZLluyVtpcoIYnrelJc1pGgSJEidtf5li1b7F7v1auX/P3337Jw4ULZuHGjXL16VVq0aBFn20P0MhBQmT9/vjx79swuoDpv3jzJkSNHrO5U3MB+2fqrO67tjh07yr///iuXL1+O9NqMGTOkbNmyUrx4cdHr9aodg/YOgmy4WYY6iaPgni28vnjxYhk/frwcP35c1aFatmypglxEiYpGbtGuXTutadOmds+1aNFCK1WqlPVvo9GoDR8+XAsKCtKSJEmiFS9eXFu4cKH19bt372rvvvuulj59evV63rx5tenTp1tfx+H8/PPPNb1erx06dMj6PNaL9TuznnPnzqnPsX3Yvtd2XUuWLIn0fJEiRbShQ4faPVe6dGnt66+/drhfBg0aFGl9sHPnTq127dpaunTptICAAO21117T9uzZY/fee/fuaV26dNEyZsyo+fv7q3X//fff1tc3b96sVa1aVX3H7Nmza5988on2+PFj6+vPnz/XvvzyS/Wan5+flidPHm3atGnqtfXr16ttWblypdp+X19f9Rzeg8/JkCGDWmeVKlXUtlpY3rd8+XKtWLFiapkKFSrYHQ9H8J6ff/5Za9y4sZYsWTK1X+Cvv/5S5wg+J1euXNrgwYO10NBQ9VrOnDkdHifsl44dO6rzJGXKlFrNmjW1/fv32+3zEiVKaFOnTlXngE6nc+l9s2fPVuvGcWndurX28OFDu3Nr5MiRal9inwYGBmrffPON9fWLFy9qb731lpYqVSotTZo0WpMmTdQ5R96pfv36WrZs2eyuKwucT3DhwgV1nJMnT67OKxz/69evO31e/fLLL1qWLFnUuWULn9m+fXv17xkzZqhzqnPnzlr58uXttgHXBq5LV87BpUuXqvP3wIED2q5duzQfHx9txYoV1tcjllnVq1d3ufwHfN8ff/zR7nNxXTZr1kxLmjSpKuOxLWAymdR19cMPP9h9xr59+9T7Tp06Fe02EHl7PcnT60iWsiwq9+/fV3UJ220+duyY+oxt27a94t4jit3rsGjRotpvv/1mfX7u3LnqWrC9XmKjTmz5/bZwdB3hd7FgwYLq/QUKFNAmTpxo93pcXtuoc2fKlEkbNmyY3TofPXqkpUiRQps0aVKU+3LcuHGqjREdfPeZM2dGuwzqClj/+++/r+pSOXLkUHWDmzdvWutX2L+or1jcvn1be/vtt7WsWbOq+gSO57x58+w+F3WXnj17RlknIXInBqXiqbKFQjNz5syqcLZA4x2F7qpVq7QzZ86oghoF8IYNG9Tr3bp100qWLKkKHRSe//77r7Zs2bJIgSIUUA0bNoyyUI5uPWFhYdqiRYvUZ504cUK7du2aqkg5G5RCo7Bs2bLa5cuXVSNq3bp1qtDeuHGjw/2CQh3rx+dhXXjA2rVrtTlz5qgK29GjR1WwBD8KloYqflgqVqyoAlH//POP+h4ISCGIBKdPn1aFNArXkydPav/995+q2H7wwQfWdbdq1UoFTRYvXqzev2bNGm3+/Pl2P6T4wcLn4/Pu3Lmj9ejRQxXwWM+RI0fUfkXDFq/Zvq9QoULqfQcPHtQaNWqkfgBDQkKiPD/wHgTXUIHGtqAhv2nTJtVAx48VnsPn4XMQmAL8GNWrV099D9vjhGAegls4T/DdP/vsMxXcs2wjfvCxb/DevXv3qoa3s+/DsUQjAecvtg/n8FdffWX9HgjyYX9gm7HPEBhEZQLw/bFfOnTooPYLjisaEKhgBAcHR7lvyDPhvEBAExW8qOA6RZmF4PDu3bu17du3a2XKlLEL4sR0XqGhiQARrk/bdds+Z6nUXrlyRVW+LJXMiEEpV87Btm3bqopx4cKFVfljC5VufC7Wj2vPco3ERlAKFVhUHBFkQnmDfWP5/G+//VZtjy0sg6A9UUKvJ3l6HQllGW4qIYiOm0goW/BbboF6Dd5nCdhboIE5ZsyYONmnRC97HeKcfP31163P49/4vbK9XmKjThxTUAqBMVxTuO7Onj2r/ps2bVq7QE5cX9tffPGFuimEdo0F6uuobzhqIwHqI6jrtGnTJtr9jfoH6vG2N3gd1RXwnSdPnqzq5127dlXtA9Tj//jjD7W9uJmF/WzZRrTDcBMLN67wXX/66SfNYDBoO3bssH4ug1LkSRiUchMUiigMEAxAAYhCDxH9P//803q3AZWZrVu32r0PjaF33nlH/RsBA0vPAEcshTJ+GLAuNO4g4l2NmNZj+RGJWHFytK6I8PlozOF19C5Aw3HWrFnR7ht8Tkyd9tC4RS8LS0+o1atXq/2HgtgRfB/0orKFAAne8+zZM/U+rBOVVkcs+wA9lSzQGwR3OXG3yAI/qvhB/v777+3eZwluAX6c8cO1YMGCKL8f3vPpp5/aPYcKQMQGPwJ1+HGO6gcX3xE/VDgOtvBjih4nlh98fA8EtVx9H84d2x9O/FBbGgx4Hue2JQgVEbYdP762P+oIBGDf4HiSd0HFBuctgrpRQSUUZRF6J1mgfML7LHdTYzqvLOc5AkkWOCdx3Vl6T9lWavv27avlz59f3d2MGJRy5RxEMAzPIxj+4MEDu9csd1RR2XOGK0Gp/v3725U5eO5///uftZJrW6lE+YNeITHdZSXy9nqSN9SR0DBHAxE3etDorVSpkgo4Wco21B1QJ4qoXLly6oYOkSew/F6hjojr8Pz58+qBnkW3bt2yXi+xVSeOKSiFemjEHj7oNYTry13XtqVHo22v62rVqmnvvfdepP2H3kn4flgeZRLaHNHBTXvcjMK+xE19tAW2bNkSqa5guy4EzPD5AwYMsD6H3paWG/xRQcAON5wtGJQiT8KcUm6EPCtIjLtjxw5p166dtG/fXiW5hNOnT6tkd5hJDckvLY/Zs2dbcy517dpVjfHGrBRI/rl161aH60Fy8bZt20rfvn0jvebMel4FxkRv375djanes2ePjB49WuWWwbhqVyDxZ+fOnVWupFSpUklAQIBKoo5cL4D9iGTq+fPnd/j+AwcOqPHctt+xbt26YjKZ5Ny5c+r9BoNBJROMDsaKW2D/INdTlSpVrM8hGSuSIiN/lq1KlSpZ/502bVopUKBApGWiW5flOwwdOtTuO2CfIE8FjmFU3xv7yZJE1fLAd7Y9vjlz5lT5bFx9H2bmSJkypfXvLFmyyM2bN9W/8f2Cg4Pl9ddfj3LbcP7h/ZbPx75BnoLYOPfIvcz1wOjhnEBuCjxsyyfMtGN7PUR3XgESjC9atEidXzB37lx5++23VQ6HiPr06SO3bt2S6dOnv9I5+Pvvv6scFbdv31Z5HmKChKe21w628WUgN4UFZtVE2WfZF0ig2rBhQ+t3Q24a7JO33nrrpdZF5C31JG+oI2GSFlyLuIZR31i5cqVKmPzHH3/E0t4hch/UEfF7g7o0cifh3+nTp3dbndiSyxbrQV4n2+vxm2++cXg9xtW1jQkLkAfS8tuLz8JvPrYrIuTC27t3r0pUjs9Frl1A+8V2vcOHD1fPv/baayqHHiaHQS6pI0eOSLVq1VQuy6jqBpitHIoVKxbpOUt9AZO14DOwDPY51rl69WprO4rI0/jE9wYkJmhg5M2bV/0bBRuShv/666+qUENAAFasWCHZsmWze58l8TcqPJhZAhUdJN1D4x8Bn1GjRkVaF2azQ8Am4gx5zqznZSEh4ldffSVLlixRP16WQhQVTGxj7dq1nf4sVEaR5G/cuHEqgIJtw4+aJSk3EiNGB9/zww8/lB49ekR6DUka8YPi7DFzl4jrwnfAcXSUBBUzdDiC96BB72i2Mdspdx2ty5n3ocJhC412BPqcPSZIiu2osW4bICPvgIAxjr8zAZuYRHdeWWa/QRAM5Va5cuVUZRAVP0dwvmImG1w7jRo1eqlzEBVENGonTZqkEqdj9p59+/ZFW0YiqGw7G5+lghjb+6JTp07y/vvvq++PhgJmLmKydEro9aSiRYt6XR0JZRHWYalvYCIX1GEQqLL9XcVNOLxG5Gk6dOigJiuC6CY5iiuW63Hq1KlSoUIFu9dwY9mRuLq20VbDTMPYD/jtxWx5jm5s41rGA4EsBIMQYBowYIC6qWRbR8Brtr/7WA4P3FhD0A03pfFvPz8/6zK29YKonrPUFzD5DNpQY8eOVYEplK2YfIqTG5GnYlAqnuAOPwI4iKC/++67KrqPQhER7Oh676DhhIANHii8vvjiC4cVLstUpFgHCk4LZ9ZjKQBtp0R3Bu6Y4BGx9wJ+OGwbVc7477//1CxYDRo0UH9funRJ9ViwQLALM2FghgtHvaVKly4tR48etVZuI0IBjW3C7DfOBsuwH7FvsG0IlFm+M6Z1R0FvC73FLDOU3Lt3T21noUKFXNgD5u9w4sSJKL9DVO+5fv26mokMvU/i+n0RgxQITOFuDxrOjtaxYMECNcsKen+Qd0OFCr0BUEFD8DdioBMNL5zzuHbxsPSWwnWJ11AWOQtBWARnEUxCAw93WXE+RQUVx59++klVyFw9B1EuIAiFBi3uuGI2HTSIMavPyJEjoywjce67cq2+LJSJ2NcImGGWnk2bNsX5Ooniu56E31BvqyOhEYyeEggiAwLiaETiN9LSSx6/8ViXbU8SIk+aDRNBDAQ88Hvv7joxbu4gmIMbRegx7Yy4urYxi27Pnj3VDIToXYWemZZAUFQsbR/0aEb92tk6ArYVM36jF7dlm1yF44L6y3vvvWfdFux3V+peRO7EoFQ8QjdvVJjQqPv888/VA9MFo+CoWrWqPHjwQBUqaDyhgoVGESo1mHIYBdzy5cujLdTRWwB3FzAEC3fTAcNWYloPflxQ0OLz0QBCYwvdPlHBsu1hZBkGh8YpfmzwfhT0+E54Dz4HQR8U3mPGjHE5wDFnzhzV++Dhw4fWz7TAetDlFRU7fDYKevTYwHbjRxR3FypWrKh+mBAgQSMOjWHcPcW07gi84LviLhAar7gbizus6PaKHx5H8Bn4EcK2WL7z999/r7oDR+zCizscGAqHH9Svv/5adXlu1qyZS/sAxxs9PbAedOlFBR3Djw4fPqzuojiCABsqt1gXtg0BO0w5jTtDzZs3jzRE8FXfFzFwgP2OHib4EUWXbgyjQldk7B9UKHDnBj+S2D8Yfol9jqlw8R78Td4FZReOM7rr45giWIyKFK4zBE1wzSEAjGOPu3V47eOPP1bXrzPnlC18Bq4HnE+WSlZ05yLulqKXRMTPiOkcRCAL68ADMHx42rRpat0ob/BdEdRCeYSgEN6D9WE5d0CQH0EzlO8oJ9mYpcRQT/rll188vo6E96FXJ17H7+egQYPU9frOO++oz0YZgd9CBNlQh8DnIYCOaxj1FSJPg/PXMswuYs8kd9WJ8VuOG1+4flC/x7W9e/duFdyyDI2L62sb8F98Fj4b7RL8DttCD030ekRvbiyLOgT2DepI0d3srVGjhiojUCfCPkK9CQE1DGV+lRu4qB/8+eefahhzmjRpVFsJ28egFHms+E5qlVhEleh2xIgRaipVJAxE8t2xY8eqRLxIeIfn69ata525Don9MLMCEuhhFgZ8HmaiiC75OBJl205ZDDGtB4YOHapmvcHsWpb3WhIARjUdMiDBHma4Q6JDJEXEOkaPHm2XWNiZROeYFQ4J//AZ+fLlU7NpRUwKjGSJSGqKGeKwHKY7xbSzFkikXKdOHTV7FRKnYiY9zF5lgeSDvXr1UonDkXzUdvroqJId4j2Y/hbJhaOb/hYJ2TEzID4XU9RbZrhzNXE8kqVWrlxZHXMkIsdnTZkyJcpE54CkqthGHAMcX8wwiNk/LMmmo5q2+mXeh+OB42KBxNOY3QTP4TOQ5NU2WTvODyTCt+y/3LlzqxkbIyaSJu9x9epVNesVjjnO92zZsqkZcCwJQTH7lGXKYkxW8NZbb2nXr1+3vt+Z88pybuFaxbWCmWRsRUyUCphJBzPVRUxOGt05iAkQcK3ZJm61wDIofy2TASChP64RJGK2nU3wVROdRywH8L3w/Wzh+2NZSzJZosRST/LkOlLr1q2t9QmUg/gbs9BGrEN8/PHHaoYyJF1u3rx5tImJiTzlOnRU74yNOnFMic4Bv8mYWRPvx7WDGWdtJ1mJ62vbAknS8ZkNGjSItF8w2ziSr+O7WNouffr0iXbSKMt24n0os/A+1Ekwq+Ht27ejrCs4+s4RJ2BBGwnHCm0gzO6NSVRQ97E9tkx0Tp5Eh/+L78AYUUKBnEy4u4E7OLY5I4iIYgtyamF4IYZFvmzuKiIiIiIiT8DZ94iIiLwAhi0gl97gwYPVsCYGpIiIiIgotiBXKYaiI58bhrNGnDQgqk4ZyNuKvG1IqYNZO13FoBQREZEX+P3331XOCySKR+4OIiIiIqLY8uTJE5Vr2dkZN5G7rWHDhmqkEHJNY6ID5HNevXq1S+vl8D0iIiIiIiIiIlLQU2rJkiXRTkyASa4wMRYm4rJ4++231Q1UTAjkLPaUIiIiIiIiIiIip23btk3N4m6rbt266nlX+Li0NBERERERERERud3z588lJCTEqWUxpx16PNlC7ic8YsP169cj5TjF3w8fPpRnz55J0qRJnfocBqWIiIiIiIiIiDw8IJU0ZTqRsKdOLZ8iRQp5/Pix3XODBg1Sk+Z4EgaliIiIiIiIiIg8WAh6SIU9Ff+i7UUMftEvbAyRx4dnyKVLlyQgIMD6dGz1koLMmTPLjRs37J7D31ifs72kgEGpWGAymeTq1auSMmXKSN3jiOjVoNvpo0eP1NSker1np8FjWUAUd1gWEBHLAiLytnIgTvgkEZ0h+uCSpjPvFwSIbINSsalSpUqycuVKu+f+/fdf9bwrGJSKBQhIBQYGxsZHEVEUEOXPnj27R+8flgVEcY9lARGxLCAibykH4oROZ37EtIyLMNTv9OnT1r/PnTsn+/fvl7Rp00qOHDmkX79+cuXKFZk9e7Z6/aOPPpIJEybIl19+KR06dJB169bJH3/8oWbkcwWDUrEAPaTAr8Yg0fkkiY2PpDi2f1pH7mMv8fjRIylXLI/1OvOKsqD2cNH5sizwBrt/bhPfm0AulAWViuf1rrKg+kDWC7zE/umd43sTyJV6QdHc3lUW1PtedL7OD2Wh+LNj7Fvc/V5SDlQtmc8ryoE4odObHzEt46Ldu3dLzZo1rX/37t1b/bddu3Yyc+ZMuXbtmly8eNH6eq5cuVQAqlevXjJu3DgVIJw2bZqagc8VDErFAsuQPQSkGJTyDinjqAsjxR1vGBprLQt8k7Dy6SVSpmRZ4G28qixgvcBrsF7gfbyqLPBNynqBl2C9wLt4QzngTT2latSooYZGRgWBKUfv2bdvn7wKBqWIiIiIiIiIiLyC3omeUN6Ta4tBKSIiIiIiIiIib6A3mB8xLeMlGJQiIiIiIiIiIkrEw/fiC4NSRERERERERESJONF5fGFQioiIiIiIiIjIG+jYU4qIiIiIiIiIiNxNx55SREREREREREQULz2l9DEv4yU4fI+IiIiIiIiIyBsYDOZHdDTOvkdERERERERERLFJx5xSRERERERERETkbjrmlCIiIiIiIiIiInfTsacUERERERERERG5m449pYiIiIiIiIiIyN107ClFRERERERERETupjeYHzEt4yV84nsDiIiIiIiIiIjIGU4M38MyXoJBKSIiIiIiIiIib6Dj8D0iIiIiIiIiIoqXoJQ+5mW8BHtKERERERERERF5Ax1n3yMiIiIiIiIiInfTcfgeERERERERERG5m449pYiIiIiIiIiIyN30BvMjpmW8BHNKERERERERERF5Ax2H7xERERERERERkZvpdDr1iGEh8RbsKUVERERERERE5AV0DEoREREREREREZHb6V48YlrGS7CnFBERERERERGRF9CxpxQREREREREREbmbXq8XnV4f7TJaDK97EvaUIiIiIiIiIiLyAjr2lCIiIiIiIiIiIrfTMacUERERERERERG5mY49pYiIiIiIiIiIyN10OnNgKvqFxGswpxQRERERERERkRfQ4X8xBaW8KCrFoBQRERERERERkRfQcfgeERERERERERG5nV4nOn30PaG0GF73JPr43gAiIiIiIiIiInK+p1RMj5cxceJECQoKkiRJkkiFChVk586d0S4/duxYKVCggCRNmlQCAwOlV69e8vz5c5fWyaAUEREREREREVEiDkotWLBAevfuLYMGDZK9e/dKiRIlpG7dunLz5k2Hy8+bN0/69u2rlj927Jj8+uuv6jO++uorl9bLoBQRERERERERkTfQOflw0ZgxY6Rz587Svn17KVy4sEyePFmSJUsm06dPd7j81q1bpUqVKvLuu++q3lVvvPGGvPPOOzH2roqIQSkiIiIiIiIiogTWU+rhw4d2j+DgYIefGRISInv27JHatWtbn9Pr9ervbdu2OXxP5cqV1XssQaizZ8/KypUrpUGDBi59H86+R0RERERERETkBXRODM+zvI48T7Yw1G7w4MGRlr99+7YYjUbJlCmT3fP4+/jx4w7XgR5SeF/VqlVF0zQJCwuTjz76yOXhewxKERERERERERF5Ab1erx4xLKT+c+nSJQkICLA+7e/vH2vbsWHDBhk+fLj8/PPPKin66dOnpWfPnjJs2DAZMGCA05/DoFQiZXpyS0IPzRMJeSLik0R8i70j+pRZ7JbRNJOEnfhbTLePi2gm0afOJT5FWopObz5tjDePSNiJZeo1Xcos4lvsXdH5JImnb5SwnT1zWnp93Enu3r0tAQGpZMyEqVKgUGG7ZS5dPC+9unWWIwcPSGDOIPlnk/1Y3t/nzJCJ40aJZjJJ5Wo1ZPion8TX19fN34Q8ienxTQndN0sk5LGIb1LxLdlW9AFZI5cDRxaL6eYREZ1BdH7JxadEG9GnyKheDzu1WoyXtougXDD4iG/R1qJPExRP3yhhO3fmtHzWvZPcu3NHUgYEyKgJUyV/wYjlwAX5vHtnOXrogGTPGST/27DD+trWTRtk5LD+8uTJE3X3rFadetJn4DcxV2oocdQJDv9uXydIkTlyWXByuU2dIEh8CpvrBFpYsITunymmh5dFNKMkeX14vH2XxODsmVPSq6tNnWDitEh1Auvv/tgfzL/7r9W0/u6bTCb5dlA/2bDmHwkzhknZCpVlxOjx4ufnFy/fhzyH6fENCd093Vwv8EkqvmXbiz4gW+Sy4NBCMd04LKLTi84vhfiUbiv6FJnEeOOwhB1eFL5s8EPR+acS/9cHxsO3SfjOnT0tX3TvLPfumusF3/80JVK94PLFC/Jljy5y5NABCcyRU5avD68X7N21QwZ+2UP9Gz1cypSvJAOHj47VoAXFEZ0TOaNevI6AlG1QKirp06cXg8EgN27csHsef2fObF8nsEDg6f3335dOnTqpv4sVK6bqmV26dJGvv/7a6Toma6KJVNiRP8SQvZL4v/aV+OSuJaGHfo+0jPHyDtEeXha/yp+JX9W+6AMoxgub1GuqAnp4vviW6iD+r32tfnDCTv8TD98kcejbu5u0addBNu86LF17fCa9u3eOtEyKlAHy5deDZcLUWZFeu3jhnIwaMVQWr1grW/Ycldu3bsrcWb+6aevJU4UdnCuGnFXF//Uh4pP3DQndPzvSMqbrB8V094z41egv/jX7iz5DAQk7vtT82oNLEnZ+k/i91kf8a3wtPkE1JPTQ/Hj4JonDV591l3fadpT1Ow/JRz0+U8GniFKmTCmffzVIxv0yM9JrAalTy/ipc2TN1n3y99qtsmfXdlm0YK6btp48WdjRhWLIXlH8q/UTn1wx1Akq9Ra/Kn1UbddSJxC9QQy5aolf2Y/cv/GJUN9e3aXNBx1l8+4j0rXn59K7m7kxEOl3f/gQWbxynWzZe0xu37whc2dOswarDh3YL//buEM27DioGg2/Th4fD9+EPE3YvjliyPWa+L/xrfgUqCehu2dEWsZ07YCY7pwWv9cHiX/tIaLPWEjCjixRrxkyFRV/PP/ioU+dUww5KsTDN0kc+n/+ibzdtoOs3X5QPvzkMxV8iihFypTSu+8gGTs58rEsVKSYLPlniwpUrdy4S+7cviVzZ0xx09aTp82+5+fnJ2XKlJG1a9dan8NNDPxdqVIlh+95+vRppMATAluA4XzOYlAqEdKCH6nGpCFrGfW3PlMJ0Z7fV3dK7ZZ7dFX06fKru6A4qfXpC4nxym71mun2MdEHZFd3RcCQo4oYr+2Nh2+T8CGAdHDfXmnR6l31d8MmzeXqlcty7uwZu+XSpEkr5StWkaTJkkX6jBVLl0ideg0lY6bM6li+176zLF20wG3fgTwP7l6a7l8UQ/by6m99llKiPbunek/Z04mYwkRMoerHRQt9LrokaWxeM4qEhZg/M+yZzWsU2+XAof17pflb76i/6zduLlevXpHzEcqB1GnSSjlVDiSP9BlFi5eUHEG51L+TJEkihYsWV3dQKXGz1gmyWOoExZ2rE2QoJMare9RreM6QLp/qcUluqBPs3+OgTnDabrkVSxc7+N3/Q7129PBBqVa9lmqA4LWatevKogXzeOgSOe35QzHdOy+GwIrqb33WMqI9u6t6T0WCeoEx7EW94Jnokkb+7dee3RfTzWNiCHTcmKVXLwsO798rzVqa6wX1GjWTa1cuO6wXlK1Y2WG9AG0Gy6iJ0JAQef78mcuBDEo4QSno3bu3TJ06VWbNmiXHjh2Trl27qp5PmI0P2rZtK/369ROLxo0by6RJk2T+/Ply7tw5+ffff1XvKTxvCU45g8P3EiFUNnX+AaLTm08UddImTaOel+QZrMvpArKL8dI2MeSsJqL3FeP1/erHSX3Gs3t2jU9d0rQiwQ9FMxmtn0uxA5XNjJkzi4+Pj/V4Zc0eKFcvX5RcufM49RlXrlyS7IE5rH8HBuaUK5cv8RAlYuoadlQO4Bp/MTQP9JmLif7OCQle3UcN69ElSSV+VXqbX0uVXXzy1JLgtf1FfJOrhqlflc/i7TslZKhoonFpWw5ky5Zdrl6+JEFOlgO2bt64Liv//kumzw0fZkGJk8M6QZLUkesEqQLFeGmrGHJUjVQnIDfXCTI5qhNckly581qXw2989sCc1r8DcwRZf/eLlywtv82cJh907ipJkiaV5Uv+lMuXGKBO7HA94zferixIlla0p6gXhCc+1mcpIfpbJyR4ZW9zvSBpavF77ctIn2e88J+qQ+iSxDxsiFx37eplyeCoLLjiWr0AN6c+bNtKLp4/KzXq1JM27SP3tiLP40qic1e0bt1abt26JQMHDpTr169LyZIlZdWqVdbk5xcvXrTrGdW/f3+1Hvz3ypUrkiFDBhWQ+vbbb11aL4NSFCVDtvKq4RqyY4KIwVfdIZU7DDgRJSba/YuiPbwq/m+MUJXPsKN/SeiB38WvTHsxPbktxmv7xf/1oaoRG3Zug4TsmSb+VT+P782maDx69FA6tXlTPureS4qXMveOIYqJIWs51WgN2TnRXCdIm09Ed4I7zgu1eretXL50UVo2qi1JkiSVqjVqiWH9mvjeLPIS2r3zoj28Iv71R4n4JlE5pEL3zRG/cuFDytGDynhhi/iUMPfiIc+VPUdOWbFhhzx5/Fh6f9xBVq9YKo2bvxXfm0WxmFPKVd27d1ePqBKb20JQFLP54fEqOHwvip2NiN/9+/clIVJ3QF/0agLV9Vb1fEptv5xOJ7756ol/lc/Fv2JPNVRP9+JOibln1T3rsupuqc2dVoo9WbNll5vXr6sEhJbjhTuiWbOH93yKSbZsgaoCanHp0gXJlt1+elBKXGWBuoYdlQPo9WgDScz16QuIzjeZ6HR61a3fdMfcEDVd26cSoFrKDnTR1+6eEQ3d+ilWZUE5cMO+HLhy5bK6K+qKx48eSbtWTaRO/cbS6eOePEpOStBlgaM6AXpPOaoT5K0n/pU/E/8KPVQidF2EZOjkpjrBDUd1AvuyAL/xtr2fMBmK5Xcfx/KzvgNk9aadsvSfjZK/QCEpECE5MiXCsiBpWtGeP7AvC57eVb2lbBkvbhN9hoKi83tRL8hZWUy37APUptsnRDOGiT5TUbd+h8QkS9bscstRWZDt5er3yVOkkEbN35Jli5gbNDEP34svLx2UqlGjhnz66aexuzXkFjr/lOaheS9yQZhuHFDddfU23fRBM4aKFvrU/O+QxxJ2dq1KgAr69AXVLDuWcebGi/+JIUspHsE4kD5DRilaoqQs/sOc72HFsiWSJWs2p4fuQYMmzeTfVStURRY/Wr/NmCpNWrSKle1r2biODOrHnjHeRg3XwXCcyzutASY0Qi2z6lmXS57eXLl8EWgy3Tgk+pTmGfp0ydKrJOha2HPra7rkGa0zdFLslgNFipeUJQvNCaj/9/cSyZIlm0td9HEXtF3rJlK91hvyyWd9Y/3wtG7yhgz5mmWB19YJrlnqBAedqxOcC68TkJvrBMVLOagThA/dgwZNmjv43Tf3fnj+/Lncv2++sXj3zm01Q1/XHuZh2bGhZSPUCziU29tgmJ0udQ7zjLooC67uUTewLPljrcslzyCmW8fD6wXXDkaaoc94fosKViFoRXFbL/jrT3O9YNXyvyRzVtfqBcg/FRoaqv4dEhIi/6xcJgUKM5DoDfQ6vRpGF+3Di66/OG05qK6bRqN1rGtcw8XE6Wyd41uklYQemifGs2tEfPzV9M+AGfX0GYuKIWNRkbDnErJzgpp1TzRNDDlfMz+PHyRMGV20tYTum66mhsbdUt9i5qSbFPtGjpkovbp1lvE/fi8pUwbI6AnmmTE+7/GRvFG/kXo8e/pUqpUrJiEhwfLo4QMpWySPvNn6Hek38BvJGZRbevfpL83r11Tvq1jlNXnvg8iz9cQVlgWeybfEuxK6b7YYT60yTwNfqq16PnT/HNFnLi6GzCXEEFRdtEfXJWTDtyI6g6q0+hQ3lxf6LCVFf/+ChGz6TgSBKIO/+JbpEM/fKuEaPnqCfP5JZ/n5x+/VbJs/jP9FPd+nZ1epXa+h1HlRDtSsgHIgRJUDFYvlkeat3pU+A4bJjCkT5cDe3WqmlFUrzDMoNmzSQrr3xkxq7sGywDP5FnlLzbhnPLvWXCco+rZ6PvTwAtFnLBJeJ9g18cV4AEudoIj1M4L/+0EFqyQsWJ5vGCL6tHnFr3ibePxWCdfIHyeY6wRjRr6oE0wNrxPUayhvNGhs/t3vO0Ca16uhXqtYtbpKdg4oG95qXEc1WjCzUscPu6vyw51YFngm1ANC90wX44mV5npBGXNy49A9M9VvviFrSTHkrinao2sSsnbwi3pBKvEp9b71MxC8Nl3dK361h8TjN0kcvhk1Xr78pItMGvuDmmXv+3HmekG/Xl3l9boNpXY9c72gdqXi1npBlRJ5pdlb78oX/YfKti0bZNa0SWLQG8RoDJNK1WrIJ73Dk1hT4sspFV90mitz9b3wwQcfqIzstpBt/fz581KzZk1ZuXKlSnZ16NAh+eeff2TmzJmqm+tff/1lXR69rPbv328dl4gfxZEjR8qUKVNUUq38+fOrzO0tW7aMcjuCgoKkY8eOcurUKfXZLVq0UOvasmWLygq/e/duSZ8+vTRv3lxGjBghyZObZx2YM2eOjBs3Tk6cOKGeq1WrlowdO1YyZjT3EMA24Xvcu3dPUqe2777uyMOHDyVVqlTiX3uECtaQ5zs9r2t8b0KC0KtbJ1n4+292z23bf1wuXbwgrZrUldkLlsoPwwfL8aOHZe6i5bLw9zny8MED+fW3hdbl0cvqyOED8uff/1rLgp/HjZK5s36VmzdvSFBQbjlx/Kg8ePBAAgICvKMsqD9GdJyFyiscn/5BfG9CgvBZ986yaL59WbB573GVQPWdZnVlxvy/ZPTwIXLi2GGZvXC5/DkfZcF9mTonvCxAL6ujhw7KgmX/WMuCST+Nlt9n/yq3bt5QjeyT3lYWvD6c9QIvcXp+t/jehASh18eoF8yxe27bgRPmekHjN2T2H0vlh29f1AsWr5CF88xlwa9z/7Quj15WRw4dlD+X29QLxqJeMM1cL8iVR04cO+JdZUHj8awXeIkjk3mT3VvyY5bMkznaciAheviiTMnR9Q/R+0eecd2WKfipXJzUyiv20Uv16UJhXalSJencubNcu3ZNPQIDw8ev9u3bV7777js1jWDx4sWd+kz8IMyePVsmT54sR44ckV69esl7770nGzdujPZ9o0aNkhIlSsi+fftUEOvMmTNSr149efPNN+XgwYOyYMEC9QNkm6wL3RSHDRsmBw4cUD9UCKYh0EZErhkyYrSUKVdR3m3bQfYeO68etmPZRwztL/0GDpP12/dLoSLFnPrMCT9+L3/OnysjRk+QdVv3SruOH6nncR2zLCDyTIOGj5LS5SrIO+93kJ1HzqkHct9YjBw2QPoMHCZrtqIscG5owM9jf5DFC+bKt6PGy79b9sr7HT9Uz7MsIPKCekG7DrL3+AX1sKsXDOkv/QZ9I+t3HHC+XjDme/lzwW8yYswEWbdtn7RjWUBEiZwugeWUeqlxdYjOYZhcsmTJJHPmyEkuhw4dKnXq1HH684KDg2X48OGyZs0aFeyC3Llzq4rnL7/8ItWrV4/yvbiD8dln4ePWO3XqJG3atLHmu8qXL5/89NNP6jMmTZokSZIkkQ4dwoeXYD14vVy5cvL48WNJkSKFU9uLh23EkigxCghIJb5+vpI0aTI1RXVEn/cbKK/VrO305+G6whDF+YtXSpnyFdVzzd96W776oofMmDFDGjRoEOV7WRYQxXNZ4OsnSZIldVgW9O4zQKrVeN2lsmDi2O/lt0UrVAMXmrV8WwZ80ZNlAZEHC0iFeoFf1PWCr16mXjBS5i/5n0294B356nPWC4go8dIlsOF7cZLsqWzZsi4tf/r0aZXjImIgC2NfS5Uq5dK60PsJPaTmzp1rfQ4jFNH1F0MMCxUqJHv27JHBgwerZdH9Fq/BxYsXpXDhwk716hoyhOOkiWJSvGRpl3YSEi5i7Ps7bzYMf/LFCGNcv9FhWUDkuYqXcq0suHDOXBa83zI8z40l2wDLAiLvVbxkmZerF7SwuSnFsoCIEjmdzvyIaZlEHZSyjMu2QCLFiKmrLJn+AT2UYMWKFZItm/3sDf7+/i6tC5/14YcfSo8ePSItmyNHDnny5InUrVtXPRC4ypAhgwpG4W8EwZyBsei9e/e26yllO3zRk4Tsmyk+QTVEnyZIQk+tEuPFLSohISA5uV+J8MSEoAU/kuD/vhd96iDxK91RPWd6cFFCjy0R7eEVNTW85XkL093TEnp8mYgpBLlPVYJUrM9484iYbh5WCdHJdR9+8K50+biHujM4+rthMuvXXyRzFvOsZ/kLFpIJU8x53e7cviWfffKhXL50UcJCQ6VkmbJq6FvSpEmtn/Xs2TNpULOSunv5zybzbGtHjxyS4YO/lt8WLouzw5MsYlmgi1wWhIWFlwVPnpjLglnzl0jmLOay4MnjR1KvRkU1vDc6LAuiF7JrqvjkeV30aXNL6PHlYjy/0Trluy5lFvF7kaA87NwGMZ7fLIIZOzSjGHJWFZ/c5hm2NM0kYYcXiunGYZXs2JCnlvjkqhHj+/CahD4Tn/z1nTpvyN7HHd6Vjl17WHsMYea9sd9/a72Wfp23WAJz5FR5l/p/0UMunDurfmPfbddROn70iVpmwpiRsmLZYutnXjp/Tlq/94EM+OZ7OXbkkHw3tL/MWmBOfh4XkiZzVC+wXyYs1DyTk21ZMH3eEmu5h+ca1mRZ8CpC9s8Sn6Dq6jc+9DTqBP/Z1wmKvxe5TrD1B3OdoNSLMuLKTgk7/peaPl69zyep+JXvFuNrqk5w64iaaIVezocfvCNdPu4ZXi+YNtmmXlBYJky1z/d6+9ZNqV2ljJQpV8Gas+mPebNlUN/PJDBnkPo7Veo0svBvcx63NatWyD+rlsv3Yye5r17goI2AukykesGCv8LLgsePpV71CqwXvIKQHZPEJ+8bok+XR0KPLhXj2fVqpj3QBWQVv3LmpPhhp9eI8fymFxMciPjkryeGHOZRLaa75yT04HzRHlwSfcbC4lcpPFVK2IX/JOzAfDWDr/pM32Ti99oX6t/GawfEdO2A+JY2T65CrunesY10+KiHGjI/qM+nsmfnNutrZ06flD4Dv5UPOn8sz589k68//0SOHNqvXsuRM0hG/DhJ0qXPINv/2yQd3mkmufPks773z5UbJEnSpHL8yCEZOay/zJgfd3UCenU6PcrP6KNOmj4RBKUwfA8z6zkDgZ/Dh9GICYck576+vurf6J2E4BOCQ9EN1XNG6dKl5ejRo5I3r/3UuBZIvn7nzh2V88oSSEKyQ1dgW2MKlnkC0/0LIqFPVYDIwpC1jPgWah7le0KPLBRDhiKihT6xmzret2AzMT28Iqbbx+yW154/kJCD88Sv7Idqylg1PazRXJnArDxhp1eJ6cmtSFNLU/T27dkl9+/ftXZVh+Yt35YhI0ZFWvan0SPVVNAzf1+srsm2rZupSqcl5wIMH/K1lK1QSQ7sM0/5DYWLFFPn8X+b1kuV18yz8r0MP18/MZqcKwvSps8gx48ftXvuyKED4vOiLMhfoJDapiuXL0mlKq+p5x69GB6bPXt4fhpnsCwIZ7p3XiT0iQpIWRiylxPfopEbh4bsFayBJi30mQRvGCb6dPlEnypQTJd3qhl3/F4fooJMwRuHiz5dftEHZI32fQhQhawbIoZcNZjo1UX79+6S+/fuWQNShw/uV0mCf1/yP8mUJas8fvRIDAaDeu2bAX0kX4FC8susBfL0yRN5s2EtKVu+kpQoXVbNsGeZZQ/DYSoUza2GwwHyuuA3feumDVL5NfMxfBn4DJOT9YK06dLLyWNH7J47eviA+PiYy4J8+QuJn7+/XL1ySSpWqWZNagosC16xTpDapk6QpXT0dYKjqBMUVrNp2VIz670IUkUU1WuqTnBmNesEr1IvuBehXvAW6gWjo3xPn17d5PW6DeT+3Tt2z1euVt0usbgFZvFEsOvsmVN2jVVX+fn5Ot1GQFlwPEJZgCTnkeoFl1gviC2mu2dFQp6ogJSFIbCi+JYw/ybY0gVkE7/qfVVQSXt6V4LXDRVd2jyiT5FRBbR9i7cW0/1LYrpxKNJ79RkK2AWqrOvKUkLCji0V0+Mbqu1AzjuAOsH9eyogBUNGjrW+duvGdalerrCaVRcwScjzZ0/lfxt3qWFc/Xp/LFMnjpW+g75Vr+MaX75+R6R1FFR1An/ZunmDVK728nUCilu6BDZ876XjZ5jVYseOHSpJ+O3bt61D4KLK9YLAD3o6YBaMQYMG2QWpUqZMKZ9//rlKbo5Z/ZCsfO/evTJ+/PhIs/zFpE+fPrJ161aV2ByBL6xv6dKl1kTn6C2FijM+++zZs7Js2TKV9DwhMl7aJvqszg+ZCLu8Xd3d1KUJb7gCelPoU+c0T/ke8T0X/1OBLsuPik7vY9foNGQuKcbL21/peyRGv82cJs3edK6HGQqcx48fqWsQvf3wA5Qla3iPw80b1sr1q1dV5TWipi1ayW8zf32lbc2eI6eqLF+6eF7u3om+LKhSrYYc3LdH/pz/m5w9c1pGjRgqJ46FB6kwne2H3T+VIV9/qWbvOX/ujBw5fFC9Nm/ePJe2i2VBOOOFzaLPVs6p/WY3a6AxRMQm4Gi8skcMOauITqcXnV9yde0br+yK8X0oF/QZConxsrmXHjlv3qxfpalNWTDt53HSqWsPFZCyXDNJk5lnX0GPp5q161p7I1SoVEWWLIx83fyzcplkyZZditkMr23SopWa2epVZA/MKftVWXAhxrIAFd2D+/fKogVz5dyZ0zLmu2FyMkJZ0KXbpzKs/5eqvEDvr6MsC16J8fI20WeJPiVC5DpBukh1gldhyFSCdYJXqRe8CCQ74/c5M1TPCJQDrmjUrKX8PnuGuK1e8JptveDUi3rBkQj1gl4y5OsvwusFhw6o11gveDnGc5tEH2gOasTEkLGQCkiBLlla0SUJEO3ZXevf6maXwfU+DoZs5cy9q8klCDTh99oRTAxSrUZtyWDJ5abTybNnT1XP6bCwMHWzKrNN+yA6jVu0Uusizx++p4vhkeCDUggi4e4sejlZhsBFBUPjMDPel19+qRKKP3r0SNq2te+yicAQlkG+JuR9wgx6GM6XK1cul7YLs/1hxr6TJ09KtWrVVE6qgQMHStas5go8thVTwi5cuFBtO3pMYQa/hAjD6vSpcto9Z7x+QIL/+0FCdk4U451T4cs+vSPGi1vFJ3/UiaQd0Z5cVw3QkF0/q88NPbpItLDwJPC4I2u6czIWvk3igm61pcqUt3tu+bLFUqdaOWnVtK78t3mD9fmeX/RTlbRSBXNKyQKBkjd/QXmjvjkPy4MH9+XbwV/LiNE/OVxPmfIVZMum9a+0rR917yUGvUFqViolxfNllyuXoy4LarxeR3p+3k9tU6PXq6gu+G++3cZumS++GqyWmTD2B6lZsaR0aWeuhOfMaX8ux4RlQTjT7ZN2PSbBeHWvBG/4RkK2/ijG2yciv7Z+qASv6S8+eeqo3k6AiigaqRa6ZOlEe3YvxvcBKq6mCOsh58qCkmXCA4qnThxXvYdaNa4jDWpWlNEjhlh7JBQtUUqWLlqgGoAY1rtp3Rq57OC3+Y+5s6R1m3Z2z+GuK+6KvorO3T4VvcEgdaqUktIFAlWPx6hUr1VHPvmsn4wY8rU0qVNVDdNt0dp+Gu7P+g2STz7rq2bhq125pHT9gGXBqzDdPRO5TnDjRZ1g18+R6wSXtolPPsdDbk33zknw1lESvOMnMV7f7/Rrqk5wN3w95LztWzZGrhcsXSx1qpaVVk3s6wUXL5yT32ZMlS/7D3X4WTu3b5U3qpWTpm9Ul+V/LbJ7DUP90IP6lesFBoP6DS+eN1sM9YI3pOcXX8m3g76SRrVQL3gUuV7w9WBV18HsvDUrlGC94BXht1gfIdhsvLJbgtcOlpDNo8R467jD9xlvHhUtBCMwnGubme6cluC1QyR4wwgxXrYflYJeWqab9qMvKGY7/tssJUo7vsm48PfZ0srmt/3dth0leYqUUr5wTqlQJEgePXwgbV/MaA0Xz5+TJq9XkmZvVJXfpv9i91mly1aQba9YJ6C4pdfrnHp4C50WcSA3uQw5pTAjoX/tEaLzSeIxe/D56s/Fv+Zg0fmZZxTUgh+K+CYXnd4gpntnJWTfDPGv1EskSRoJ3fWz+ORvpHpEhV3eKaabhyLljnL0fMieaaIFPxC/cl1FDP4Seuh30fmnFN+CTdXrpic3JWTHBElSy3HFKL6cntdVPFnuzAGy6/AZNe4bbt64LmnSplNDXndt3yqd2raWFWu3qJ4JyDV18vhRGfrdGJUMtH2bN9VQv3fbdpBPPvxA6jZoIo2atpCtWzbK4K++sOaUAvSswrpOX72vZqb0RBi+Vygoozx48EACAgLEK8qC+mM8apja8+WfiH+d4eratAy7Fb8U5rLgzhkJ2TVZ/F/rq4JMttAwDd05WXzLdhR9iswSvH6Y+JZoYx0GGHZuo2qA+pX+INr3WSqzYSf+Fv9q5iFknuL4dPtt9zT5s6aSbQdPW8uCeq+VkyxZs8ukGfNU8KnTey2lbsMm0q5TV9Uj4dtB/VQvgvTpM0hgzlxy984tNZzP4vKlC1K7cinZfvC0pE5jzvtjKQuwruOX73luWfDooRTLlcm7yoLXh3tMveD5P1+If41BUdQJzknI/hniX/HTyHWCK/jtP2wdkqeFPBYx+InO4KeG34Ts+UX8SrRVAafoXrPWCXZOlCQ1PW+ymNPzzbmvPFXuTCll15GzUdcL3m8lK9b+J9kCc6ggVb9B30jpsuXVcP7VK5ZZh+uhnMDMeOhheerEMXm3RSOZPHOeCkbB2dMn5c2GtWXfiagDSR5RL8iZwbvKgsbjPaZe8Pyvj8S//g8R6gQoC3zEdOeUhGz/Wfxr9rerE5geXJaQrePEr1wX0ae3H9qJ/FGmq/vshuohH50qC3z8xfTwqoT896P4VfhI9GnNQwZNj65LyKbvJUnDMeJpjky2v0HiSQplTy1b9p+ylgMWu7b/Jz06v69eswzpX7Nqufy18HcZPfFX0en18mWPLiqX3Gf9BpuHw2uapAxIJdeuXpaO77SQbr37SMOmb1rrBFjX0Yt3xd+D6wQl82T2inIgLsqUAp8tFoO/fY6+iIzBT+TE6BZesY+8KP0VuczgZ83vZMkNhcon4A6JPmU2MT24JBL2XEyPrqoEqM83DJWwE8vUXZSQnT/HuAokRdRnKKy69uKzDVlKmfNWWJjCRGcw5wUg56HCGPz8ufVvTKtsycFWrmJlKVqshBzYt1f9PfvXX1QQCj9C6OaOseQIQAEqqt8M7CsVS+SXbp3aqgroa+WLWT8X68D7MKSVEjBcgyabsiBJqvCyIF0ec74o2+v2BX2ydOqOqOm6ebg1hvdqz8Jzk2hP71gTo0b3PsUUKjo9zzNXJUFZEBxeFmTNFij1GzdTyUgxRK9eo6ayb/dOa26W0ROmyqqNO+W3RSvU0F4kP7a1cN4cqVOvkV1ACrAOlgWJuU6Qy75O8PiahByYLc83DlPBZFUn2GVOfo2gFoJO6n0pMokhfUEx3T8f42sK6wRxVy8oXkIO7N+rAjYYyvtxh/ekYvH8MmxAX9m4fo20blrXWk5YhvwiB12tOnVl946t1s99/jzYYwPTFEdlgaoTmIfgmXNB5jDnonxBBZW2/iS+pT+IFJCKCgJeCEipz0TeyUzFVM+p8A8NtZYV5GqdIHxEim0P6Bat37MGpGD+nOnyRoMmKqiEej5SAWzfsulF6pwAFZAC3Ohq3OItFdiKWCfABEnk2TmldDE8vAWDUgmYLmVWdVfSQnt+3/pvJB83PbqiZt3CnZskr38rSWoMVA+fAk3Ms+yV/zjGdSBJKrriqwTnqkvwcfXjY13n4xtqO8g1BYsUVTNoWFy9ctn6b+RiQp6lQoWLqr9zBOWS9WvNM+dg3PjGdf9KwUJF1N/bD5y0PiZOm60qoJt2hiejPHXyuBQoVETNfkMJly4gu+q1YGE75M70+Ka6A4pkpurvR9fClwt+pBqjlteQo8544T81C58W8kSMV/eIIVvZGN+nnnt0XXSpnMtlQOEKFi4qZ0+HD3dq+mYr2bR+jeolhRwRm9avVYnK4d7dO9aZbZEQ/Z///S3vt+8SfqxNJvnz99lq1r2ITp88LvlZFiRo+L03Pb3lXJ2g1jeSpPoA9fAp0NhcJ0CP6AjvU9f63dOiS5ktxtfUc6wTvDQkH466XnBKJQcvVLiIBKRKJYfPXpPtB0+qx4Bh30n1mrVlwdLVatlrV69Y34cZOzHsr2jxknZlQeGixV9+Q8nj6VKhTnDd+jcSmFugroBZt611AhWQGqdmyjNkMtctnWFbz0BPLNOt46JLlSP8uUfX1HaQ63WCczblgKXH0Kq/l0jLd+1T46C3NPLKYlAUHuv/XWW9UXXzxjVrrjfkpV33z/+kSLES1veeQZ2gYGG2DzyYLoHllHrp2ffI8xkyF1cNQ0P6Aurv0JMrRHt42TyHpE4vvoXfFH3yjDF+DhqtyDeheloYQ+X5+sHik6e2+OSoqu6uGjIWlZD/RqnPxJTSvkXeCn/v7eNiyBxeyJFz0Ntpw7o1Uq3G6+rv778dJIf27xODj4+6c/HN92Mld17z3SrMyNev9yfyepUyauYr5IZBImRnbFj7jzRsEvXMS5QwqB6MN4+KIUMh9XfosaWiPbgoojOoXyzf4m9bJyswnl0nobibqe6aamLIXUslOlWfE1hBtPvnJWTtQDU9tE+e10X/ouIa3fsA6/cp2Dhevr83a9CkuWxa969UrV7LmnwUAac6VUqrsqBcxSrS/kPzkIn9e3fLkH6fqXIiRYoUMnHab5IxcxbrZ23ZuE514Xc02+bGtf9Kg8YsCxIyJBlXv8np8qu/Q0+ttK8TFHKuToAJTjCcT9DLStPEkLO6GNLli/E1UOvPxIDHy2jYtLlsWPtveL3gG9QL9obXC34YJ7nzmo9tdGZNm6wC1pjpEo3Szl172JUJqBc0eDF7FyVMhmxlxHTjiBgymgMUoUeXiIbe0pY6Qck2ok9pHnofdvB3Ndtu2OFF6gE+Rd8UQ6ai5iF4m0djjJC5fbDyC/Ep0EB88tSUsDPrxXRt/4uywCSGvHXs6wTXD6vtINfUa9xc3Ziq8qJOAMuXLFQ5JTETt62eX3wtX3/WXeq/Zr55iPLhm1Hj1b9XLf9L5s2cJgaDjxiNYVK/cXNp+U54UAv1DqyLPJder48xaKh5UacD5pRKwDmlkHA8ZPs48avY09qF1q3rD3mshgD6Ve5t7RbsKTw9pxQSgDetV0OWrd6ohujEBYwXb1CrsvyxdJXqzu+pmFPq1Wlhz1XyUr9qX8RLWYBeVKEH5op/1c/F03h6Tik1GUCDmrL4fxvitCxoUruKzFvyP88uC5hT6tXrBDt+Er8KPeKvTrBrkvhV6uVxdQJvyCml6gV1q8uyfzbFWVmAfFPIR7Vy/TaPHtbPnFKxUCfYMEL8anwVP2VB8CNznaTWAI8sCzw5pxTKgbca1pI/V66P0zpBszpV5bfFKz2+TpCYc0oV7bvUqZxSh79r6hX7yHvCZ+Qy/ND4FGpmnbrV3ZBvBr2mPPEHx9MlT5FCBn/7vVy8YJOLI5ZdunBeJUL15B8cih0IlvsUbSna09vxskvRjR8J0unlyoIB33yvplaPy7Kgz4BhLAsSQ52gYNP4rRMUbsk6wavUC4b/EKf1gvPnzsp3YyZ4dECKYqlOUPxt0Z7EU53gyS3xLfU+y4KXLAe+HjYyTusEly+ely8GDGWdwMPpxImcUuI94/cYLUjgLN304wNm7aGXZxmuE1fy5MuvHpQ4GDIUjL91vxgiQC/H0XC72MSyIPFgncC7xXW9ALP1UeJgO5TO3Swz+JJn1gkwzM+ZocAUv3RO5IxiTikiIiIiIiIiIopVOidm1/Om2ffYU4qIiIiIiIiIyAvo2FOKiIiIiIiIiIjcTa/XqUd0tBhe9yTsKUVERERERERE5AV0HL5HRERERERERERup3Mikbn3dJRiTykiIiIiIiIiIm+gY08pIiIiIiIiIiJyNx0TnRMRERERERERkbvp2FOKiIiIiIiIiIjcTceeUkRERERERERE5G56vV49YlrGW/jE9wYQEREREREREVHM2FOKiIiIiIiIiIjcTsecUkRERERERERE5G465pQiIiIiIiIiIiJ307GnFBERERERERERuZvuRW+pmJbxFt6Tkp2IiIiIiIiIKBEz6HVOPV7GxIkTJSgoSJIkSSIVKlSQnTt3Rrv8/fv3pVu3bpIlSxbx9/eX/Pnzy8qVK11aJ2ffIyIiIiIiIiJKxMP3FixYIL1795bJkyergNTYsWOlbt26cuLECcmYMWOk5UNCQqROnTrqtT///FOyZcsmFy5ckNSpU7u0XgaliIiIiIiIiIi8gF5nfsS0jKvGjBkjnTt3lvbt26u/EZxasWKFTJ8+Xfr27RtpeTx/9+5d2bp1q/j6+qrn0MvKVRy+R0RERERERETkDXThvaWieliSSj18+NDuERwc7PAj0etpz549Urt2betzer1e/b1t2zaH71m2bJlUqlRJDd/LlCmTFC1aVIYPHy5Go9Glr8OgFBERERERERGRF9DpnHtAYGCgpEqVyvoYMWKEw8+8ffu2CiYhuGQLf1+/ft3he86ePauG7eF9yCM1YMAAGT16tHzzzTcufR8O3yMiIiIiIiIi8gK6F/+LaRm4dOmSBAQEWJ9HMvLYYjKZVD6pKVOmiMFgkDJlysiVK1fkhx9+kEGDBjn9OQxKERERERERERElsJxSAQEBdkGpqKRPn14Flm7cuGH3PP7OnDmzw/dgxj3kksL7LAoVKqR6VmE4oJ+fn3Pfx6mliIiIiIiIiIgoXun1OqcerkAACT2d1q5da9cTCn8jb5QjVapUkdOnT6vlLE6ePKmCVc4GpNT3cWlLiYiIiIiIiIgoXuh1Oqcerurdu7dMnTpVZs2aJceOHZOuXbvKkydPrLPxtW3bVvr162ddHq9j9r2ePXuqYBRm6kOicyQ+dwWH7xEREREREREReQGdTSLz6JZxVevWreXWrVsycOBANQSvZMmSsmrVKmvy84sXL6oZ+SyQRH316tXSq1cvKV68uGTLlk0FqPr06ePSehmUIiIiIiIiIiLyAjqdTj1iWuZldO/eXT0c2bBhQ6TnMLRv+/bt8ioYlCIiIiIiIiIiSsQ9peILg1JERERERERERF5A70TOqJfJKRVfGJQiIiIiIiIiIvICegaliIiIiIiIiIjI3fQ68yOmZbwFe0oRERERERERESXyROfxgUEpIiIiIiIiIiIvofOemFOMGJQiIiIiIiIiIvICOvaUIiIiIiIiIiIid9MzpxQREREREREREbmbjj2liIiIiIiIiIjI3Qw6nXrEtIy3YE4pIiIiIiIiIiIvoNPFnOjci2JSDEoREREREREREXkDHYfvERERERERERGRu+nYU4qIiIiIiIiIiNxNr9OpR0zLeAvmlCIiIiIiIiIi8gI69pSiqBye0VlSBgRwB3mBoOaj4nsTyEla2HOv21enZnwgASwLvELmWl/H9yaQk7SwYK/bV6fmfsSywEuwLPAe3lgWHJ30DtsIXiJHk5HxvQmUQNsHscnA2feIiIiIiIiIiMjddEx0TkRERERERERE8TF8Tx9DyigvSinFnFJERERERERERN5A70RQKqbXPQkTnRMREREREREReQEdh+8REREREREREZG76dlTioiIiIiIiIiI3E2nizlnFHNKERERERERERFRrPLR6dQjOkYvikoxpxQRERERERERkRfQsacUERERERERERG5m150oo+hJxSW8RbsKUVERERERERE5AV07ClFRERERERERETupufse0REREREREREFB89pfQxDN/zojznHL5HREREREREROQNDHrzI6ZlvAVzShEREREREREReQHdi//FtIy3YFCKiIiIiIiIiMgL6JlTioiIiIiIiIiI3E3PoBQREREREREREbmbTqdTj5iW8RZelP6KiIiIiIiIiCjx0uuce7yMiRMnSlBQkCRJkkQqVKggO3fudOp98+fPV4GwZs2aubxOBqWIiIiIiIiIiLyATufcw1ULFiyQ3r17y6BBg2Tv3r1SokQJqVu3rty8eTPa950/f14+//xzqVat2kt9HwaliIiIiIiIiIi8gI9e59TDVWPGjJHOnTtL+/btpXDhwjJ58mRJliyZTJ8+Pcr3GI1GadOmjQwZMkRy5879Ut+HQSkiIiIiIiIiIm+gc6KXlIsxqZCQENmzZ4/Url3b+pxer1d/b9u2Lcr3DR06VDJmzCgdO3Z86a/j89LvJCIiIiIiIiIit9GLTj1iWgYePnxo97y/v796RHT79m3V6ylTpkx2z+Pv48ePO1zHli1b5Ndff5X9+/e/xLew3VYiIiIiIiIiIkpQOaUCAwMlVapU1seIESNiZRsePXok77//vkydOlXSp0//Sp/FnlJERERERERERF5A78TsepbXL126JAEBAdbnHfWSAgSWDAaD3Lhxw+55/J05c+ZIy585c0YlOG/cuLH1OZPJpP7r4+MjJ06ckDx58jj1fRiUIiIiIiIiIiLyAnqdTj1iWgYQkLINSkXFz89PypQpI2vXrpVmzZpZg0z4u3v37pGWL1iwoBw6dMjuuf79+6seVOPGjVM9tJzFoBQRERERERERkRcw6HRiiKGrFJZxVe/evaVdu3ZStmxZKV++vIwdO1aePHmiZuODtm3bSrZs2dQQwCRJkkjRokXt3p86dWr134jPx4RBKSIiIiIiIiIiL6CzyRkV3TKuat26tdy6dUsGDhwo169fl5IlS8qqVausyc8vXryoZuSLbQxKERERERERERF5Ab0TM9a9bOgIQ/UcDdeDDRs2RPvemTNnvtQ6GZQiIiIiIiIiIvICOp1OPWJaxlswKEVERERERERE5AV0Lx4xLeMtGJQiIiIiIiIiIkpgs+95AwaliIiIiIiIiIi8hE4SDgaliIiIiIiIiIi8gF6vU4+YlvEWDEoRERERERERESXy2ffiA4NSREREREREREReQMfZ94iIiIiIiIiIyN10nH2PiIiIiIiIiIjcTceeUkRERERERERE5G565pQiIiIiIiIiIiJ30+t06hHTMt6Cic6JiIiIiIiIiLyATmd+xLSMt2BQioiIiIiIiIjIC+hFpx4xLeMtGJQiIiIiIiIiIvICOvaUIiIiIiIiIiIid9O9+F9My3gL9pRKpM6eOSU9Puood+/clpQBqWTcpGlSsFCRSMvNmz1Dxv/4g5hMJqn6Wg35bsx48fX1ld9/myXTJo+3LnftyhWpWLmqTJ+70M3fJOEzPb0toUf/FAl9IuKTRHwLtRR9ikx2y2iaScJOrxLT3ZMimkn0qXKKT4GmotP7iOnZXQk9PA8LiWhG0SXLKL4Fm4vON2m8fSfyDGdOn5KundvLnTt3JCAglfw85VcpVDhyOTB75nQZO/p7VQ68Vr2mjB43QZUDO3dsk949uqtlwsJCpWKlKjJy9Fjx9/ePh2+T8Klr+cwKkbBnIgZ/8c3TQPTJMtgto2mahF1cL6b758xlQcps4pOrruj0BjHePythFzeGLxv6VHS+ycW/+Afx8G3Ik7As8C4sCyiunD19Srq/aB+gXvDTZMftg7mzZ8hPY8ztg2rVa8hIm/bBlEkR2gdVqspMtg/ihOnpHQk9vkgk9Km5jVCwueiTO2gjnPlHTHdPvWgj5BCf/I1ftBHuSeiR+ep5PHTJM4hv/qZsI3gBXQLrKYXZBCkR+qJnN3nvg06yde9R6f7p59Kza6dIy1w4f05GfjtYlq5aJ9v3H5Nbt27InJnT1GvvvNdO1m7ZbX1kyJRJWrR6Jx6+ScIXdvwvMWQrJ/6VPhOfnK9J6LE/Iy1jvLpHtEdXxa9cd/Gr0EvFxo2XtqrXdP4B4lf6Q/Ev/4n4V/hUdP4pJezcmnj4JuRpPv2kq7Tr0Fn2HDwmn372hXzcpWOkZc6fPyfDhw6S//27QfYdPiE3b96Qmb9OVa8VLVZC1m/ZLlt27JGtu/bLrVs3ZdqUSfHwTRKHsHOrxZCxpPiX7CI+WStI6JmVkZYx3jwg2pMb4lfsA/Er0UnVSIzXd6vXDKlzi3/x9tYHKq6G9IXj4ZuQp2FZ4F1YFlBc+fzTbtK2fSfZvu+ofNLrc+nxkeP2wXffDJZlq9fJzgPH5NbNGzJ7Rnj7YP1/u62PjJkyyZtsH8SZsJNLxZClrKrf+wRWldDjSyItY7y2V7THV8WvbFfxK9/DXC+4vF29hjaBX6lO4l+um2on6PxSStj5dXG3wRRr9DqdGGJ4eNPsewxKJUJoOB7Yv0datn5X/d2oaQu5euWynDtz2m655UsXS936jSRjpsyi0+mkbYcu8tefCyJ93t7dO+X2rVtSt0Fjt32HxEILeSymR1fEkKmk+lufoahowQ/UnRG75R5fE33avOquB46VPl1+MV7fr15Tzxl8zcvhTogxVAWtKHG7dfOm7N+7R1q/00b93aRZC7ly+ZKcjVAOLFuySOo3bCyZMpvLgQ6dusiihfPVa8mSJVN3RiEkJESeP3umlqHYp4U+EdOT62LIYL5jrU9bQLSQR2J6fs9+uae3RJ8qSPWMUmVB6txivH0k8ufhvQ8uWD+PEi+WBd6FZQHFZftg/z779sGVK5cj1Qv+RvugQSPJ9KJ90K5DF1nioH2wZ5e5fVCP7YM4bCNcFUOmEupvfYYioj131Ea4Lvo0ecLbCGnzifEG2wgJpaeULoaHt2BQKhG6evmy+iHx8TGP3kQBlS17oGqQ2sLf2QNzWP8OzJEz0jKWIX4t325jbZxS7MGPC+5ioIFpOVY6/9SiPb9vt5wuZTYx3j4mWthz0UxGMd48JJpNY1UzhUnwzvESvPkbMT27LT65a/MwJXK4ljNlzmJXDmQPDJRLly7aLXfp0iUJzBFeDuTIGSSXL4WXAxcunJcqFUpLnsBMEpAqlXTq0tWN3yLx0IIfic43heh0+vCywC9AtOCHdsvpkmcS471TooUFm8uCO8dVIDsi463Dok+TWw3fo8SNZYF3YVlA7mwfZHfUPkC9wLZ9kDOK9sEctg/iEn7bdX4p7NsISVJF+s3XpcwqxtvHbdoIh+3aEaqNsGuiBP83QkzP7ohPrlpxut0UuzmldDH8z1swKEWv5MmTJ/LX4j/k3feZkyQ+GbKUVnc+QvZOVQ9dsvToImV9HXdH1PC9ql+pHDTGKzvjdXsp4ciZM0j+27FXTpy7IsHBwfL30shdx8l9DBmKiT5Vbgk5Ok89dEnSRvqpR94p482DYshQnIeGYg3LAs/CsoDiu32wZNEf0qYt2wfxzZC5lLmNsP9X9XDYRsDwvcp9RJ8svRiv7orX7SXn6HXOPbyF1wWlnj59Km+++aYEBASoiPD9+/Y9RhzZsGGD08smBlmzZ5cbN65LWFiYtYGCOxzoLWULf1+26TVx6eKFSMv8/dciKVCwsHpQ7DPf8Xik7mxYjpUWfF90SVLbL6fTiW/u2ubAU9mPRJ88o+iSZ4z8eXofMWQpI8br+7z+cLEseDW4lm9cv2ZXDlyOcPcTAtF76mJ4OXDxwnnVoyqiFClSyJtvtZY/5s97xS0jR9BjUgt9bB6CaykLQh6qnHGRyoLAqua8UUXfF33SdOYKqA3Tw0uiaWGiT50rQexslgWvhmWBd2FZ4BjLgbhpH1x21D6I0Kv60gW2D+KDzj+VGsJn10ZQIyxSRa4X5Kol/mW7iX/pLurmdJRthMylrUP7yLPpEnNPqRo1asinn34q8WnWrFmyefNm2bp1q1y7dk1SpbK/8ChmGTJklGIlSsmfC+ZZc0dlyZpNcuXJa7dcoybNZfX/lsvNG9dVQTd7+hRp+mYru2V+nzND3n2/PXd7HFHdctHt9sUPhOnWYfVjo0+Wzm45zRgqWugz879DnkjYhU0qKbr6+9k90Ywh5n9rJtVtV5ci8yttV/DeqRJ6crnEJ5YFryZDxoxSvGQpWfD7XPX3sr8WS9Zs2SV3hHIAuab+t+JvuXHdXA5MnzZFWrRsrV5DnonQ0FBrTqnly/6SokWLveKWkSMYZqdLlkmMt8z5oUx3T6iEpPokaeyWQzd8dNFX/w59KmFXd6ik6LaMt9BLqph1KOCrCD4yT0LPx+/ECSwLXg3LAu/iiWUBy4GE0z4oHqF9kDVbtkj1AtU+WLlcBbBQL5g1fYo0j9A+wOx8bdqyfRDnbYQUWcR444D623TriLpRFWMb4eJmlRRd/f38vn0bAe2M5K/WRiD30CWwnFLmQcOxSA0LMBqt45Fj25kzZ6RQoUJStGjROPn8xOKHsRPVjHs/jf5OUqQMkHE/m2fT6t39Q5W8EEnLc+bKLV/0GyiN36ihXqtc9TVp276z9TNOnzohhw8dkLkLl8Xb90gMfAs0UzPuGc9vME/3WuhN9XzoscWiT19IDBkKiRifS8jeaebSR9PEEFhZDOkLqeWQHDns4L8vPs2k8k/55o/7pPQoC9T0si/Gusc2lgWvbuz4SfJxlw4y5ofvJGXKAJn4i3n2nE+6dlHJzRs0aixBuXJLv/6DpO7r5iBn1WrVpX2nLurfmzasl18mTRC9wSDGsDB5rUYt+aJf/1jYMnLEN3ddNeOe8eo2EYO/+OZpoJ4PPfM/0afJK4a0+UTCgtWwPWtZkKWsGNKENyiQa8p096T4Fe/gtp2sygLRYiUI5gjLglfHssC7eGNZwHLAO4waN1E++aiTjB31naQMCG8f9HrRPkDSctQLvvxqoDSqY24fVKn2mrTtELl98HsLtg/imm+BJmrGPePFTeayoGBz9Xzo8b9En76AuS1gDFbD9syTHGliyFZJDOkLquVMj6+Hz8itaaJLmUV88zWM8+2mV6d70VsqpmW8hU4z/0rE6IMPPlB3I22dO3dOzp8/LzVr1pSVK1dK//795dChQ/LPP//IzJkz1XC5v/76y7o8elnt379fDacDk8kkI0eOlClTpsj169clf/78MmDAAGnZsmWUPbU2btxo/bt69erqs+bMmSPjxo2TEydOSPLkyaVWrVoyduxYyZjR3DURy2Ab7927J6lTp7Z28X348KGsWLFCPTdt2jQZPXq0+k5BQUHSo0cP+fjjj53aifgc9Ng6dem2KsDJ8wU1HxXfm+C1Qo7+Kabre+2e86v0hUqsHrpvmviWaCdhZ/8V7fEN8S3ZXk1FK2HPxK/4+9bl0cvK9Pia+JfuHH535sImNY4ds4JhuJFPUE0xZCym7vQGbxoqDx48UMN2vaEsuHj9rnVbybNlrvV1fG+C1wo5vUJMtw/bPedX8iOVZDX02O/iW6ClhF3erGYE9C3YWoy3DqnKsV+BFtbl0cvK9OSm+Bd5N/zG1tXtYrx5QN3R1SVNIz7ZKoshXUHViA7ePZZlAcUJlgXeUQ6o1yOUBZ5aJ7CtF5y5zDaCt8jRZGR8bwI5QbUPtnxrVydIDB6+KFP+t+e8JE8R/fd+8vih1C8T5BX7yOnuTCjUT548qXooDR06VD2XIUMGFZSCvn37yqhRoyR37tySJo19F+KojBgxQn777TeZPHmy5MuXTzZt2iTvvfee+lz8oES0ePFitZ7Dhw+rf/v5+annMXxk2LBhUqBAAbl586b07t1bBdEQKIsIgbKGDRuq/Cf//vuvmtJ87ty5MnDgQJkwYYKUKlVK9u3bJ507d1Y/YO3atYv0GUjmi4ftyUGUWPjmbyQhT2+LPkUm8cn1YhY/v+TW2f7CzqwWn7z1VYJlnW9SMY90j57xwkYxXt8vPgWaqoCU6f45CT260DxMIWXWSMuzLCCKf75BtSXk+V2Vn8Inu3kogPgms878E3Zpo/jkqKly4Ol8kojxVsyfiZ4fGJbkk+sNVYYg/1Xo6eWi802mhixFxLKAKHGVA/oA+7yHnlQOANsIROQOOieG5yXI4XuIyKGQRyGdOXPksaYIVNWpU8fpFaPQHj58uKxZs0YqVaqknkNAa8uWLfLLL784DEqlTZtWrR/bYbsNHTqEdz3GZ/z0009Srlw5efz4sfpxsUBvrNatW6sA2Lx586w/WoMGDVJ3QVq0MN+1yZUrlxw9elRth6MfHQTThgwZ4vR3JUpIUKkUDMnT+6qEqxEhUKWGDjgJeS/Czm8Qv1IdRZ/KXNnUJ00r2v0LEnZ1pxq+GBHLAqL4p/PxF9GhLPBRuS0i8sleTQwuJFNXZcGV7eJXqLXoU2ZTz+mTpBbt0WUJu7FffHPVjfQelgVEiasc8HMQlPKUcgDYRiAi9w3fi54XxaRiL6dU2bJlXVr+9OnTqptsxEAWkuXiboQr9uzZI4MHD5YDBw6oLrgYFggXL16UwoXDZ4XDusqXLy8LFiwQg8FgnbIU+Sg6duyo7n5YYOaJqJKo9+vXT91tse0phRmqiEhEH2CuRDpLe3pHxBQqIfun279gMqqx7a5gWUDkOfQuTqigeluiLDi2IMILRoe9pKLDsoDIMySmcgDYRiAid9CLTvQxdIXCMokuKIWurLb0ev2LpIbhLLM0Ae5SAMZsZ8tm34j19/d3er340ahbt656oJsthv7hxwZ/I8BlC91yFy1apO5yFCtWzG47pk6dKhUq2M9KYvlhigjb58o2xqdObd+Wj7p/KmXLV5R/V6+UH74dIsePHZF2HT+UYd+Nti43bfIEmTNzmpo2FI9uPT+Tlq3bqNeePXsmX3z6sRw6sE/9nTMol4yZMEXSp8+gpofv+XEnOXxwv+TIGSRrt+y2fubRwwdl2KCv5fdFf8fDN/cOIYfmiU+OqqqHkPH2cZVsELmYDNkqqGFytjBrXti5tSpJIfgWbyf6pGkk9MRSMT0In5oXeRt88tQTn8DKYnp4WeVv0h5fE33avHZ5nTDbRtiJZWJ6dEVV9vTpC4tPnrrq+CPfU9jp1eJX8gPXv5TefIfRAp8XKXGdZjOo78WsH37F24pEnMbWhSTpLAuc165Na+nW41MpX6GSjPhmiEybMkmyZDGXw4UKF5apM+ZYl13612IZ+e1Qa3k+f9FSyZkzSD7/9BPZvm2rdblTJ4/LkG+/k48+/kRWrVwu/1uxXMZNnOzCViVuISf/Ep8s5VTPAOO9My9ysNwWQ6aSaniMRei5f8zX7Avaszvik6OG+GQJvzGkoUF3aJbqveBf3Dz7Ed6D95oXMIo+ZXbxCaqtpoCO9rUnNyXs4gbxK2Q/s5JT9L4OyoIIpYFmbiQqRnMdwa9gSxE/+16YOvTEcBLLgtgvC6K73lFH+PSTrnJwf3gdYcKkqZI+QwY5fOigDO7fT/5cusKFrUrc4rosMD64IGEXN4qYzL+9+tR51PvUb//TW+ayIPQpfoBFnyKL+OSqIzq9r+rBFHLkN/Er9I65d5SzElE54E1thI5oH3T7VMpVqCj/rlopI4cPkeNHj8gHHT+Ub0aGtw+ie+3mzRvyZa/ucu7sGQkLDVWTIX3YrYd6bf7c2fJ1n96qbQCpU6eRJSvMk+78878Vanbv0T9Ncvv39hYhR+aLT/bK5vbBnRMSdm6daE9uiCFrefHNZ57YwAJDXcPOr7M0D8S32HuqfYB8bKEnlqgZ9lRQN2V28c3fRHQG8zWJUQrGFzli9RmLiW9uc0eRsGt7Jez0StG9mNETqTj8Spp7HRpvnxDTnePiW6CpO3cHOZCoe0qhOytm1nMGCn+M7baFJOe+vuYLAXcoUGjjB8LRUD1nHT9+XO7cuSPfffedtbfS7t3hgRFbWAbddV9//XWV3BDbkClTJsmaNaucPXtW2rQxB2ESir17dsn9e/dUQAowpeuPE6fI338tUj/WtgoULCx/r94oAalSyZXLl6ROtfJStlxFCcqdR+bMmCrPnj6VDdv2qcrEZ598JD+PGy0Dh5ln7Orbf4g8fPhAvhs20O4zCxctLv7+frJl43qpWr2mW7+7N0COBAl7ah2yhlxKvgXfFONNJAG1rzCZHl2VsLP/iF+pTmq6VyT5tAwUtv1h0IIfSfC2H1SCcPWZfinFN39D9X7TnZN2n4kfI1QM/cr3UBXC0INzxHTrsHovKqIYome8e0YMafNE2nZzxdCmEhkd3+Qq0Gb3fR5fM3f3x2clz6i6/WvPH4ghTe5Ib7dMaR0TlgXO2bNrp9y7d1c1Qi3eav2ufPfDmEjLHti/T74ZPECWrfxXsmTNKo8ePbJWxkeNHW9d7sb161KicF5p3uIt9Xe9Bo1kxDdD5czpU5Inr/NDORMr0+OrImHPrUNVUBH0zd1AjHePRyoLfHO9Yf23FvJYgvf9Yk3+a4EGJz4Ls+pY6JJlFL+ibVWgFwHG0JNLxHhjn2r8Rvea/sX1iYasIVXOSNuuAsfOzZci4pNMNa7tvvuTm6rxa97GdKpc0IIfisHBEB1V7jmBZUHslwXRXe8zfp2i6ghbd+1XdYQeH38oP40dJUO/HSlFixUXP39/2bhhnVSvUcvJLUu83FIWYBbffE3MQ+IQaDo2X00D74N6g84gvkF11HWPCUhCT/0tYVd2iG9gVRWkNqQvImHXdopvYLV4KwecxXIgant3m9sHCEhZ2gdj0T5YErl9EN1rA/t9IQUKFJKZcxeq1xq9UV3KV6wspcqYA6NVqlWX2b8virT+N+o3lO9HDJWzp09JbtYRIsENZQl9Ft4+SJpOfAs2E+PNI1G0D9aIX8n2kdoHYRc2qvf6FXvPfD0fnKOCUD7ZKojp/nkx3jwofuW6q2svZO9UMabKIYZ0BdR79alziV+xyO1iQ/oCKgBmenpH9LhWKf7oElZUyqX5mTHrxI4dO1Ry89u3b1u7wTqCGS4QHJo9e7acOnVKjcu2DVKlTJlSPv/8c+nVq5ea1Q9dZPfu3Svjx4+PNMtfdHLkyKGCZXgfAkvLli1TSQ2jgmTsCD5h+/CDBcgPhTHgGGuOZO6YQXDGjBkyZkzkipk3mTN9qrR4623r33ny5pcixUqIwSdyLLJajVoqIAXZsgdKhkyZ5MqVy+pvVDJxJxQ93dBt+cmTx5I1W3b1Wpq0aaVCpSqSLJl9TzmLZm+2ltkzzNPJkj3jlZ2iz1TS+rc+WXrRY7iag2nTwy5uEQMqhf7mmRNwl1Jn8Iv8mdf3ij5tPmuuJ12SVKIPCFSNyojQewq5n1TvOL1B9aQyXjff6QZDphJivLrT4WHDTDimB5fF9OyeuhODH7uo6NPkFu3RFTULn+npbQk9a+4NZv0sH3/13UJPr3ixzB119zfs0lbzzH1OYlngnBm/TpWWrd5xatmJP/0o3T75VAWkLOU28nZE9Pvc2VKr9huSySaXR/M3W8rsmZiCmGJivLFf9Ji2+QXkVFPBoBh+otGQRMXRNo+L8cF5NYOlPl0Ru2VxZ9Ta8xA9FU1hTr0GhnSF1DY6ovNPpYLMpucPRAt9GqmHtC1UsLUn19R2m57dldBLm0V7Fp71WGfwV3eBQy+sUzN0mZ7fE9OT6xJ2fY95xi4nsSyI/bIguutd1RGePnVYR4CWrd6Wmb+yHuApZYE+eSYVkFLHTu8j+mSZrEnJw9eH44qeUpmtr1nLAsyIF+E6ZzngXVAvt2sf5MsvRaNoH0T3GkZEvF63vnXETKXK1WTh/LlObUOT5i3lt9kR0jaQgpmo9ZmK27cPUkTRPrj0nxgCKztuHyAgYQw219FNRtVz0rIcboCj9yWWVQHnLKXFeOOgU0fAkLGIGK857gBC7qPTmYfvRffAMgkyKIUgEu6So4eRpRtsVNA9dsCAAfLll1+qpIK4w962bVu7ZRA8wjIICBUqVEjq1aunhvMhkaCzsB0zZ86UhQsXqu1CbygEnqLz448/SqtWrVRgCkGoTp06qSlfEYhCt1303MJnurIdnmjrlk1Sqmw5l9+3af1aeXD/vpQsbb7T8X77zqqHWdG82aRY3uwqh1aHLs5Nh4teWugpRZGZ7p0TfUB4xT062pObqvtt8N4pErxzvISe/ddhIMh4bY8YsjqX302HoQE3D6s7pbizYrx1VLRn9+0akKZ7Zxy+15CjmroTE7JjrJqOFb2comJIl18MQTUl9MwqCdn9s/qBNGSxzxvnk7uO+ATVlLALG9RnhuyfKaY7J6xdh53BssA5WzZvlLLlyts9t3Txn1KlQmlpXL+2bLK5Xo8fOyaXL12UBm/UlGoVy8o3QwY67C372+yZ8n4789AQC9yB3bie176zvSb1KSLPMhkTBGoMGYvb9SrENWTbg8JuPc8fSPDB6RK8+ycRBIAylXbqNdXT4uEFh59pyFLeXBYcnCbBe8ar3g1RMaTOLYZslSX04gYJOTxb3fE1pC8aKSkypn5HouOQA9Mk5NhCVQ7p/M2NaGewLIj9siC66719xy6SImVKyZszi+QLyioPHzyQLh91sy8LNqxz+vglZu4qC6zLhTwW490TYkgTuUc0hvijJ4XtxCUIeqEBqz2z7+nEcsC7/Ld5k5R+ifZBRMVLlpbFf8xXnRRu374l69f+I5cummdlh53btkrNKmWlQe3XZNmSP+3eW658RdnM9oFDmIEaw+idgZQdqIMH7/tVgndPlFCkAXnRPvDJWUO0Z3cleOv3Erz1OzVDpuFF0BvBZsyIaYH6tm0AGmlBgndNVO0OtBVsYQZM072zTm0fxX1HKV0MD2+h06K7rUlOQZAGSQ9PXbotKQPMEWhPkCNDCtl77JzK/WTrhxFDVaXRNqeUxbEjh6TNW01l0q9zVA8oWL3yb3XnY8KUmSpXWM+uHVXOiL4Dhlrf99/mjTKw32d2OaUA4/axHedvPJQkSZKIpwhqHn3g0h2erx8g/lX6RJqtBj2J0H3fNqdU8I6fVK8n36LvqpxSoQdniz5DEfHJXsnuRyzk8HzzZ0a4mxJ2bY+Ybh2NkFMqVMLOrFJdeMXHX/0A4kfGv3x38+umMAneMFD8qw+xjj+PD6hcB28aKg8ePJAAD7q+oisLLl6/67HbmjF1Mjl66oLK92IZipM2XTo1tHr7tv/kvbdbyrrN2yVHjpxSuXwpyZ4tu8ya94eqdL7zVjNp1LiZdOka3uDc+t9m6fD+u3Lk1Hm7PBunT52UBnVqysnz4TlPPFHmWl/H9ybI8x2jxL/0x2q6c1uhl7aIGJ/b5ZGxbbyGnFom/qW7Wq/3kNN/iyFNPjWEx/jgooRdWGvNIxOxsYnp1bGcIX3hGF/TTEYJ3jlK/Mt/phqk8QXB8+DdY1kWxENZEN31vnL537Lg99/kl19nqTrCx106SFBQLuk/eJi1HoB1Xb/72KPqAYm9LMD1hKF7aKT6ILhs+5rJqIbxotEacZ3Bh+eIT/YqKsgcX7yxLDhz2XPaCNnTp5D9xyO3D74fbm4f2OaNiu61O3duy+Cv+8ihg/slfYaMqm1w5/YtNZwPryVNmkz1rj554pi0atZQps36XcqWN+fnOnPqpDSp/7ocOX1JPE2OJiPjdf3PNw4W/0pfiM7PfhRK6Ll15vaBTU6p4F0TVE9F3yLo+aZJ6KG5qrelT/aKaugtbmr75GuocrWFHp4r+ozFxSdrWQk59JsKZBte9Mgy3jkpYRc3iX+pTmoEhKAHtcFPDa0NOThL/Aq/LfpU5jQ5GPUQsu9XSVKlj8Qn1T7Y8q1XlANxUaasO3BRUqSM/ns/fvRQapXI4RX7yKWeUuRd8GMQ/Ny5fDxw4vhRea91c/lxwhRrQArmzPxVGjRuqiqTGCrZotU7KgjlDKwfFVfL9LpkA4GeCMNkoqKG4WUoYh5mY/BT/zY9sP8hD7u6R3W/jRiQivIzDb7im7+x+Jf/RPxLdxGdb3JzficLbBs+y4Vk4+T5kiZLJs+Dw8sFDMGx5PqrWKmKFC9RUvbt3aP+DsweKI2btZCkSZOqrvmNmzSX3bt22H3enFkz5J333o+U+PX58+eSJGlSt3wnr4dAj5NlgUUYejBkKGp3vauJDS6sl+d7J0no6WWqN0Pw/sjDplCGqMbq7aPOvaa2TWfNA0eJryyI7nqfNX2qNGrSzFpHeKv1O7J50wa7soD1AM8qCzRjsIQc/0MFrhwGpE4tVQ1in5yvR14hZsaNx+A0xc6170r7ICrp0qWX8ZN/lQ1b98ifS/+nhgoVLFjY+ppluH/+AoXk9Tfqyc4d4ZMloOxB3YJesX3gj/ZBYZv2QWFzztoXaUKQigNlA4b1qbbD/XPW96kE6DazXuI59ZpfcusQQAznNaTNL6aHNqOjTGGiM7AMiG86J//nLRiUSsAKFS2m7kQ4A3cx0ENq1NifpXot+7tiuPOxYR26g2rqsWb1/6RgIfscBVHBDD0FCxdRd0/Jni5FZnW3wRn4UTHdPaW65KLCaLp72m6aZdwtUEnKs5RxejfjPegVAcjvgjsqmAnQ+vqTW6JLnsnpIBd5hyJFi8npkyesf1+5bM4dB0hMfujgASlSxDykqmXrd2T92n9VLynkilm39l+VuNj2bs2yJYvkvbaRe+OcPHHcblmKGhKNm57fdamXgAlDbjKYJzSwSFK6q/Xhm7eJ6JKmF/+S5qnMkZ8JZYd6v8koxrun1Hpjek099+yO6JJl8KrcBBS7ZUF013vOXLlV2WCpI/yz6n9SqHD4+04eP6b+Zj3AM8oC/O5jWCx6OmF2L7vPQzLk08tUMnSfXPUiXfOqDhJ8T5UH5L0KFymmejO/qrt37lhnVscM3f9bvkzad/5I/X3t6hW7Wfq2bNwgxYqH51E9deK4FC5qf96SmS65K+2D4qpNYN8+yGL+nKRp1O85mF87Zc0Zp89QVOWwQ3mAkRHI4WoZAmw7DB9DfE33z4ruxWeq556ifRDeBqH4odM59/AWDHMmYI2atlANytdqmu90bd6wTnp07SiPHj1UFcflSxfLd6N/kroNGkv/L3vLI3TLHfyVekD/wcOlZu035PO+A+Tznl2lekXzj0m+fAXk+3E/q38/ffpUqpQpIsHBwfLo4QMpVSiXtGz9rnw9+Fv1+ro1q9V2UGS4s4kfCEPavOpv493TEnrsTxE1w5SmEpf65m8qhgyFVMJDzLARsmOcKmH0qYNUYkMLJCfUpcyqkiHaMj25JSH7fzVPsWwKlef/fafGmKNbL8aZhx7+3dwbCjPu5Gsg+pThuSxMd0+KIYNzwUfyHk2bvylr1/wjNV4En4cNGSAH9u0Vg8FH9WYY9eN4yZsvv3rtzbday4H9e6VimeLqtUqVq8pHL6Z7hsULF0iJUqUdzrC35p/Val0UM8x2g7uXhlRB1gTFoWdWmGfZ0VA2nBDfoDesuV2Md46pCiGSEjvL9OCCGK/vMddQNJPoUwVZG6TRvWZ+/awY0ppn5KHEWRZEd733+3qg9Oz+kVQqW0L9nS9/ARk7Pnyq9zX/oixgPcBTygLj9d3mCQdMoWK8e9K6XuRyM905pn77EXQKOTTTmlPOkptKe3RZ9MmziM6HPVy8WeMX7YPqL9oHmzask08+Cm8f/L10sYwc85PUa9A42tf27dklX33ZW3x8DJIiRUqZNmueZMpsDl5MnzpJVq34W3x8fdWNrQ+79ZBqNjNxr1vzj9oOigx1b9M9tA/Mud6M985I6LHFKieruX1wRKX4wNBbfcZiYnp0TUJ2YYZUvehT5xRDdvOsij55G0rYyWUSjNc0TU18ZHjx225Ik0u0jMUkZNcE9bc+Y1E1sx7gJrXp9rEXvaM19R7b2bFV24XtgwQ9+d7EiRPlhx9+kOuYbbdECTWhXPny9r1qLaZOnaomtrNMaFemTBkZPnx4lMtHua3MKZVwc0o9efxYGr3xmiz/d7MaeuNuyCNRt0ZF+fPvf1Q3Xk/iCTmlVD6HPb+IX9mPHM6kF5/UNNG7fha/Uh0jjWl3+7Ywp1Ssevz4sbxRq5r8u35LnJULd27flsYN6siGLTs8fuiuJ+SRUT0XjvwmfkXe88CywCghh2eJX6G3I+W5cfu2eGEemcReFqAeUKNqBfl75b+SLr1n1QMiYlkQs5BTS8WQobgYUsfvREDeWBZ4Uk4pXPsN67wmK9fET/sA+aZaNHpD/t243SPrCPGdU0q1D/ZNFT+k1vC0OkHIEwk5MEP8ynwU78N4E3tOqY2HLjmVU6p6sUCX9tGCBQvU5HSTJ0+WChUqyNixY9WEcidOnJCMGW3SvLzQpk0bqVKlilSuXFkN4x85cqQsWbJEjhw5ItmyZXP6e3FcTgKWPEUKGTp8lFy8YB4/7G5YL3pMeVpAylNgfLdPvgaqx5Kn0Z7dE588deM9IEWxDzNpjhg5Si6cj7ty4dzZM/LjTxM9srLpiVDp9MlZy27mG0+hBd8Xn8Dq8R6QIu8sC/DZg4d+6/EBKU/h0WWBKUz0KXPEe0CKYufaHzYi/toH58+elR/Gso4Qbfsgb31VF/c02vO74pu/SbwHpEhEj5EzTjxcNWbMGOncubO0b99eChcurIJTyA83ffp0h8vPnTtXPv74YylZsqQULFhQpk2bpnpHrl271qX18oxK4KrVqBVv686br4B6UNQsQ/c8jT55BhE8KEGydNmPK5bZdch5luE6nkafNJ0IHpQgxXVZgKF8eJD3lwVohPpkLhXfm0Gx5LV4bB+UKefasJ7EyJDGPHTP02AIIHnf8L2HD8PzhIG/v796OOrdvGfPHunXr5/1OeSDrF27tmzbts2p7UJqH+SaS5vW+WHlaj0uLU1ERERERERERPEbldLF8MBs2oGBasif5TFixAiHH3n79m0xGo2SKVMmu+fxN/JLOaNPnz6SNWtWFchyBXtKERERERERERF5Ad2L/8W0DFy6dMkup5SjXlKx4bvvvpP58+fLhg0bVH4pVzAoRURERERERETkBXQ68yOmZQABKWcSnadPn17Nvnvjxg275/F35syZo33vqFGjVFBqzZo1Urx4cXEVh+8RERERERERESWs0XtOwwRFZcqUsUtSbklaXqlSpSjf9/3338uwYcNk1apVUrZsWXkZ7ClFREREREREROQFdDqdesS0jKt69+4t7dq1U8Gl8uXLy9ixY+XJkydqNj5o27atZMuWzZqXauTIkTJw4ECZN2+eBAUFWXNPYZZPPJzFoBQRERERERERUQIbvueK1q1by61bt1SgCQGmkiVLqh5QluTnFy9eVDPyWUyaNEnN2teyZUu7zxk0aJAMHjzY6fUyKEVERERERERE5AV0TgzPe4mYlNK9e3f1cARJzG2dP39eYgODUkREREREREREiT0qFQ8YlCIiIiIiIiIi8gK6F/+LaRlvwaAUEREREREREVEizikVXxiUIiIiIiIiIiLyArqENXqPQSkiIiIiIiIiIm+g0+nUI6ZlvAV7ShEREREREREReQOdE8PzvCcmxaAUEREREREREZE30HH4HhERERERERERuZ0uYUWlOHyPiIiIiIiIiMgL6F78L6ZlvAWDUkREREREREREXkDnRE4pL8pzzqAUEREREREREZE30DEoRURERERERERE7qbj8D0iIiIiIiIiIoqXPOe6mJfxFswpRURERERERETkBXQJa/I9BqWIiIiIiIiIiLyBjjmliIiIiIiIiIjI/XQJqq8Uh+8REREREREREXkBHXtKERERERERERGRu+l15kdMy3gL9pQiIiIiIiIiIvICuhf/i2kZb8GgFBERERERERGRN9AlqJRSDEoREREREREREXkDXcKKSTEoRURERERERETkDXRMdE5ERERERERERO6mY04pIiIiIiIiIiJyNx17ShERERERERERkbvpGJQiIiIiIiIiIqL4GsCXUFKd+8T3BhARERERERERUeLrKaWP7w0gIiIiIiIiIqLEhz2liIiIiIiIiIi8gC6B9ZRiUIqIiIiIiIiIyGsySuliXMZbMChFREREREREROQF9DrzI6ZlvAWDUkRERERERERE3kDnxOR6DEoREREREREREVFs0nH4HkWkaZr676NHj7hzvIQW9jy+N4GcpIUF211n3lEWPIzvTSEXzy/yfJqRZQHF4fnFssBreGdZwDaCt2AbwTt4U/vA2xKdT5w4UX744Qe5fv26lChRQsaPHy/ly5ePcvmFCxfKgAED5Pz585IvXz4ZOXKkNGjQwKV1cvheLLD80JQunCs2Po6IorjOUqVK5RVlQZF8QfG9KUQJFssCIvK2sqBkIbYRiBJrOeBNo/cWLFggvXv3lsmTJ0uFChVk7NixUrduXTlx4oRkzJgx0vJbt26Vd955R0aMGCGNGjWSefPmSbNmzWTv3r1StGhR57+PlljDi7HIZDLJ1atXJWXKlKLzprkXY/Dw4UMJDAyUS5cuSUBAQHxvDiXS44UiCj84WbNmFb1eL56MZQF5ApYF8Y9lAXkClgXxj2UBxTeWAwnveKZKlUqu3b4fY3sPy2ZJn1oePHjgdNsQgahy5crJhAkTrGUY2peffPKJ9O3bN9LyrVu3lidPnsjy5cutz1WsWFFKliypAlvOYk+pWICGcvbs2SWhwkmckIIcCV1CPF7ecgeEZQF5EpYF8YdlAXkSlgXxh2UBeQqWAwnL40ePRB9DZxgsYwlO2fL391ePiEJCQmTPnj3Sr18/uzKsdu3asm3bNofrwPPoWWULPav++usvl74Pg1JERERERERERB7Mz89PMmfOLPlyBTq1fIoUKVRPJ1uDBg2SwYMHR1r29u3bYjQaJVOmTHbP4+/jx487/HzknXK0PJ53BYNSREREREREREQeLEmSJHLu3DnVq8nZNCgR0ws56iUV3xiUoijhhEUk1RNPXIqMx4viCs8t78LjRTy3iGUB8XeGWA4k3MBUkiRJYv1z06dPLwaDQW7cuGH3PP5G7yxH8Lwry0eFic6JiIiIiIiIiBKxChUqSPny5WX8+PHWROc5cuSQ7t27R5no/OnTp/L3339bn6tcubIUL16cic6JiIiIiIiIiMg5SFrerl07KVu2rApOjR07Vs2u1759e/V627ZtJVu2bDJixAj1d8+ePaV69eoyevRoadiwocyfP192794tU6ZMEVdw+B4RERERERERUSLWunVruXXrlgwcOFAlKy9ZsqSsWrXKmsz84sWLakY+215R8+bNk/79+8tXX30l+fLlUzPvFS1a1KX1cvgeERERERERERG5XXiYi8hGUFCQ6q4XHzBDACKsFDvHCVN+Irrtyn7lMSBXzrG4wvPQeSwLKK6xLPAOLAsovs+vuMI6gfNYDpC34fA9ogTs2LFjMmTIEFmyZIlUrFhR0qRJE9+bRETxgGUBEbEsICKWA+SJGJRKwEJCQsTPzy/BrSuhict9d+bMGfXfpk2bqjtMlDixLPAOLAvIm8+x+FxXQsOygLz1/IrPdSU0LAcoMeHwPS9Ro0YNNRUjHqlSpZL06dPLgAEDRNM0u66aw4YNU1nxAwICpEuXLur5LVu2SLVq1SRp0qQSGBgoPXr0UFn0LW7evCmNGzdWr+fKlUvmzp0b4/Z88MEH0qxZM/n2228la9asUqBAAfX8pUuXpFWrVpI6dWpJmzatCoacP3/e+r5du3ZJnTp11PbjeyBb/969eyWh8KTjhGF7WB6QkM4SlHqZYzBo0CDJkiWLHDx40KltpcRxjgHLAs8/TiwLEiZPOseAZYHnHyeWBQmPJ51fwHLA848TywHySBp5herVq2spUqTQevbsqR0/flz77bfftGTJkmlTpkyxLpMzZ04tICBAGzVqlHb69GnrI3ny5NqPP/6onTx5Uvvvv/+0UqVKaR988IH1ffXr19dKlCihbdu2Tdu9e7dWuXJlLWnSpOo9UWnXrp3anvfff187fPiweoSEhGiFChXSOnTooB08eFA7evSo9u6772oFChTQgoOD1fvWrl2rzZkzRzt27Jh6vWPHjlqmTJm0hw8fWj8bp+WSJUs0b+RJx+nRo0fajBkz1P68du2aerh6DEwmk9a9e3ctKChIO3XqlHrNmW2lxHGOAcsCzz9OLAsSJk86x4BlgecfJ5YFCY8nnV/AcsDzjxPLAfJEDEp5UWGGgA+CBBZ9+vRRz9kWZs2aNbN7HwIOXbp0sXtu8+bNml6v1549e6adOHFCBSB27txpfR3BCjwX048OAhmWYBMg0IEAlO024nUUjKtXr3b4OUajUUuZMqX2999/J5iglCcdJ+zHmGLPUR2DhQsXqqAitv3y5ctObyslrnOMZYF3HCeWBQmPp51jLAu84zixLEhYPO38YjngHceJ5QB5GuaU8iJIVG2bF6hSpUoyevRoMRqNYjAY1HNly5a1e8+BAwfUkCvbrpyIOZhMJjl37pycPHlSfHx8pEyZMtbXCxYsqIbfxaRYsWJ248SxrtOnT0vKlCntlnv+/Lk1t9GNGzekf//+smHDBtXdFNv+9OlTuXjxoiQUnnacInL2GPTq1Uv8/f1l+/btqpuxs9taqFAhl7eJvPscY1ngHccpIpYF3s/TzjGWBd5xnCJiWeDdPO38YjngHccpIpYDFJ8YlEpgkidPbvf348eP5cMPP1TjjyPKkSOHKsxic10oFB2NZc6QIYP6b7t27eTOnTsybtw4yZkzpwp6oFBGMr/ExJ3HKSJnjwHyTv3++++yevVqadOmjdPbSp6BZYF3YFlACekcY73g5bEsoLjEcsA7sBygxIpBKS+yY8cOu7/RgyVfvnzW6LojpUuXlqNHj0revHkdvo5oelhYmOzZs0fKlSunnjtx4oTcv3/f5e3DuhYsWCAZM2ZUCfoc+e+//+Tnn3+WBg0aWBOj3759WxISTz9Ozh6DJk2aqMSJ7777rtr2t99+26ltpbjn6ecYywLvOE4sC7yfp59jLAu84zixLPBunn5+sRzwjuPEcoDiE2ff8yIYXtW7d29V2KAHy/jx46Vnz57RvqdPnz6ydetWNdvD/v375dSpU7J06VL1N2DWvHr16qm7pigsUah16tRJzeDgKvSmwTAvzLi3efNm1a0UQ8RwN/by5ctqGRS+c+bMkWPHjqn14T0vsy5P5unHyZVj0Lx5c7Vs+/bt5c8//3RqWynuefo5xrLAO44TywLv5+nnGMsC7zhOLAu8m6efXywHvOM4sRyg+MSglBfBFKHPnj2T8uXLS7du3VRBZpkuNCrFixeXjRs3qu74mE60VKlSMnDgQMmaNat1mRkzZqi/q1evLi1atFCfid5OrkqWLJls2rRJdf/H5yC3UMeOHVVOKUvPqV9//VXu3bunIv/vv/++Cli9zLo8macfJ1ePQcuWLWXWrFlq2cWLFzu1rZS4zzGWBd5xnFgWeD9PP8dYFnjHcWJZ4N08/fxiOeAdx4nlAMUnHbKdx+sWkFNq1KghJUuWlLFjx3KPeTAeJ+I5RiwLiL83xHoBsRwglgNEzmFPKSIiIiIiIiIicjsGpYiIiIiIiIiIyO04fI+IiIiIiIiIiNyOPaWIiIiIiIiIiMjtGJQiIiIiIiIiIiK3Y1CKiIiIiIiIiIjcjkEpIiIiIiIiIiJyOwaliIiIiIiIiIjI7RiUIiIiIiIiIiIit2NQioiIiIiIiIiI3I5BKSIiIiIiIiIicjsGpYiIiIiIiIiISNzt//yK8Y3GXsbfAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 1600x360 with 5 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def aggregate_confusion(run_name: str) -> np.ndarray:\n",
    "    cm = np.zeros((2, 2), dtype=int)\n",
    "    for fold in load_run(run_name).get(\"fold_results\", []):\n",
    "        fold_cm = fold.get(\"test_metrics\", {}).get(\"confusion_matrix\")\n",
    "        if fold_cm is not None:\n",
    "            cm += np.asarray(fold_cm, dtype=int)\n",
    "    return cm\n",
    "\n",
    "\n",
    "conf_rows = []\n",
    "for run, label in RUN_LABELS.items():\n",
    "    cm = aggregate_confusion(run)\n",
    "    tn, fp, fn, tp = cm.ravel()\n",
    "    conf_rows.append({\n",
    "        \"label\": label,\n",
    "        \"run\": run,\n",
    "        \"true_real\": int(tn + fp),\n",
    "        \"true_fake\": int(fn + tp),\n",
    "        \"false_positive_rate\": fp / (tn + fp) if (tn + fp) else np.nan,\n",
    "        \"false_negative_rate\": fn / (fn + tp) if (fn + tp) else np.nan,\n",
    "        \"fake_detection_rate\": tp / (fn + tp) if (fn + tp) else np.nan,\n",
    "    })\n",
    "conf_df = pd.DataFrame(conf_rows).sort_values(\"false_negative_rate\")\n",
    "display(conf_df.style.format({\"false_positive_rate\": \"{:.4f}\", \"false_negative_rate\": \"{:.4f}\", \"fake_detection_rate\": \"{:.4f}\"}))\n",
    "\n",
    "show_runs = [\"p2c_resnet18_facecrop\", \"p3_convnext_tiny\", \"p3_resnet50\", \"p3_mobilenetv3_small\"]\n",
    "fig, axes = plt.subplots(1, len(show_runs), figsize=(4 * len(show_runs), 3.6))\n",
    "for ax, run in zip(axes, show_runs):\n",
    "    label = RUN_LABELS[run]\n",
    "    cm = aggregate_confusion(run)\n",
    "    norm = cm / cm.sum(axis=1, keepdims=True)\n",
    "    im = ax.imshow(norm, cmap=\"Blues\", vmin=0, vmax=1)\n",
    "    ax.set_title(label, fontsize=10)\n",
    "    ax.set_xticks([0, 1])\n",
    "    ax.set_xticklabels([\"pred real\", \"pred fake\"])\n",
    "    ax.set_yticks([0, 1])\n",
    "    ax.set_yticklabels([\"true real\", \"true fake\"])\n",
    "    for i in range(2):\n",
    "        for j in range(2):\n",
    "            cell_text = f\"{norm[i, j]:.2f}\" + \"\\n\" + f\"({cm[i, j]})\"\n",
    "            ax.text(j, i, cell_text, ha=\"center\", va=\"center\", fontsize=8)\n",
    "fig.colorbar(im, ax=axes.ravel().tolist(), shrink=0.8)\n",
    "fig.suptitle(\"Confusion matrices for selected Phase 3 models\", fontsize=13)\n",
    "fig.savefig(FIGURES_DIR / \"06_phase3_confusion_selected.png\", dpi=180, bbox_inches=\"tight\")\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "26c72a45",
   "metadata": {},
   "source": [
    "## 4.1 Balanced class metrics\n",
    "\n",
    "The dataset has more fake images than real images, so ordinary accuracy and binary F1 are influenced more by fake performance. For a detector, that can hide the cost of false alarms on real images. The next cell recomputes two class-balanced metrics from the aggregated confusion matrices:\n",
    "\n",
    "- balanced accuracy: average of real recall and fake recall, so real and fake each contribute 50%.\n",
    "- macro F1: average of F1 for the real class and F1 for the fake class, again giving both classes equal weight.\n",
    "\n",
    "These are not new model runs. They are a different readout of the same saved predictions. This view is mainly used to check that choosing ResNet50 is not just a consequence of the 1 real / 3 fake class mix.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "8319a9b0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style type=\"text/css\">\n",
       "</style>\n",
       "<table id=\"T_3b334\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th class=\"blank level0\" >&nbsp;</th>\n",
       "      <th id=\"T_3b334_level0_col0\" class=\"col_heading level0 col0\" >label</th>\n",
       "      <th id=\"T_3b334_level0_col1\" class=\"col_heading level0 col1\" >run</th>\n",
       "      <th id=\"T_3b334_level0_col2\" class=\"col_heading level0 col2\" >real_recall</th>\n",
       "      <th id=\"T_3b334_level0_col3\" class=\"col_heading level0 col3\" >fake_recall</th>\n",
       "      <th id=\"T_3b334_level0_col4\" class=\"col_heading level0 col4\" >balanced_accuracy</th>\n",
       "      <th id=\"T_3b334_level0_col5\" class=\"col_heading level0 col5\" >real_f1</th>\n",
       "      <th id=\"T_3b334_level0_col6\" class=\"col_heading level0 col6\" >fake_f1</th>\n",
       "      <th id=\"T_3b334_level0_col7\" class=\"col_heading level0 col7\" >macro_f1</th>\n",
       "      <th id=\"T_3b334_level0_col8\" class=\"col_heading level0 col8\" >standard_accuracy</th>\n",
       "      <th id=\"T_3b334_level0_col9\" class=\"col_heading level0 col9\" >standard_f1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th id=\"T_3b334_level0_row0\" class=\"row_heading level0 row0\" >0</th>\n",
       "      <td id=\"T_3b334_row0_col0\" class=\"data row0 col0\" >ConvNeXt-Tiny</td>\n",
       "      <td id=\"T_3b334_row0_col1\" class=\"data row0 col1\" >p3_convnext_tiny</td>\n",
       "      <td id=\"T_3b334_row0_col2\" class=\"data row0 col2\" >0.8838</td>\n",
       "      <td id=\"T_3b334_row0_col3\" class=\"data row0 col3\" >0.9685</td>\n",
       "      <td id=\"T_3b334_row0_col4\" class=\"data row0 col4\" >0.9262</td>\n",
       "      <td id=\"T_3b334_row0_col5\" class=\"data row0 col5\" >0.8935</td>\n",
       "      <td id=\"T_3b334_row0_col6\" class=\"data row0 col6\" >0.9650</td>\n",
       "      <td id=\"T_3b334_row0_col7\" class=\"data row0 col7\" >0.9293</td>\n",
       "      <td id=\"T_3b334_row0_col8\" class=\"data row0 col8\" >0.9473</td>\n",
       "      <td id=\"T_3b334_row0_col9\" class=\"data row0 col9\" >0.9650</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_3b334_level0_row1\" class=\"row_heading level0 row1\" >1</th>\n",
       "      <td id=\"T_3b334_row1_col0\" class=\"data row1 col0\" >ResNet50</td>\n",
       "      <td id=\"T_3b334_row1_col1\" class=\"data row1 col1\" >p3_resnet50</td>\n",
       "      <td id=\"T_3b334_row1_col2\" class=\"data row1 col2\" >0.9092</td>\n",
       "      <td id=\"T_3b334_row1_col3\" class=\"data row1 col3\" >0.9679</td>\n",
       "      <td id=\"T_3b334_row1_col4\" class=\"data row1 col4\" >0.9385</td>\n",
       "      <td id=\"T_3b334_row1_col5\" class=\"data row1 col5\" >0.9067</td>\n",
       "      <td id=\"T_3b334_row1_col6\" class=\"data row1 col6\" >0.9688</td>\n",
       "      <td id=\"T_3b334_row1_col7\" class=\"data row1 col7\" >0.9377</td>\n",
       "      <td id=\"T_3b334_row1_col8\" class=\"data row1 col8\" >0.9532</td>\n",
       "      <td id=\"T_3b334_row1_col9\" class=\"data row1 col9\" >0.9688</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_3b334_level0_row2\" class=\"row_heading level0 row2\" >2</th>\n",
       "      <td id=\"T_3b334_row2_col0\" class=\"data row2 col0\" >EfficientNet-B0</td>\n",
       "      <td id=\"T_3b334_row2_col1\" class=\"data row2 col1\" >p3_efficientnet_b0</td>\n",
       "      <td id=\"T_3b334_row2_col2\" class=\"data row2 col2\" >0.9292</td>\n",
       "      <td id=\"T_3b334_row2_col3\" class=\"data row2 col3\" >0.9503</td>\n",
       "      <td id=\"T_3b334_row2_col4\" class=\"data row2 col4\" >0.9397</td>\n",
       "      <td id=\"T_3b334_row2_col5\" class=\"data row2 col5\" >0.8941</td>\n",
       "      <td id=\"T_3b334_row2_col6\" class=\"data row2 col6\" >0.9628</td>\n",
       "      <td id=\"T_3b334_row2_col7\" class=\"data row2 col7\" >0.9285</td>\n",
       "      <td id=\"T_3b334_row2_col8\" class=\"data row2 col8\" >0.9450</td>\n",
       "      <td id=\"T_3b334_row2_col9\" class=\"data row2 col9\" >0.9628</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_3b334_level0_row3\" class=\"row_heading level0 row3\" >3</th>\n",
       "      <td id=\"T_3b334_row3_col0\" class=\"data row3 col0\" >ResNet34</td>\n",
       "      <td id=\"T_3b334_row3_col1\" class=\"data row3 col1\" >p3_resnet34</td>\n",
       "      <td id=\"T_3b334_row3_col2\" class=\"data row3 col2\" >0.9067</td>\n",
       "      <td id=\"T_3b334_row3_col3\" class=\"data row3 col3\" >0.9384</td>\n",
       "      <td id=\"T_3b334_row3_col4\" class=\"data row3 col4\" >0.9225</td>\n",
       "      <td id=\"T_3b334_row3_col5\" class=\"data row3 col5\" >0.8670</td>\n",
       "      <td id=\"T_3b334_row3_col6\" class=\"data row3 col6\" >0.9529</td>\n",
       "      <td id=\"T_3b334_row3_col7\" class=\"data row3 col7\" >0.9100</td>\n",
       "      <td id=\"T_3b334_row3_col8\" class=\"data row3 col8\" >0.9305</td>\n",
       "      <td id=\"T_3b334_row3_col9\" class=\"data row3 col9\" >0.9529</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_3b334_level0_row4\" class=\"row_heading level0 row4\" >4</th>\n",
       "      <td id=\"T_3b334_row4_col0\" class=\"data row4 col0\" >ResNet18 facecrop reference</td>\n",
       "      <td id=\"T_3b334_row4_col1\" class=\"data row4 col1\" >p2c_resnet18_facecrop</td>\n",
       "      <td id=\"T_3b334_row4_col2\" class=\"data row4 col2\" >0.9027</td>\n",
       "      <td id=\"T_3b334_row4_col3\" class=\"data row4 col3\" >0.9288</td>\n",
       "      <td id=\"T_3b334_row4_col4\" class=\"data row4 col4\" >0.9157</td>\n",
       "      <td id=\"T_3b334_row4_col5\" class=\"data row4 col5\" >0.8531</td>\n",
       "      <td id=\"T_3b334_row4_col6\" class=\"data row4 col6\" >0.9472</td>\n",
       "      <td id=\"T_3b334_row4_col7\" class=\"data row4 col7\" >0.9001</td>\n",
       "      <td id=\"T_3b334_row4_col8\" class=\"data row4 col8\" >0.9223</td>\n",
       "      <td id=\"T_3b334_row4_col9\" class=\"data row4 col9\" >0.9468</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_3b334_level0_row5\" class=\"row_heading level0 row5\" >5</th>\n",
       "      <td id=\"T_3b334_row5_col0\" class=\"data row5 col0\" >MobileNetV3-Small</td>\n",
       "      <td id=\"T_3b334_row5_col1\" class=\"data row5 col1\" >p3_mobilenetv3_small</td>\n",
       "      <td id=\"T_3b334_row5_col2\" class=\"data row5 col2\" >0.8692</td>\n",
       "      <td id=\"T_3b334_row5_col3\" class=\"data row5 col3\" >0.9336</td>\n",
       "      <td id=\"T_3b334_row5_col4\" class=\"data row5 col4\" >0.9014</td>\n",
       "      <td id=\"T_3b334_row5_col5\" class=\"data row5 col5\" >0.8405</td>\n",
       "      <td id=\"T_3b334_row5_col6\" class=\"data row5 col6\" >0.9444</td>\n",
       "      <td id=\"T_3b334_row5_col7\" class=\"data row5 col7\" >0.8924</td>\n",
       "      <td id=\"T_3b334_row5_col8\" class=\"data row5 col8\" >0.9175</td>\n",
       "      <td id=\"T_3b334_row5_col9\" class=\"data row5 col9\" >0.9443</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x1a61d103ad0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x450 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA3kAAAG4CAYAAAD42y7tAAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjEwLjksIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvJkbTWQAAAAlwSFlzAAAPYQAAD2EBqD+naQAAfiBJREFUeJzt3Qm81OP7//G7fdGufdGKVkqbCkURWZKiiBYtKmlDKqkUZWmVtKC0oBBZoiRCSmlBJJWopJUUpX3+j/f1+9/znXOa016znNfz8Tg0cz4zZ5bPfOZz3dd1X3eKQCAQcAAAAACAuJAy0g8AAAAAAHDmEOQBAAAAQBwhyAMAAACAOEKQBwAAAABxhCAPAAAAAOIIQR4AAAAAxBGCPAAAAACIIwR5AAAAABBHCPIAAAAAII4Q5AFAlChSpIhr0aKFi3WvvPKKS5Eihfvtt99cLNN7offkXNFr1rFjRxfv9Dz79et30rfT/qTbav86njVr1rjrrrvOZc2a1W4zY8YMF8nnGC+fCQCxgyAPAE6SP2EL/cmdO7e7+uqr3UcffcTrGWX27t1rJ9zz5s2L9EPBOdK8eXO3YsUK9+STT7rJkye7SpUq8doDSFZSR/oBAECs6t+/vytatKgLBAJu69atFvzVq1fPvf/+++6mm26K9MNDSJD3+OOP279r1ap1wq/Liy++6I4cOcLrGGP+++8/t3DhQvfoo49GLDOqx5A6NadYACKHIxAAnKIbbrghQYagVatWLk+ePO71118nyIthe/bsceedd55LkyZNpB8KTsH27dvt/9myZYvY65c+ffqI/W0AEMo1AeAM0UllhgwZjhrBHzx4sKtevbo7//zz7fcVK1Z0b7311nHv76+//nIPPfSQK1eunMuUKZPLkiWLBZbfffddgu1UhqiS0TfeeMPK0woWLGgnmbVr13Zr16496n4XLVpkGcfs2bNbMHPJJZe4ESNGJNhm1apVrlGjRi5Hjhx2Xwpm33vvvaPu68cff3TXXHONPS/93SeeeOKEs1+a86bntWHDBguK9e8CBQq4UaNG2e9Vbqf71mMsXLiwe+211466j7///tt16dLFFSpUyKVLl86VKFHCPf3008HHoDlQuXLlsn8rm+fLa/18Kf8YfvnlF3tNMmfO7Jo2bZrknDzdr14rvSd6XXTf119/vVuyZElwmzlz5rgrrrjC9gfd98UXX+x69erlTtSrr75qt9H9a1/54osvgr/77LPP7PG/8847R91Or49+pyzW8UqN58+f7zp16mSPX4/zvvvucwcOHLDXs1mzZrZv6Kd79+6WqU4cBD/44IPB11yPVft44u3279/vunbtan9Dr+stt9zifv/997CPa9OmTe7ee++1QRLdZ5kyZdz48ePdydL7qn1FHn74YXuu/j1cv36969Chgz1e7a/6PN5+++1HzZM7E6/R8eYdqpw0Z86c7uDBg0f9TnMJ9RgB4HSQyQOAU7Rr1y63Y8cOO8Hbtm2bGzlypPv333/d3XffnWA7BQU6wVXwoJPEqVOn2snlBx984G688cYk73/dunXWMELbqixUJaFjx451NWvWdCtXrnT58+dPsP1TTz3lUqZMaYGhHtszzzxjf1NBXWgAooAqX758rnPnzi5v3rzup59+sseiyz5wq1GjhgVcPXr0sCBLAeStt97qpk+f7ho0aGDbbdmyxeYhHjp0KLjduHHj7AT6RB0+fNgC16uuusoerwIcldjpvlRup8d/2223uTFjxtiJdbVq1ey18GWYei0UIOgE/IILLnALFixwPXv2dJs3b3bDhw+3E/TRo0e79u3b2+PWfYkCW0+Pv27duhaYKVjJmDFjko9X2VoFAXrMrVu3ttt++eWX7uuvv7ZAWK+dXl/dv8p5FbAo0P7qq69O6PX4/PPP3bRp0yy40G1feOEFCyIXL17sypYta+WmCq70Ovn3wdN1xYsXt9foeB544AF77xX46rHrfVMgo9dPr+PAgQPdhx9+6J599ln7u3rtRfu69mUFm3otypcv72bPnm0Bld6HYcOGBf+GXp8pU6a4u+66ywY5Pv3007D7u/bryy+/PNh4Ru+Z5rbq/nfv3m1B/InS+6vnoeDyzjvvtMBdgbZ888039vyaNGliAxIK7rRv6DXV5ynx+36qr9GJuOeee9ykSZPstQst7dZnSq9T3759T/i+ACCsAADgpEyYMEHD9kf9pEuXLvDKK68ctf3evXsTXD5w4ECgbNmygWuuuSbB9YULFw40b948eHnfvn2Bw4cPJ9jm119/tb/Tv3//4HWfffaZ/f1SpUoF9u/fH7x+xIgRdv2KFSvs8qFDhwJFixa1v7Nz584E93vkyJHgv2vXrh0oV66c/f3Q31evXj1w4YUXBq/r0qWL3f+iRYuC123bti2QNWtWu16P9Vj0XLXdwIEDg9fpcWXIkCGQIkWKwNSpU4PXr1q1yrbt27dv8LoBAwYEzjvvvMDq1asT3G+PHj0CqVKlCmzYsMEub9++/ajbJn4Muk243+m18j799FPbtlOnTkdt61+/YcOG2Tb6myfL70dLliwJXrd+/fpA+vTpAw0aNAhe17NnT9sH/v777wSve+rUqcM+x3D7bt26dRO859WqVbPXvF27dsHrtL8ULFgwULNmzeB1M2bMsNs/8cQTCe63UaNGdvu1a9fa5W+//da269ChQ4Lt7rrrrqPei1atWgXy5csX2LFjR4JtmzRpYvuS//xof9Jt9RyOxW/37LPPHvNzKAsXLrRtJ02adMZeI0n8HP19+s+EPte6XePGjRPcbujQofY31q1bd8znCADHQ7kmAJwilRUqM6YfZSyU1VL24u23306wXWhma+fOnZZlu/LKK92yZcuOef/K5Cgz5zNef/75Z7D8L9xtW7Zs6dKmTRu8rL/hM4KyfPly9+uvv1pmJPF8JWVRfImoMgl33HGH++effyxTqR/9bWW71JpeGRtRFkMZmCpVqgTvR1kYX+54ovSaeXpcen7K5OkxeLpOv/PPRd588017jiqZ849TP3Xq1LHXK7TM8XiU6TseZTH1OoXLsvjXz7+u77777ik1bVEWTiWanjJG9evXt4yPnpMoY6RSyNCSX2X/lFVMnEVOirJk/jFL1apVLUun671UqVJZdjL0Ndd7ruuVaQyl8k3d3neX1XaSeLvEWTndRq/rzTffbP8OfR+1v+mzcrzPyYkK/RyqTFL7tMp79Z6F+xun+hqdCH2u9TlRCbQ+Z6HZWGU9fbYaAE4VQR4AnCIFNwoo9KMTtpkzZ7rSpUtbyZnKMj2VQioY0hwrzXHzJYQ6gT0WBQkqf7vwwgst4NMcHt32+++/D3tbBQShFPz4wFI070xUWpYUlRbqRPaxxx6zvxX644Mblab6OU56bImdzHwiP68tlNY2Uzld6Am2v94/F1HAOWvWrKMep96P0Md5PJpDqb93PHr9VCKr9zApjRs3tlJXBa6aX6bSQJW6nmjAF+71vOiii6w01TcUKVmypKtcubIFBJ7+rX1MQcuJSLyv6LUVlYIe6zXXe67XQHPsQpUqVSr4e/9/BTIqHz3WvqHnpDluKoVM/D5q0OJk3scT6XjZp0+f4FxC/3nS3z+Rz9OJvkYnSsG6HpOfX/nzzz+7pUuXWiknAJwu5uQBwBmik1pl8zQHTwGImkdovpbmMGnOmeZXaS6cujZOmDAhbCORUJrzo2BLDSkGDBhgwYX+hrIh4YIGZRXCSdwU4lj8/WpenzIp4ZxoIHEiknrMJ/Jc9FivvfZaa3wRjoKjExGaMT0T2SJlEDVnTUG/glBl2dRA5uOPP07yeZ1KgKA5lGpkoqye5ow9//zzZ+V1P5n952T5/U0ZSDUjCSd0/uTp0Bw7fe70+VHG1C+UrkD8ZD5PZ+o10oCQsraqAtD7qf8rEx+awQaAU0WQBwBnkErmRA1YRKVoylap3E7BhKeTzeNROZ6CxpdffjnB9co8KAtxsnxW5YcffghmuxIrVqyY/V+BaFLbeOpiqGA2MWUkzgU9H73Ox3uciTOCp/P39D6qpPVY2TwFjOpsqp+hQ4dasK4mMgr8jvdYw72eq1evtqYgoRlPBSbdunWz5TqUDdL7pSzi2ab3/JNPPrESw9Bsnrqx+t/7/ytwUvYzNHuXeN/wnTdVinq81+Z06fOkQHLIkCHB6/bt22efp0hRcKf3UY2CNOijxjQ+Aw8Ap4NyTQA4QzTPR9kajcb78jWN+ivI8POpRF391DXzeHTbxBkCzUPzc+JO1mWXXWZzfdR1MvGJrf87uXPntm6D6uKpE8/EfMmgqHOhMkjq/Bj6+9AywrNJGQ8tF6DAKzE9Px9w+66Jp3sy37BhQ3ud/MLq4V4/BYCJqQOlKON2PHo+ofPDNm7caPP71FY/NIOkIF8dPpX90eutDpynEvifLL3n2pcTZw1VVqz9XI9J/P+fe+65BNtp3wul56TXVYMhGnw41v52usJ9ntQRN/Szea6pA6heN2VlNa/vROdUAsDxkMkDgFOkJhM+g6F5QxqJVyZGywloTTvRyLyyOToJVyt5baeGLSp51Ny6Y1FrdbXh19wkNWPQunE6offZtpOlDJPmAqrJhQIP3a/KR/Uc1PrfB0t6fFpOQGvBtWnTxv6e2twrAFF5oF+nT2WSkydPtuemk1S/hIKyOMd7bmeC2varcYVeJ61pp9I3reGm10lZGwXTCnxUQqnSOJVNqoRTWTjNSzzW3MRwlFXVfCkFLnqf9byVrVJJrn6nuZh6v1Suqfddr4Peb5Xpas6fXtPj0WNSmWzoEgoSLrBUFkhrGYrKec8F7Tt6rspM6vW99NJLbWBDgajKIH22WPuXAhg9fs130/47d+7csOs2aukPZTnV2ET7m94rBcsKdpU1DBc4nwrtJ9pfVaapv6H9Wfev9fIixa+zqMEbNYA51pIqAHAyCPIA4BSpiYOnkkw1xFAQpTXbPM3FUrmlTmR1EqxMmhbr1gny8QIhLaCtoEXBowIUZeI0z0tB5KlSAKETagUNKltTkKITc51cezoB1uLe2kZrwqkLoTJ8FSpUSPCcFSDqvjTXSc9PJ8vt2rWzxhyhHQjPFmXotK6cyiF1kqx1xxRcK5DTY/eNMuSll16yx6n109QUR01kTjbI82W2miOm91RBpv6GuisqiBHNv9R7q4W81SFSQabW8kv8eJKibTVfTNtrkXi9F3oPws1LU8Cl0j69h/q754IGChRYaz/QPqnXQ4uNa604ddgMpddAQYwGJpS51mdB+2/ixiVqUKNssAJkdaZVYKh9SXNa9Vk5UzRXVtk8PR6VaapBjoK8pOaenisK1tWcSZnp0JJuADgdKbSOwmndAwAAOOdUjqqAWsFe4nmbiB3Kgt56662WAfbLngDA6WJOHgAAMUjZMc1ZUyYIsevFF1+0kugTKecFgBNFuSYAADFk0aJFVuqreXgqoVWJJ2LP1KlT7X1UCatKSc9UF1gAEMo1AQCIIWoyo66aam6i+XqnMrcQkaegLlOmTLb0xZgxY1zq1Iy7A4iTck3Vn2sugeYU6GB3Ii3F582bZ80HNDlZ3en0BZeYOsNpIrgaIahbV2h7bwAAYpm+9zQfT81xCPBil1oiaL1BNQUiwAMQV0Geusap/bKCshPx66+/WnthtW/+9ttvrVNd69atE6yRpG5fWlhUndPUfln3r85ZamMNAAAAAPEuaso1lcl75513rMNUUh555BGrXQ9dMLVJkya2wO2sWbPssjJ3lStXDi7UqtbSates1tmn03YcAAAAAGJBTBWAa+HSOnXqJLhOWTpl9ERrHy1dutT17NkzwZo+uo1um5T9+/fbj6fAUIuvap0eJkIDAAAAiKZSb013U5wTF0Heli1bbNHUULq8e/du999//7mdO3e6w4cPh91m1apVSd7voEGDbOFZAAAAAIh2GzdudAULFoyPIO9sUeZP8/i8Xbt2uQsuuMBevCxZskT0sQEAAACAKLmlqWiZM2d2xxJTQV7evHnd1q1bE1ynywrEMmTI4FKlSmU/4bbRbZOiTp36SUz3S5AHAAAAIJocb0pZRLtrnqxq1aq5uXPnJrhuzpw5dr2kTZvWVaxYMcE2ml+ny34bAAAAAIhnEQ3y/v33X1sKQT9+iQT9e8OGDcEyymbNmgW3b9eunVu3bp3r3r27zbF74YUX3BtvvOG6du0a3EZlly+++KKbOHGi++mnn1z79u1tqYaWLVtG4BkCAAAAwLkV0XJNLeSqNe88Py+uefPmttjr5s2bgwGfFC1a1JZQUFA3YsQIm2yoRUTVYdNr3Lix2759u+vTp481ailfvrwtr5C4GQsAAAAAxKOoWScv2iY0Zs2a1RqwMCcPAAAgNqjL+sGDByP9MIBTliZNGusxcrpxSkw1XgEAAAASU85CFVx///03Lw5iXrZs2axp5Oms102QBwAAgJjmA7zcuXO7jBkzntbJMRDJwYq9e/e6bdu22eV8+fKd8n0R5AEAACCmSzR9gHf++edH+uEAp0XLwokCPe3TxyrdjJslFAAAAIBQfg6eMnhAPMj4//fl05lfSpAHAACAmEeJJuJFijNQbkyQBwAAAABxhCAPAAAAOMdq1arlunTpclr3MW/ePMv6xEJX0SJFirjhw4dH+mEkGzReAQAAQFxqPvDdc/a3Jvaqf87+FnA8ZPIAAAAAIJHTaXwSaQR5AAAAQAQcOnTIdezY0WXNmtXlzJnTPfbYY7ZWmjd58mRXqVIllzlzZlsc+6677gquoRbOn3/+6e68805XoEAB69BYrlw59/rrrx9VJtqpUyfXvXt3lyNHDrvffv36JdhG5Z/33Xefy5Mnj0ufPr0rW7as++CDD4K/nz9/vrvyyiut3X+hQoXs/vbs2RP8vR7jzTffbL8vWrSoe/XVV4/7WnzzzTfu2muvtddBr0fNmjXdsmXLTupxffXVV/b8MmbM6LJnz+7q1q3rdu7cmWS5aPny5RM8d5W+jh492t1yyy3uvPPOc08++aQt0dGqVSt7Hno+F198sRsxYsRRj3/8+PGuTJkyLl26dLa+nd5Xuffee91NN910VPCo5RFefvlld7YQ5AEAAAARMHHiRJc6dWq3ePFiCxyGDh3qXnrppQTBwIABA9x3333nZsyY4X777TfXokWLJO9v3759rmLFim7mzJnuhx9+cG3btnX33HOP3X/iv6sgZtGiRe6ZZ55x/fv3d3PmzLHfHTlyxN1www0WME2ZMsWtXLnSPfXUU8H12n755Rd3/fXXu4YNG7rvv//eTZs2zYI+H9SIHuPGjRvdZ5995t566y33wgsvHDM4lX/++cc1b97c7uvrr792F154oatXr55dfyKP69tvv3W1a9d2pUuXdgsXLrT7UaCpIO1kKOhr0KCBW7FihQVo+rsFCxZ0b775pv3NPn36uF69erk33ngjeBsFhvfff7+93rrde++950qUKGG/a926tZs1a5bbvHlzcHsFplr0vHHjxu5sSREIHS6A2b17t40g7Nq1y2XJkoVXBQAAIEopsPn1118t06LsTqzMyVPGSYHPjz/+GGyZ36NHDwsQFEyEs2TJEle5cmULfDJlymSNV66++mrLVmXLli3sbZRFKlmypBs8eHDw7yrw+fLLL4PbVKlSxV1zzTUWNH388ccWTP3000/uoosuOur+FLQosBo7dmzwOgVUyrwpm7dhwwbLdimw1GOVVatWuVKlSrlhw4adcLMZBVd6Tq+99po9h+M9LmU59bf1WMJRJk9/O/TvK5N36623BrN5eh/0ez3OY1FAu2XLFgtgRZnTli1buieeeCLs9srwKYBV9lSUKTz//PPdhAkTTnqfPtE4hUweAAAAEAGXX355gjXRqlWr5tasWRPMPi1dutSyURdccIGVbCqQEgUz4eh2yvypTFOlmAoEZ8+efdT2l1xySYLLKi/0mTZlxJS5ChdIibKKr7zyit23/1FZpIIyBSYKwpSdVEbRU5CZVBDqbd261bVp08YyeApiFMD8+++/wcd+vMflM3mnq1KlSkddN2rUKHs+uXLlsuc7bty44OPS6/bHH38c828rMPYBnZ7nRx99ZFnCs4numgAAAECUUVZMwZN+NKdNAYYCC10+cOBA2Ns8++yzVvapuWcK9FSSqcxU4u3TpEmT4LICTQVponlnx6LAS/PiNA8vMQWjq1evPoVn6yzTpTmFevyFCxe2uW0Kev1jP97jOt7vU6ZMmWC+Y1KNVfSahZo6dap76KGH3JAhQ+zxKNjW66xS1xP5u9KsWTPL0qqMdMGCBZah05zGs4kgDwAAAIgAHyh4fi6ayiFV4qigRyWUam7iyzWPRfPV6tev7+6++267rMBNQZfmqZ0oZfl+//13u124rNlll11m5aR+zlliytqpoYyykL5c8+effz7uWn567Jq7p3l4ojl9O3bsOOHHpd/PnTvXPf7442HvX0Fy6Lw4lT0q83g8elzVq1d3HTp0CF6neYmegj6Vgupvq3Q2HJVmqixU2TwFeirtPNso1wQAAAAiQJm5bt26WRCkLpgjR450nTt3DmbF0qZNa9etW7fO5uqpFPNYFCCqgYqyRSqbVMZN5YEnQyWhV111lTVW0X0pEFJ5oZqHyCOPPGL3r3lpKpFUeem7774bbLyi+XhqzKK/rSBWwZ7KFY+X8dJjVzdRPW7drmnTpgluc7zH1bNnT+vQqWDs+++/tyBZDVF8oKg5h7p/zUVUcxRlDn3TluM9LgXXKntVgKkOqPo7oTSnT5m+5557zl4PdQXV+xZKr4Ea3uj56W+fbQR5AAAAQASojO+///6zxifqzqgATx0afeZJc9/U1VGZOGX0fPOUpPTu3dsybSrpVIMVLY+gDNLJmj59umXhtByD/rYahvh5gsqYff755xbwqOSwQoUK1nEyf/78wdsrY6XLCsxuu+02e05aMuBYtJyAGsjo8asjqMpBE9/mWI9L2T01Z9GcwSpVqlhppYJPzQ/0QaAej5q43Hjjjfa6FC9e/LivhYJVPQd1wqxataplV0OzeqKgTSWyykSqyYr+hoK9UHXq1LG5j3pvQl+rs4XummHQXRMAACA2HKsTIRAtNJdRXTgVACtoPJYz0V2TOXkAAAAAcBZoXqRKRlXOqQ6jWj7hXCDIAwAAAICzNO9SGTkt/6DyW18+erYR5AEAAADAWaDOm4mXbjgXaLwCAAAAAHGEIA8AAAAA4ghBHgAAAADEEYI8AAAAAIgjBHkAAAAAEEcI8gAAAAAgjhDkAQAAAEAcYZ08AAAAxKW2bzU7Z39rXKNJLjnp16+fe/zxx4+6fs6cOa5OnTruxx9/dH369HFLly5169evd8OGDXNdunSJyGNNjgjyAAAAgGTq4MGDLk2aNKd02zJlyrhPPvkkwXU5cuSw/+/du9cVK1bM3X777a5r165n5LEihso1R40aZSvBp0+f3lWtWtUtXrz4mDth//79XfHixW37Sy+91M2aNSvBNocPH3aPPfaYK1q0qMuQIYNtO2DAgIisNA8AAACEU6tWLffAAw9Ydit79uwuT5487sUXX3R79uxxLVu2dJkzZ3YlSpRwH330UYLz3FatWgXPcy+++GI3YsSIo+57/PjxFoClS5fO5cuXz3Xs2DH4uxQpUrjRo0e7W265xZ133nnuySeftOt1nc6b06ZNa/c7efLk475xqVOndnnz5k3wo9tL5cqV3bPPPuuaNGlijwPJKMibNm2a69atm+vbt69btmyZBW1169Z127ZtC7t979693dixY93IkSPdypUrXbt27VyDBg3c8uXLg9s8/fTTtpM+//zz7qeffrLLzzzzjN0GAAAAiBYTJ050OXPmtCSHAr727dtb5qt69ep2bnzddde5e+65x7JicuTIEVewYEH35ptv2rmwyiF79erl3njjjeB96jz4/vvvd23btnUrVqxw7733ngWLiUstdQ6t3997773unXfecZ07d3YPPvig++GHH9x9991ngeZnn312zl8TnBkpAhFMcSlzpyhfAZnfcQsVKmQ7eY8ePY7aPn/+/O7RRx+1Hddr2LChjWRMmTLFLt900002EvLyyy8nuc3x7N6922XNmtXt2rXLZcmS5Qw8UwAAAJwN+/btc7/++qtlt1TpFStz8pTJU2buyy+/tMv6t84/b7vtNjdp0v/d15YtWywTt3DhQnf55ZeHvR9l6bTdW2+9ZZcLFChgAdoTTzwRdntl8pQ91Bw5r0aNGpb5GzduXPC6O+64w7KKM2fODHs/ChRVLadzbK906dJhq/JUtae/yZy809+nTzROiVgm78CBAzYRUxMzgw8mZUq7rB05nP379x/1RLVjzZ8/P3hZIx9z5851q1evtsvfffed/f6GG244a88FAAAAOFmXXHJJ8N+pUqVy559/vitXrlzwOiUuJLTKTVOdKlas6HLlyuUyZcpkgdmGDRuC2/3xxx+udu3ax/y7lSpVSnBZ1W8K9ELpsq4/FpV1fvvtt8Gf6dOnn9DzRhw3XtmxY4eNWPid19PlVatWhb2NSjmHDh3qrrrqKqsZVjD39ttv2/14ygAqwi1ZsqR9WPQ71Ro3bdo0ycei4FE/nm4PAAAAnE2JG54oyxZ6nS77ajeZOnWqe+ihh9yQIUNctWrVbN6e5r0tWrTIfh+aVTsWzcU7EzT/LnEpKKJDxBuvnAxNLL3wwgstgNNOpfS00tHKAHqqSX711Vfda6+9ZrXMqnUePHiw/T8pgwYNsrSn/1HJKAAAABBNvvrqK6ta69Chg6tQoYIFWL/88kvw9wr6VBqpRMjJKFWqlN134r+l8kvEpohl8jTJVJm2rVu3Jrhel9WZJxylpWfMmGF1qn/++afN0VPmTu1ZvYcfftiuUycfUcpba3MokGvevHnY++3Zs6c1gAnN5BHoAQAAIJoo2aH5erNnz7b5WuqA+c0339i/Q+fKqTlh7ty5bbrSP//8YwGbel4kRefPmoOnwFFTp95//32rlku8PMLJTs1Scxj/702bNllJp0pMyf7FcSZPmTjVE4eONCgVrctKPx+L5uVpUumhQ4es9rd+/frB36n7UGhmTxRM+jR3OGrrqomLoT8AAABANFHXSzVmady4sTUwVNJDWb1QSmoMHz7cvfDCC9ZMRU0J16xZc8z7vfXWW61iTtVvuo262U+YMMGaw5wqzQ1U0KifzZs3233r361btz7l+0SMdNfUEgraEbUjValSxXZIlVtqTp7m5jVr1syCOWXhRPXGGgUoX768/V8jFeo8o7LMbNmy2TYtWrSwUQfdp3ZSLa+gFrJqD6vlFE4E3TUBAABivxMhkFy7a0asXFM0CrF9+3Zb40OtXxW8aXFz34xFnYJCs3J6wlorb926dZbqrVevnqWpfYAnWg9Pi6FrVEMdhlTSqVEP/Q0AAAAAiHcRzeRFKzJ5AAAAsYFMHuLNvlheJw8AAAAAcOYR5AEAAABAHCHIAwAAAIA4QpAHAACAmHes5bKA5LYvR7S7JgAAAHC6ay+rG7vWZcuVK5ddTpEiBS8qYo76YWrheK0+oH1a+/KpIsgDAABAzNLJsLoQasFtBXpArMuYMaO74IILEiwld7II8gAAABDTlPHQSfGhQ4fc4cOHI/1wgFOWKlUqlzp16tPORhPkAQAAIObppDhNmjT2AyR3NF4BAAAAgDhCkAcAAAAAcYQgDwAAAADiCEEeAAAAAMQRgjwAAAAAiCMEeQAAAAAQRwjyAAAAACCOEOQBAAAAQBwhyAMAAACAOEKQBwAAAABxhCAPAAAAAOIIQR4AAAAAxBGCPAAAAACIIwR5AAAAABBHCPIAAAAAII4Q5AEAAABAHCHIAwAAAIA4QpAHAAAAAHGEIA8AAAAA4ghBHgAAAADEEYI8AAAAAIgjBHkAAAAAEEcI8gAAAAAgjkQ8yBs1apQrUqSIS58+vatatapbvHhxktsePHjQ9e/f3xUvXty2v/TSS92sWbOO2m7Tpk3u7rvvdueff77LkCGDK1eunFuyZMlZfiYAAAAAkMyDvGnTprlu3bq5vn37umXLllnQVrduXbdt27aw2/fu3duNHTvWjRw50q1cudK1a9fONWjQwC1fvjy4zc6dO12NGjVcmjRp3EcffWTbDRkyxGXPnv0cPjMAAAAAiIwUgUAgEKG/bZm7ypUru+eff94uHzlyxBUqVMg98MADrkePHkdtnz9/fvfoo4+6+++/P3hdw4YNLVs3ZcoUu6zbffXVV+7LL7885ce1e/dulzVrVrdr1y6XJUuWU74fAAAAADhTTjROiVgm78CBA27p0qWuTp06/3swKVPa5YULF4a9zf79+61MM5QCvPnz5wcvv/fee65SpUru9ttvd7lz53YVKlRwL7744jEfi+5XL1joDwAAAADEoogFeTt27HCHDx92efLkSXC9Lm/ZsiXsbVTKOXToULdmzRrL+s2ZM8e9/fbbbvPmzcFt1q1b50aPHu0uvPBCN3v2bNe+fXvXqVMnN3HixCQfy6BBgywi9j/KJgIAAABALIp445WTMWLECAveSpYs6dKmTes6duzoWrZsaRlAT8HfZZdd5gYOHGhZvLZt27o2bdq4MWPGJHm/PXv2tJSn/9m4ceM5ekYAAAAAECdBXs6cOV2qVKnc1q1bE1yvy3nz5g17m1y5crkZM2a4PXv2uPXr17tVq1a5TJkyuWLFigW3yZcvnytdunSC25UqVcpt2LAhyceSLl06q2kN/QEAAACAWBSxIE+ZuIoVK7q5c+cmyMLpcrVq1Y55W83LK1CggDt06JCbPn26q1+/fvB36qz5888/J9h+9erVrnDhwmfhWQAAAABAdEkdyT+u5ROaN29ujVKqVKnihg8fblk6lWBKs2bNLJjTnDlZtGiRrYFXvnx5+3+/fv0sMOzevXvwPrt27eqqV69u5Zp33HGHrbs3btw4+wEAAACAeBfRIK9x48Zu+/btrk+fPtZsRcGbFjf3zVhUYhk6327fvn22Vp6aq6hMs169em7y5MkuW7ZswW20JMM777xj8+y0cHrRokUteGzatGlEniMAAAAAJJt18qIV6+QBAAAAiDZRv04eAAAAAODMI8gDAAAAgDhCkAcAAAAAcYQgDwAAAADiCEEeAAAAAMQRgjwAAAAAiCMEeQAAAAAQRwjyAAAAACCOEOQBAAAAQBwhyAMAAACAOEKQBwAAAABxhCAPAAAAAOIIQR4AAAAAxBGCPAAAAACIIwR5AAAAABBHCPIAAAAAII4Q5AEAAABAHCHIAwAAAIA4QpAHAAAAAHGEIA8AAAAA4ghBHgAAAADEEYI8AAAAAIgjBHkAAAAAEEcI8gAAAAAgjhDkAQAAAEAcIcgDAAAAgDhCkAcAAAAAcYQgDwAAAADiCEEeAAAAAMQRgjwAAAAAiCNREeSNGjXKFSlSxKVPn95VrVrVLV68OMltDx486Pr37++KFy9u21966aVu1qxZSW7/1FNPuRQpUrguXbqcpUcPAAAAANEj4kHetGnTXLdu3Vzfvn3dsmXLLGirW7eu27ZtW9jte/fu7caOHetGjhzpVq5c6dq1a+caNGjgli9fftS233zzjW17ySWXnINnAgAAAACRF/Egb+jQoa5NmzauZcuWrnTp0m7MmDEuY8aMbvz48WG3nzx5suvVq5erV6+eK1asmGvfvr39e8iQIQm2+/fff13Tpk3diy++6LJnz36Ong0AAAAAJOMg78CBA27p0qWuTp06/3tAKVPa5YULF4a9zf79+61MM1SGDBnc/PnzE1x3//33uxtvvDHBfQMAAABAvEsdyT++Y8cOd/jwYZcnT54E1+vyqlWrwt5GpZzK/l111VU2L2/u3Lnu7bfftvvxpk6daqWfKtc8EQoc9ePt3r37lJ8TAAAAACTbIO9UjBgxwso7S5YsaQ1VFOip1NOXd27cuNF17tzZzZkz56iMX1IGDRrkHn/88bP8yAHg9DUf+G7MvYwTe9WP9EMAACBZiWi5Zs6cOV2qVKnc1q1bE1yvy3nz5g17m1y5crkZM2a4PXv2uPXr11vGL1OmTDY/T1T+qaYtl112mUudOrX9fP755+65556zf4dm/LyePXu6Xbt2BX8UKAIAAABALIpokJc2bVpXsWJFK7n0jhw5YperVat2zNsqS1egQAF36NAhN336dFe//v+NFNeuXdutWLHCffvtt8GfSpUqWRMW/VtBZWLp0qVzWbJkSfADAAAAALEo4uWaWj6hefPmFohVqVLFDR8+3LJ0KsGUZs2aWTCnkkpZtGiR27Rpkytfvrz9v1+/fhYYdu/e3X6fOXNmV7Zs2QR/47zzznPnn3/+UdcDsYZSPQAAAJy1IO/vv/92b731lvvll1/cww8/7HLkyGHNTtQ0RUHZiWrcuLHbvn2769Onj9uyZYsFb1rc3Ddj2bBhg3Xc9Pbt22dr5a1bt87KNLV8gpZVyJYt26k+FQBnUdu3msXk6zuu0aRIPwQAAIBzF+R9//33tjRB1qxZ3W+//WaNUBTkqculgrJJk07u5Khjx472E868efMSXK5Zs6Ytgn4yEt8HAAAAAMSrlKdaYtmiRQu3Zs2aBB0slVX74osvzuTjAwAAAACc7SBP68/dd999R12vMk2VXAIAAAAAYijIUzfKcAuGr1692pY4AAAAAADEUJB3yy23uP79+7uDBw/aZS1Krrl4jzzyiGvYsOGZfowAAAAAgLMZ5A0ZMsT9+++/Lnfu3O6///6zZiglSpSw5QuefPLJU7lLAAAAAECkumuqq+acOXPcV1995b777jsL+C677DLruAkAAAAAiKEgTyWaGTJkcN9++62rUaOG/QAAAAAAYrRcM02aNO6CCy5whw8fPjuPCAAAAABwbufkPfroo65Xr17ur7/+OvW/DAAAAACIjjl5zz//vFu7dq3Lnz+/K1y4sDvvvPMS/H7ZsmVn6vEBAAAAAM52kHfrrbeeys0AAAAAANEY5PXt2/fMPxIAAAAAQGSCPG/p0qXup59+sn+XKVPGVahQgbcEAAAAAGItyNu2bZtr0qSJmzdvnsuWLZtd9/fff7urr77aTZ061eXKletMP04AAAAAwNnqrvnAAw+4f/75x/3444/WYVM/P/zwg9u9e7fr1KnTqdwlAAAAACBSmbxZs2a5Tz75xJUqVSp4XenSpd2oUaPcdddddyYeFwAAAADgXAV5R44csUXRE9N1+h3gtX2rWcy9GOMaTYr0QwAAAADObZB3zTXXuM6dO7vXX3/d1sqTTZs2ua5du7ratWuf+qMBAMSdWBzsEQZ8AADJak6eFkPX/LsiRYq44sWL20/RokXtupEjR575RwkAAAAAOHuZvEKFCrlly5bZvLxVq1bZdZqfV6dOnVO5OwAAAABApNfJS5Eihbv22mvtBwAAAAAQw+WaWibhueeeC1vG2aVLlzPxuAAAAAAA5yqTN336dPfee+8ddX316tXdU0895YYPH34qd4vjaD7w3Zh7jdJdFOlHAAAAACQvp5TJ+/PPP13WrFmPuj5Llixux44dZ+JxAQAAAADOVZBXokQJWxA9sY8++sgVK1bsVO4SAAAAABCpcs1u3bq5jh07uu3bt9uaeTJ37lw3ePBgN2LEiDPxuAAAAAAA5yrIu/fee93+/fvdk08+6QYMGGDXaZ28MWPGuGbNYnPRWwAAAABItuWa//33n2vevLn7/fff3datW933339vmb08efKc+UcIAAAAADi7QV79+vXdpEmT7N9p0qSxRdCHDh3qbr31Vjd69OhTuUsAAAAAQKSCvGXLlrkrr7zS/v3WW29ZBm/9+vUW+IVbPw8AAAAAEMVz8vbu3esyZ85s//7444/dbbfd5lKmTOkuv/xyC/ZO1qhRo9yzzz7rtmzZ4i699FI3cuRIV6VKlbDbHjx40A0aNMhNnDjRbdq0yV188cXu6aefdtdff31wG/3+7bffdqtWrXIZMmSw9fu0jbYFACCWxeKaqTKxV/1IPwQASDZOeQmFGTNmuI0bN7rZs2e76667zq7ftm2brZV3MqZNm2bdOvv27WsZQgV5devWtfsKp3fv3m7s2LEWCK5cudK1a9fONWjQwC1fvjy4zeeff+7uv/9+9/XXX7s5c+ZYYKjHuGfPnlN5ugAAAAAQ30Fenz593EMPPeSKFCniqlat6qpVqxbM6lWoUOGk7ktz+dq0aeNatmzpSpcubR06M2bM6MaPHx92+8mTJ7tevXq5evXq2Zp87du3t38PGTIkuI3W8GvRooUrU6aMBY2vvPKK27Bhg1u6dOmpPF0AAAAAiO9yzUaNGrkrrrjCbd682YIor3bt2pZVO1EHDhywwKtnz57B61T2qUYuCxcuDHsbLd2QPn36BNepJHP+/PlJ/p1du3bZ/3PkyHHCjw0AAAAAkk2QJ3nz5rWfUEnNo0vKjh073OHDh49aekGXNZ8uHJVyKvt31VVXueLFi9si7Jp/p/sJ58iRI65Lly6uRo0armzZskkGjvrxdu/efVLPAwAAAABiulwzkkaMGOEuvPBCV7JkSZc2bVpbn0+lnsoAhqO5eT/88IObOnVqkvepRi1Zs2YN/hQqVOgsPgMAAAAAiNMgL2fOnC5VqlS2oHooXU6cJfRy5cplTV/UREWdPJXxy5Qpk83PS0wB4AcffOA+++wzV7BgwSQfh8pFVdLpf9RQBgAAAABiUUSDPGXiKlasaCWXoeWVuuybuSRF8/IKFCjgDh065KZPn24LtHuBQMACvHfeecd9+umnrmjRose8r3Tp0llX0NAfAAAAAEhWc/LOFC2f0Lx5c1epUiWb0zd8+HDL0qkEU5o1a2bBnEoqZdGiRbY+Xvny5e3//fr1s8Cwe/fuCUo0X3vtNffuu+/aen5af09UiqkmLQAAAAAQryIe5DVu3Nht377dlmVQMKbgTUsg+GYsWvogdL7dvn37bK28devWWZmmlk/QsgrZsmULbjN69Gj7f61atRL8rQkTJtjSCgAAAAAQryIe5IlKK/UTzrx58xJcrlmzpi2Cfiwq1wQAAACA5CjmumsCAAAAAJJGkAcAAAAAcYQgDwAAAADiSFTMyQMAAMCJaz7w3Zh7uSb2+t9yVwDOLjJ5AAAAABBHCPIAAAAAII4Q5AEAAABAHCHIAwAAAIA4QpAHAAAAAHGEIA8AAAAA4ghBHgAAAADEEYI8AAAAAIgjBHkAAAAAEEcI8gAAAAAgjhDkAQAAAEAcSR3pBwAAAOJf27eauVgzrtGkSD8EADglZPIAAAAAII4Q5AEAAABAHCHIAwAAAIA4QpAHAAAAAHGEIA8AAAAA4ghBHgAAAADEEYI8AAAAAIgjBHkAAAAAEEcI8gAAAAAgjhDkAQAAAEAcIcgDAAAAgDhCkAcAAAAAcYQgDwAAAADiCEEeAAAAAMQRgjwAAAAAiCNREeSNGjXKFSlSxKVPn95VrVrVLV68OMltDx486Pr37++KFy9u21966aVu1qxZp3WfAAAAABAvUkf6AUybNs1169bNjRkzxoKx4cOHu7p167qff/7Z5c6d+6jte/fu7aZMmeJefPFFV7JkSTd79mzXoEEDt2DBAlehQoVTuk8AAAAgljQf+K6LNRN71Y/0Q0g2Ip7JGzp0qGvTpo1r2bKlK126tAVmGTNmdOPHjw+7/eTJk12vXr1cvXr1XLFixVz79u3t30OGDDnl+wQAAACAeBHRIO/AgQNu6dKlrk6dOv97QClT2uWFCxeGvc3+/futBDNUhgwZ3Pz580/rPnfv3p3gBwAAAABiUUTLNXfs2OEOHz7s8uTJk+B6XV61alXY26jsUpm6q666yublzZ0717399tt2P6d6n4MGDXKPP/74GXteAAAASKjtW81i7iUZ12hSpB8CEJvlmidrxIgR7sILL7T5eGnTpnUdO3a0skxl605Vz5493a5du4I/GzduPKOPGQAAAACSRZCXM2dOlypVKrd169YE1+ty3rx5w94mV65cbsaMGW7Pnj1u/fr1lp3LlCmTzc871ftMly6dy5IlS4IfAAAAAIhFEQ3ylImrWLGilVx6R44cscvVqlU75m01L69AgQLu0KFDbvr06a5+/fqnfZ8AAAAAEOsivoSCljpo3ry5q1SpkqtSpYotd6AsnUowpVmzZhbMad6cLFq0yG3atMmVL1/e/t+vXz8L4rp3737C9wkAAAAA8SriQV7jxo3d9u3bXZ8+fdyWLVsseNPi5r5xyoYNGxLMt9u3b5+tlbdu3Tor09TyCVpWIVu2bCd8nwAAAAAQryIe5Imap+gnnHnz5iW4XLNmTbdy5crTuk8AAAAAiFcx110TAAAAAJA0gjwAAAAAiCMEeQAAAAAQRwjyAAAAACCOEOQBAAAAQBwhyAMAAACAOEKQBwAAAABxhCAPAAAAAOIIQR4AAAAAxJHUkX4AAAAAAOJf27eauVgzrtEkF4vI5AEAAABAHCHIAwAAAIA4QpAHAAAAAHGEIA8AAAAA4ghBHgAAAADEEYI8AAAAAIgjBHkAAAAAEEcI8gAAAAAgjhDkAQAAAEAcIcgDAAAAgDhCkAcAAAAAcYQgDwAAAADiCEEeAAAAAMQRgjwAAAAAiCMEeQAAAAAQRwjyAAAAACCOEOQBAAAAQBwhyAMAAACAOEKQBwAAAABxhCAPAAAAAOJIxIO8UaNGuSJFirj06dO7qlWrusWLFx9z++HDh7uLL77YZciQwRUqVMh17drV7du3L/j7w4cPu8cee8wVLVrUtilevLgbMGCACwQC5+DZAAAAAEBkpY7kH582bZrr1q2bGzNmjAV4CuDq1q3rfv75Z5c7d+6jtn/ttddcjx493Pjx41316tXd6tWrXYsWLVyKFCnc0KFDbZunn37ajR492k2cONGVKVPGLVmyxLVs2dJlzZrVderUKQLPEgAAAACSSSZPgVmbNm0sCCtdurQFexkzZrQgLpwFCxa4GjVquLvuusuyf9ddd5278847E2T/tE39+vXdjTfeaNs0atTItjtehhAAAAAA4kHEgrwDBw64pUuXujp16vzvwaRMaZcXLlwY9jbK3uk2PmBbt26d+/DDD129evUSbDN37lzL8sl3333n5s+f72644Yaz/pwAAAAAINmWa+7YscPmz+XJkyfB9bq8atWqsLdRBk+3u+KKK2yO3aFDh1y7du1cr169gtuonHP37t2uZMmSLlWqVPY3nnzySde0adMkH8v+/fvtx9PtAQAAACAWRbzxysmYN2+eGzhwoHvhhRfcsmXL3Ntvv+1mzpxpjVW8N954w7366qs2f0/baG7e4MGD7f9JGTRokM3Z8z9q6AIAAAAAsShimbycOXNapm3r1q0JrtflvHnzhr2Numbec889rnXr1na5XLlybs+ePa5t27bu0UcftXLPhx9+2LJ5TZo0CW6zfv16C+SaN28e9n579uxpDWBCM3kEegAAAABiUcQyeWnTpnUVK1a0+XPekSNH7HK1atXC3mbv3r0WyIVSoCh+iYSkttF9JyVdunQuS5YsCX4AAAAAIBZFdAkFZc+UXatUqZKrUqWKLaGgzJy6bUqzZs1cgQIFLAsnN998s3XkrFChgi25sHbtWsvu6Xof7OnfmoN3wQUX2BIKy5cvt9vce++9kXyqAAAAABD/QV7jxo3d9u3bXZ8+fdyWLVtc+fLl3axZs4LNWDZs2JAgK9e7d29bE0//37Rpk8uVK1cwqPNGjhxpgV+HDh3ctm3bXP78+d19991nfwMAAAAA4l1Egzzp2LGj/STVaCVU6tSpXd++fe0nKZkzZ7aMoH4AAAAAILmJqe6aAAAAAIBjI8gDAAAAgDhCkAcAAAAAcYQgDwAAAADiCEEeAAAAAMQRgjwAAAAAiCMEeQAAAAAQRwjyAAAAACCOEOQBAAAAQBwhyAMAAACAOEKQBwAAAABxhCAPAAAAAOIIQR4AAAAAxBGCPAAAAACIIwR5AAAAABBHCPIAAAAAII4Q5AEAAABAHCHIAwAAAIA4QpAHAAAAAHGEIA8AAAAA4ghBHgAAAADEEYI8AAAAAIgjBHkAAAAAEEcI8gAAAAAgjhDkAQAAAEAcIcgDAAAAgDhCkAcAAAAAcYQgDwAAAADiCEEeAAAAAMQRgjwAAAAAiCMEeQAAAAAQRyIe5I0aNcoVKVLEpU+f3lWtWtUtXrz4mNsPHz7cXXzxxS5DhgyuUKFCrmvXrm7fvn0Jttm0aZO7++673fnnn2/blStXzi1ZsuQsPxMAAAAAiLzUkfzj06ZNc926dXNjxoyxAE8BXN26dd3PP//scufOfdT2r732muvRo4cbP368q169ulu9erVr0aKFS5EihRs6dKhts3PnTlejRg139dVXu48++sjlypXLrVmzxmXPnj0CzxAAAAAAklGQp8CsTZs2rmXLlnZZwd7MmTMtiFMwl9iCBQssgLvrrrvssjKAd955p1u0aFFwm6efftoyfBMmTAheV7Ro0XPyfAAAAAAg2QZ5Bw4ccEuXLnU9e/YMXpcyZUpXp04dt3DhwrC3UfZuypQpVtJZpUoVt27dOvfhhx+6e+65J7jNe++9Z9nA22+/3X3++eeuQIECrkOHDhZMJmX//v324+3atcv+v3v3bhdNDuzb62JNir0HXKyJtvc9FPvAuROt+wH7wLnDPnBm8X1wZnEsSN7HAWEfSJ77wO7//3gCgcCxNwxEyKZNm/TIAgsWLEhw/cMPPxyoUqVKkrcbMWJEIE2aNIHUqVPb7du1a5fg9+nSpbOfnj17BpYtWxYYO3ZsIH369IFXXnklyfvs27ev3Rc/vAbsA+wD7APsA+wD7APsA+wD7APsAy7KX4ONGzceM9aKaLnmyZo3b54bOHCge+GFF2wO39q1a13nzp3dgAED3GOPPWbbHDlyxFWqVMm2kwoVKrgffvjBSkGbN28e9n6VTdTcQE/38ddff1njFs33w6mPNKh0duPGjS5Lliy8jMkQ+wDYB8CxAOwDYB84c5TB++eff1z+/PmPuV3EgrycOXO6VKlSua1btya4Xpfz5s0b9jYK5FSa2bp1a7usrpl79uxxbdu2dY8++qiVe+bLl8+VLl06we1KlSrlpk+fnuRjSZcunf2EypYt22k8O4RSgEeQl7yxD4B9ABwLwD4A9oEzI2vWrNG7hELatGldxYoV3dy5cxNk0HS5WrVqYW+zd+9eC+RCKVAMrUtVYxZ15wylLpyFCxc+C88CAAAAAKJLRMs1VSKpEkqVV6qRipZQUGbOd9ts1qyZNU4ZNGiQXb755putI6dKMH25prJ7ut4He1o3Tw1aVK55xx13WJOWcePG2Q8AAAAAxLuIBnmNGzd227dvd3369HFbtmxx5cuXd7NmzXJ58uSx32/YsCFB5q537942R07/14LnWgNPAd6TTz4Z3KZy5crunXfesXl2/fv3t+UTFDw2bdo0Is8xOVMJbN++fY8qhUXywT4A9gFwLAD7ANgHzr0U6r4Sgb8LAAAAADgLIjYnDwAAAABw5hHkAQAAAEAcIcgDAAAAgDhCkAcAAAAAcYQgDwBw0tSzi75dAIBQf//9Ny9IlCDIw2nTIvZIng4dOhT8Nyf8yeszr+Vs9ANs3brV1rhF8j4m+HMBvguSLy1X9tBDD7m//vor0g8FBHk4E7SWISM3ycvnn3/ubr31VtemTRs3ceJEu44T/uT1md+8ebPr3r27GzRokPvyyy85wUtmdCJ/4MAB17p1a1esWDE3f/78SD8kRHBf0DHBnwvs27cvwe8Q/+cDv/76q/37wgsvdF9//bVbsWJFpB8WCPJwJrJ2n3zyiStYsKDbv38/L2gyMGrUKNe4cWNXqFAhd/DgQde5c2fXrl07t3v3bvs9X+rx780333QlS5Z0S5cudZ999pmrX7++69+/v/2OYD950PucKlUqO7n777//3Hvvvef+/fffSD8sRGhfWLNmjWvQoIGrVKmSu+222yyb43+H+PXhhx+6u+++27377rt2WQN/OhecPXs2x4MoQLkmTm6HSZkybOCnkdwNGzbwasaxw4cP28j9W2+9ZSUZI0eOdFOmTLHLOulXRk8BHl/q8V2GrbK8l19+2bVv397NnTvXzZo1y/Xq1cu99NJLtj8c67aILwsXLnSFCxd2r776qr3/P/30U6QfEs7B90Biv/32m2vSpInLmDGjmzBhgmvbtq2bNGmS69u3r9u1axfvSRy74YYb3GWXXeaWLFni1q9fb/tAo0aN7Hvhhx9+iPTDS/YI8pAknbAnPllT1m7EiBFuy5YtwetUmrF3716XPn16Xs0488QTT7h33nnH/q1Rex3EV61a5erWrRvcpk6dOu7mm2+2kTwd6IVsXuzyn3k/oPPnn38e9ftvvvnGRuz9dq1atbJsnoK90NsiPvzyyy9hS/B0wv/777+7O++80+XMmdNNnjzZsvuh2yC+jgv6HvDzMD0Fdvnz57dg/8orr3SXXnqp7RsaBNq2bVvEHjPOLr3HGtRVBlfnBarqEFX36Hvj008/pcIrwvgmxjEbKyQ+WdPIzLhx41zHjh2D11WpUsW+6P3Ee77cY5vKLjt16uQyZMhgX9o5cuQI/k719srmaeRW/ImfyjU1R2vBggV2mWxe7PKfeWVmLr/8chuVffzxx60cSxToaz/YuHFj8DbZs2e3QG/Hjh2W1RWyebFNpZcPP/ywlWUrc3/NNdcE59n4z/cbb7zhLrnkEvt3z5493dSpU20/mTlzJseAOD0uKIuvAR6932vXrrXrdCy4+OKL3eLFi21/qFq1qpXwKejXsQLxE9SFnt/5gP+OO+5wWbNmtbl5Cv7z5s3r6tWrZ8cBn93nvDAyCPKSOZ3Qa5Js4lIMHdDVOVHzbDSxXqUXfoTmtddes4xejx497OCu7cqWLWsjOcIJfmzSQVgnadmyZbOM3Pvvv28H6Jo1awZ/r/f6lltucaNHj7br0qVLZ/9XMKCR3GXLliUY8UfseeaZZ1zu3LndgAED7Mv7iiuusJP54cOH2+/LlCljAwCrV69O0EGtRIkStm+MHTvWLpPNi+15NjpR13FA7+ezzz5r1RqPPfaY++eff4InbJqLV7FixeBAj35Xvnx523f++OOPCD8LnEn6rKsk88knn3R33XWX/aROndoyt5qT//zzz1tFh7b5/vvvreKnaNGiNmcztAszYpeCOp3f6bxAAz76/ItKNPXer1y50gI96dKliw0IfvHFF8GMH849grxkRl/OoTX1aqChCfP6sOp3+hAr8NOBWSn4Dz74wEZ0W7RoYV30dAKvUgyN5qnpgjI+OtFXJk/lOojdfUIHYb3Xml+pL3KVYeogra55OnnT7/Wlrv1CB/l58+bZdb7hzo033mjbUrYbu5S90+DNI488Yu99t27d7IRdJ/ybNm2y+TV6zzVKq46avjxXsmTJ4sqVK2cndLTPju3vBh3nu3bt6mbMmGHvtUrwOnToYMd8fb61D+h9ViZHJ3ga/MuUKZNl/TNnzmyBoQZ9EFuUfU88585fVpWG5mBOnz7djgv6fihSpIhLkyaNNWG66KKLLLunkm1lcuTbb7+163TsQOzTd76+C6666ior0dagns/uK9ufNm1a20bfE8rsajudX2pAUMjmnXsEeclsfp3viKaTec210mjbwIEDbWTej7TowK0SLd8KV9kdbaeRfH9S17BhQ/fcc8/Z7xX8+cBQKNOKDZpXpVE3v0/ITTfdZKU46pClkdpatWrZMgn6AldXTZ3Y6brrrrsu2D3NZ/M0b0df7n50D7HDf/nee++9dpKuYD70C1llmCrLVkmOz9ro5E5Nd0IbK+ikXyP7ygYjdo8DCuw0AOjfb31fqJGCPvM7d+606zQvW//WoI8y+DqZ+/nnn62ce86cOWRvYnQervYBzaNbtGiRXef3Cb2/OnFXJt/zGbqrr77aqjl0LqHpHCrb1DmDSjYVJPr7QOwlAjzNt9PAj4L75cuX2xx8fU/oPddUjVy5ctnvFNgreyfaXt8J/jLZvAgIIK4dOXIkcPjw4eDl7du3B1q1ahVIkSJFoEqVKoHNmzcH1qxZExg5cmTg119/tW2GDBlivx88eHCC+ypSpEjgwQcfDOzatSt43Zw5cwI33nijbT9ixIhz+MxwOvbt2xdo3LhxoFixYnZ5wIABgQceeMD+PW7cuMDFF18caNasWWDZsmWB+fPnBx577LFA3rx5Ay+99JJto30mU6ZMgYYNGwZmzJgR+PDDDwMXXnhh4MUXX+SNiVEHDx60/3ft2jVQpkyZwG+//Wafab3v2bNnD5QqVSpQsWLFwOzZs2077QuXXnpp4M477wysXbs2sHz58kCNGjUCw4YNi/AzwekcBzp27Bg4cOBA8PtD3wM6vmubqlWrBkqUKBGYNm2a/f71118PLFy4MMF3TNOmTQPlypUL/PXXX7wRMWbPnj123NexvWDBgoEWLVoEvvvuO/tdhw4d7PMeeqxIfNvWrVsHKlWqFChZsmTgoosuCowfP/6cPwecOn3ek6Lzw9Dvd33v58mTJ1CoUKHA888/b9f98ssvdox46KGHAv/9959dd/nllwfq168f2LFjB29NBBDkJROrVq2yg7e+rKtVqxb49ttvg7975ZVXAuedd17g7bffDn7QM2bMGOjTp09g//79wQ/+M888Y4He4sWLE9z31q1bA/nz5w8GhaFf+IjeA/gnn3wSSJ06tb3XOoF/7733ggfqt956K7Bly5YE2990002B2267LfDHH3/Y5ZkzZwYaNWpkt82RI0egd+/ewZNDRCcN8ugzHc6hQ4fs/3///XcgVapUgQwZMgQuueQSC+Z0u48//jhwww03BIoXL27Bge5H111wwQW2Xfr06QN33HEHX+YxfhzQ5zrU559/boN5//zzj723bdq0sQHCH374IcF2/ri/d+9e+0H0Svwdre/wzp07Bx5//HEb7NOgzcsvvxyoXr164Pbbb7dt5s6da/vJV199leC2P/30k+1D/n51XND5xrH+HqKH3pvQY4LeP32X16xZ0wb8lQgQHfNF77+CfQXx/fv3D1x//fU20L9z5077fffu3QOVK1cOHkd0PvHvv/9G5LmBIC/uLVmyxL6Q06ZNayNsGmH1oy6ho3H6ULZr1y54Aq8DvkZs161bF9xG22t0T8GfRu1CTww1mq9RP0QfvUfhvmRbtmwZSJkyZXB0NlToQd/vJ0899ZSd4Ce+L33JI/rpi7ts2bLByz4bo/cz8ch8p06dbJTWZ/e9Dz74wAaKlOH1lLFZsGBB8NiB+DkOhPL7iAYD9T2QOMg7kWwAIst/XyemgVsFdFmyZAm8+eabweuViStQoIB9vqVevXo2oKN9QINBGiy+66677JwgXHYv3HWIDvqchn5W/bFB3xO1atWyYL9w4cL2nofuP7qs7wc/CKx/586dOzBlypTgOacGA5XhR+QxJy+OaW6UuuJpUrQao2juReXKla1rouZTqJ5a8ydEjVW0psl3331nl/v162dz7D7++ONgvb62VwclTbz++++/7Tpfa68mDZqzg+ij90hzLdT5SusZaT8QzbHTJGl1Qku8vl1o7bzed83JUVOV0qVLH1Wvr/0L0cu/p2qW8OOPP9p6dmqWobm0WstI+4beY33ONc9WxwE1X9G8HDVaCO2Mp/kWBQoUsAn2ocsnVKtWzeXLly8izw9n7zgQys/TVKMFNdnyzTUSY95N9PLf1+qQrTlz+nyLOqRqbqXOGdQp26tRo4ZdHjJkiF0eP368zdXXYueaf6UmHDo+6LL2j8TCXYfIrnXsr9fnVD+aS6/59+3bt3dDhw61/grTpk2z74LXX3/d5uKq264//mvebYUKFVyePHms4+6GDRtsv9IxRU3YtC9pe83RRBSIdJSJ05N4ZDbxKGri30+ePNlqpkePHp1gpE3/1zws1VL72mmVdyrDt3HjxuDtVWcdOidP/27fvr2N7icu48C59+effwZ+//33BO+t3j+VWqokV6UV+v8jjzwSWL9+vf1e86hUM594f9FIrbIzKrfQaJ3maWl+HmIza6OyS31OVVobOlqvbfX+6nplc7788ku7XvuMSnb83IoXXnjB5tpo30HyOQ7oGK+SLR0HVOGhMi3NxUN007mAz9z5/3/00UeBokWLWpXOLbfcYvvAk08+aeW1P/74o33fa06mp/1g1KhRVpL9/fff23UqvVMGb+rUqcFSPkR3LwYdDzRnOrSMWp9nldlec801Vm6pqRj6fmjSpEmCDOytt95q3wOiMl4dJ3Qc0fmeejSoEkClvX4aD5n86EKQFyf0RZxU3bsO8P4grw+7JtrrAL979267zs+jeuKJJ6ycc968eXZZpZr60L/zzjvB+/IfYH9/ug+VcPn7QuToS1qNMVQTH0p18yqf2LBhg13+9NNP7X3WJHnRCb/KtRKX5ul6lfqq6cbVV19tZRiI3nKbcDTXZtu2bfbv1atXBwM9fcZDP8v6stZE+tBjiBouaFsdL3LmzGnzbmmkkPyOAzr+6yQvV65cVsa1aNGic/hscLrHBQ3a+u94Ncrq1q1b8HcqsVOJpppt6TZqsFW6dOngPiIrVqyw/UlBwMmUASPyVq5cafOk1TSpefPmwbmT+k7Q+67SWwX5ovdfAZ4G+vwAkahUV98DmpvrEwU6H1CJphIGof0dEH0I8mKcJslqsrxq4n3Xw/fffz/4u9CDr//3mDFj7MOpD6v4LwAFippz9fDDDwcn0WqyNaJb6HusA7lG2X7++efgSX6dOnWCk6B1IlehQoVAvnz5AsOHD7cDu04CNIKrEwB9YevkXid2+rcyvgoOEL38CV3ofqATcZ2Q64RNI6/PPvusNc4QfYmrYY4ytUnx93XFFVdY19Tp06ef9eeB6DoOKHOjuVe63wkTJth3C2KHvv+VhdMJ+htvvGHzaHWu4E/y1UX3/PPPD1x55ZXBOXcK/HVu0Ldv3wT3M3bs2MCrr7561N8gaxOd9Pl9+umnLRC7++67rUGWKjQ0yO/fMw3saVDHz6UTHS+0v6jxWuh7W7t2bcv2efruCB0IQPQiyIsx+sJVyn3gwIHBAGzo0KH2Za2J0+qKp4Av9AtfI7sarfENMlS2o7IcjfCElmuKSi+VfvdBHqKTTtbUJVVZ1NASjHfffde+pH2nUx2MNWKnUX2VWalhgjK2PrvjqS1+1qxZrTGHDvLahi/w6KZuZ+qGpxK60Oy6Rt7Lly9vXRD1bzVa0r6iz7YoY5cmTZpgWWY4/nhAV7ToxnEAibNo+j4YNGiQZeC17IGyN9pGjXJUnqnrdaxXmZ6WvgmlhmqasqHyTAUEiE0KwDSIM2nSpCT3FwV0Wh5HA4Ch5ZnXXnutDQipo7Kn74x06dIlyPAhNhDkRTl/oh1agvHFF1/YB86fqKtrpk7MVVKzadOm4G21zpVO+LXmjQK30HWLtJaVRvh9KaZvq04L/NigL2q955pboZE6P29KB2v9TkG8L7tq0KCBbatSu9CyWpXjal6Fv04n/Zpr4zunIjokFWz7EzKtUxT62VZXtJtvvjl4WdkYzbfTvuIHBJSda9u2bTC757+8KbuKLRwHkqfQ+XaJ6Xiu4C5z5szBUlxRJlZldloGyWd4Pc3VU7bHDxRqgMjP1Qz9m4gNfq3b0M7X2i/88d8f51XRUbdu3QRLXmhunQYBfUWYx3lBbKK7ZpRT9yN1wvSdkGTt2rXW0dB3ylLnq/vuu8+2y5YtW/C2O3bscFdffbVbtmyZdcVSFzzfaalevXru4MGD7s0337TLvltemjRp7P/hOjIhctTRMrTj3RNPPGH/V6dEdUC9/fbbrUOeupmpS5q6qc6ePdu2adiwof2/TJkyLn369PZvdVUcNmyYW7x4sXXEkiuuuMI1adLEOi8ievjPvd5fdTfz9D7deuutLkeOHG7kyJF2nbqd7du3z11//fVu0qRJrkSJEu6BBx5wDz74oPviiy9chgwZgt1z1UFNHTdr1arlqlev7vbs2WPdFxG9OA4kT+pyqO9vdcX0xwT//f/uu++6zp07W9frzZs3u8yZM7tGjRq5LFmy2L89db/V51ydknUc8J211VFV3wU//PCDXb7xxhvduHHj3AUXXJDgMdA1NTo7ZSbeRtTZct26dXbeN3PmTHf//fe7Nm3a2HnfddddF9yPOnXqZN1yFyxYELwPdWBXB2V1Yff7iHBeEKMiHWXi2DQRWhPelXnzNdBdunSxuTKho2saiVd2LzQ9f7yRt1mzZh1VtofokjizovJLvyip9oGmTZta2a5GbFV/r2yd3nc11tEitv79VTZPI7hqvKDuaWqkcfnllx+zZA+RES6bpqYZysbqfRs5cmRwFF8ZPM2hVSmuz8Zr7oTKtnWd5uSEZvlC17fTaK+aKaiUl8ZJ0Y3jQPKlUnqV1WnuZOh+oKYamqKhhkhaq07z7VTNo+tVkaN1b9VYJ7ScX+cQ+g7Qd4Wy/dddd10gffr0tq3mbYZKKlOIcy/0fddxPvSYnvg8z5deakqPyvRVoqv5uar60DxMdcvUd4mv+tL+oHm46rbpseZp/CDIi/IPtRacVOtyzbFR+l1zsHQC79uYh37AVZKpOmyV7p1MaQVlGNHvs88+s1p5BXV+kry6Xelkfs6cOcFAQJ0wNf9SJ+46sKu1sWh+hUpzdb1OCFSSg+jny3A1r0ZBmz7fOkFr0aJFguZIKr/U3FzfVCN16tSB9957L8F9aXFa3c4vYs3nPvZwHEheFGhpgFff/54vrdZJvJbC8FMsFMypRFPl+7pO5XbqnqjzB39ffjvNsdKi1xpEZhmE2KDAq1WrVtZMS41QevXqlaD8UsFd4mO65tXpOn2P+MFhNVjSIue+q6bK+TWlx3feRHwhyIsioa2IE39YdZKuk3O1MtYozHPPPRf8nb+NGqyoljp0yQPm2sSWxO+7Op6qU56yuffff7+N6up99u+5OiUqY+eDAZ3g6wtAGR+N7Coo9POuhDlX0U9db9XK2vMNEDTaqi95BXNXXXWVZWL9PJr77rvPRu09zcXVKK4a6Hz99dfWLU/rY6n5Ctn76MdxAJ6CNX2elXXTAI8G7nSuoADOr0326KOP2om7BoJ9J1wNEN97773BNc788T/c4A7LIEQvBW8awFXGVtlXzZ/UYK06oSvQT/y9rsFfdUNNag6dls/RfYV2S9U5BeITQV6UrG0Vep3KJkJLLEKboegDriBPrY+HDBkSHJ0JnYivzlhqqqBOiqELmyI66b0P7W4VSssZ6ORdWZhwFNCnTZs2wSLlGr3zJRk6mKvDImKDvqz15avBGgV7aqqkTmcK4hXAa5Bn4sSJFvwrc69SHDXPUZm29hOtgyeacN+jR4/AZZddZsG+/k/2NrpxHICCLX8u4MuvNeCjig0FeD47r210LNDxQdfr/2qa5c8V/H1MmzYtUKBAgWBFR7h9jox+dPrmm2+CaxYPGDAgwRqlOj9UmaWm7oROvylZsqRlczUQGHpuqC7J+tEC5qrwUZn+sZbQQfxIHek5gcmNn7ysCbShTQ50/fLly1337t3dL7/84ooXL+7KlSvnhg4dGmyGIjt37nS1a9d25cuXd2PHjrXGKZo8e+edd9rvR48e7UaMGGENGp555png9Yheeu/VMEWDLhMnTrQJ8nrv8+fP77777jtrtrB161b35Zdfuo0bN9pk6Jw5c9oEeTXeyJs3r02k1j5x3nnn2e+0b2iidZUqVewyopfeXzVR0PuvY4KaIB06dMiaKalpin7UMEcNdCZMmGAT6evWrWuT6jVBfvDgwbYPFStWzPaDVq1aWWOmQYMGuV69erm//vrLFS5cONJPE8fBcSD50mc/tJnKP//8E2yaUrVqVZcrVy47P9D3hP99nTp13JgxY9zUqVOt0ZKnRmv6ad68uR3/77nnHnfRRReF/bs0U4lOOj/UcV/ve82aNV3btm2tkY73xx9/WKOsO+64I3jdr7/+ao3Yunbt6vLkyRO8/pFHHnGbNm2yHzXZUXM1fTdkzZr1nD8vRECko8zkqFOnTlYPHZpiVwtblV8oA6cSDM29U7ZO62Bp1N7r2bOnlWr59e40kqNRf43MKOsjibNClOhFl9BJzb4VtspvtXaR9gHNu9IcDM21U0MMNVXRe6zr9D4rU6ssrS/VUAMNZe1UlofYEe5zqf1BmTnNtdH6hj6j70dlNZqrEfxXXnkl+Flfvny5lWlpYVvtB1rQGNGP4wAS02dXc6/V1r5bt27BCg41TNG6djo/8N/vWrBa3wNaE9UfJ9auXWvTOu655x7Wuo1yx8qgqtRSzdL03R/aAEfzsLW8hY7zmoetKg0tg+Cna4T7flGVj8o71bgndOoGkgeCvLMkqbp30eLFKqEIXdNOXY+0OLn37bffBjJmzGgHfL/emegDrXWwEpfsqXOS776Z+O8h8vReaJF6TZrWXCntA74BhoJznbirtMbTwV3BnN5T3VYd0zRBXmsX6QtAc+2uvPLK4Pa6/aJFiyLy3HB6wd3TTz8dqFevnn0J+/lyOllTwK8BodDSLb33OtlTqaYm0Ieueagvcp0AhA4KIbpwHEBSAzwarMuXL5995jWIo6kXGrzxgwFqkqK1bTVlQ3TCPnr0aCvP09xsddVVh22tkRq67llSfw/Rwc+rTHzuqAZ7Ot8LPZfTVAwlCBT8699qxKR9Rsd+4X1GYgR5Z5kmwapZRugHVbXQCuDU9cqPyt15553WLUnZGy1arA+uJlP7A7z/8OpkXieFodcheunkvF+/fjanQpPnNZ9KlzVJ3h/A1RLf/1uL1OpEXW2tdV1ogO/fb+0TWj4hqXkWiH4K1n777Tc7oatSpYo1VNH8SY3g+8ydMrXK5vvRV5/NU9ZXAwXK+AlzaqIfxwGEG3idMGGCdbpUJqZs2bKB1157Lfg7Nc9QxkZNN0SDgqri0QBhqCVLlticXM3bSrzIOaJH4uO0Luu7XhUbb7zxRvB6f06o80EF70nd3lPGN3Q7IBRB3lmiE3KVVukgrR81UQid6Koud2XKlLHRGE2W1kiMSvKUvRkxYkSwo574A7dG5xQcavQmqb+J6KJsrdanUwfMUBqlVcdMvc/qmqkRWm2jVsbqopi49PKLL76wydTK4GTLls26bKlcF7HTSCGUSnHUNEUZWZ95W7p0qZVc+i98lWFqwEcdMkMpg6cJ9mqBnlTDHkQXjgMI/Z7WsbtZs2b2+VYzFXU31ADP999/H3jxxRftO0M/Tz31VIKMvdbEVPdkv75puGod3T9VPNFJTfX8tBrfAVVVWDly5LBMbqgHH3zQOqqGvv+JqfJDXZZV3UUTHYRDkHeG6QAbWgutk3eVU6gcSx9YvyaNyvCU3VHwJxqJU9Cn8gvxJ4ZKy+vkX0shaER/zJgxZ/oh4yxTl0N9MStzIxq1rVy5stXUK8BXG3wNBOiEP7TMRvMsdAKgTokq39MaZ1oDya+Th9ihL3O/rp3oPdTnv0OHDgm2U/ZWCxzrvdd+ohJflWBpe43oa3sFdozYxx6OAxB1utWxX0Ger9TRAI8WPFfmXiWa+p7XfGzPT8XQYJ/Wy0w88OMx0Bu9VImhKTl67/36dOq5IL6SQ9U5/txP3dMLFSpk/w4dKFSXTA0MamBA+5DOMemgjaQQ5J0lOkHTaJoapWj+nU7KNB9LrY792lb6gOqArxN9BX9a00bZneeff94CRB3INdKnSdQ6SURsWr16tZVXqvxW8yrUYEVBneZO6H3VF73KLbQeng7gnlrpaxt9sUtS694gshSAaz7MsmXLjvqdWlarxMrPxdRn22fpNadS7axDF7HVXFztG37ejUZqlfXTcUAnB2qTjdjEcSB581l3zaHXZ1zzrkNpcEc/fq6+z8zoe0BZ+9BjCmKT5s1rySNNx9Hgv1/vWN8JGtDLnj27VXKJlrxRY53QXgs6T1QiQPOydXstlbRmzZqIPR9EP4K805S4FEsjac8++6xlZXRQ1wFZ69ioBEuj8zqw6wOueXWab6WDvRY79Sfx6p6lk0GN5qnxgjpoIfb3EZXnap6d5l+p3FbzNFWOoVILzcnTSJwGAzRKq9JdZfo0qqt5WYnXQkR0UVmlSqz0PoWOpGtCvTqiKvum91dd8JTNb9euXfBLPHXq1MHyHX8s0TbqrulH8nVcSNxUCbGH40B8O5kSSc3JVhfM0GydzhH0HaC1bTXYo5N3ZX9V4TNq1KijvgeYixu9kiqZVcZO53x6T8PRHG1ldDV3X9VdOgfU4FDoe64EgILD0O68QFII8s4gf4KnAE9zpkRz7u677z6bQ+NpNF/zqh555BH7QCuD4+uu9SFWZk/d8sLdN2KTvrTVNEcdFENpAdsMGTLYwIBGbLWYtfYHlXEQ3EW3cJ/J0BbVCuLVIS30JC5nzpyW1fPbKYOrL/5Qvnw39Msd8YHjQPxJHGxp8Dap72ufzdM5gLI2moMXSif2xYoVs0E+v8j5J598chYfPU7H8c7LVJmj6otffvnFLus878MPP7Tju3/v/TJKoqBf8zB1XqipGerT4Ku4COpxKgjyToI+ZIk/1Cqr0pp24j+oGtXXgTr0pE0ndzqh95TJ0fwb35gldG2r0A8zE6jjg/YbHbRV1ufL9fwXvr7YVYqnOXeIzVFaldGodEbd0kRBnC7rM6/jQdGiRa30Rl/gmnzvKbBXd7XEc+xC5+8hfnAciF/6TleGRifomoMdWoUT7gRdZft9+/YNLo/i/fXXXzYYoDlXoTjJj17+O91/Lyi4U1ZOzbXUVVuNVVSVo2O/jgGq6PFLZoULFDW1Q+eMOjecN2/eOX42iCcEeSco3AdRH1iVV2q0RUsleKqZrlmzps3V8fNqlN3TqFwolW/qxJ61rZIHza1TeabaZocGedq31FgH0UdZ9aROtDTqqhM5HQeUdfWtrP2cGpVc+tIclVuFdtfVKK7uW/enTK7mYyB54DgQX7SUiQbq1OFQa5jNnDnTqjEUxIWufer5476qNdRB81hNlEKzPIiscEG2MnR16tSx5Q5C6Xivhii+C6rmYeq7X520/eCeGm8l/t73f0M9HVT5EbqkBnAqCPJOIrhTUKaReE2CVsdD/4FURq5WrVrBD6Q+nOqUFFpvry6JmmPl18zzpXiMziUf+nK/7bbbbNK09iVENwVtWr8y8XulkXYN7mi5Cw3maJ0rUTZeZVYanRdl+BXkhQ4AiS5r3o3WvRI6oyUvHAfiZ46VBmr8PCvNpQ3NyqlUW42zVq5cmeC73v9ft9Xt1GiD84DYPS/Uckh6r31Jpjpfqpu6XxJBA3pqnqXlr7QEhj7/Gvj3zdZEA4CU5eJsSOlwTClTpnS7du1ya9ascVdeeaV7++233TvvvOMaN27sXn/9ddvm6aefdpUqVXLt27d3y5cvt38fOHDArVu3Lng/FSpUcKVKlXLjx4+3y+nSpbP/p0iRwh05csR+EN9Sp07tmjdv7qpWrWr7B6LPsmXLXNmyZd2QIUNcmzZt3GOPPeYyZMiQYJtBgwa57777zs2dO9e99dZbrmbNmnb9ZZdd5mrUqOHee+89988//7gbb7zRruvRo4cbMGCA++abb1zv3r1tH9AAW758+ex2+ntIPjgOxOZ5QKpUqexcYM6cOW7r1q12fZo0aVyHDh1c7ty53SWXXOLSpk3rDh06ZL/T9StXrnSrVq0Kftf7/2sb3faTTz6x8wb/O0QvvY/33nuv+/rrrxOcr91xxx22P7z//vt2Wb9bunSpvcfXXnutu+KKK1yWLFnsds8884x9/nPlyuX69u3rXnvtNXfNNde4HDly2Hnlvn37IvgMEZfOSugYo8KN1PkOeWqBryyeaC0zLXeg5Q9CqR26SrQ0WqMRPK2DEmr9+vVn+Rkg2jFiG71UNqWlCpSl8+tZipog+RF6Zd00v0Ldz8K9r+qUqxFarXHk71MNlipUqBAoVaqUHTP8HF4kXxwHov9cIPQ9UqZFDdTUIVmdrzXvbtiwYcEGSpqOUa5cueBtPXXS1pzc4zXpYH+IfppDp8yr3v/OnTsHr1dppaq5lK3zXZA1917bqowztImeGvGpnNcvlaQOy+q0TjUHzpZkl8n7/yWqCa7zozIaqZPQLMv555/vypQp477//nt3yy232HUXX3yxa9Sokdu9e7d76aWXgts+99xzNnI/ePBgG83X6F6oCy64IMHfQ/LDiG100ee3fPnyNoL666+/2vuj7F3evHndjh07LFtXvHhx+/z7z+6ePXuC2TtdDj2eKEt7+eWXuxkzZtjvLrroIvfUU0/Z/cyePdtGeJXhQ/LGcSA6HT58OHguoPfor7/+ssuvvPKKZWLmz59vn+POnTtbtn/UqFH2e2XjlLVTpY8/j1i/fr1lAP3xQf9OCvtD9PHvm8/MKvMmOvbre0MVGar8SJ8+vbv55pvdhg0b3Mcff2zb3H777bYf3H333a5o0aLB88qxY8da9laVHqJzyu7du1PNgbMmZXIL7nQw1c/BgweDv/MH33fffdddf/31lpJ/8cUX7bpixYrZdbJ69ergbSpWrGgfdn1oPX2Yn3zySdekSRMr2SxUqFDYx3Ksgz2Asyc0IOvTp4/r16+ffVnri7pgwYIuT548dgJXuHBhO4mrXbu2ldJo0Eb+++8/V7JkSSvT9PwxRaVcKsNR+c1PP/3kJk+eHNwme/bsSR4PAEQHH6BpwKdcuXKuU6dObuPGje7DDz+073R97+vYoBLs+vXr2wn7Dz/8YGXaN910k5VoPvroo27JkiWuV69eVqbnB4cRGzQ4p2DfB94qr5TSpUvbIJ7OHXVs37t3r7v11ltt32jWrJkNDGp/2Llzp32nqExTv9fvhg0bZvuOBgs0CKhtgXMh2UQb/kRswYIFrlWrVu7BBx+0+mj/oe7WrZtr27atzZ0THcRbt25tJ4W1atWyEzc/n06UpdOIvD7Qb775pl23f/9++/+kSZPc4sWLg5k7ANExQu+/uPVZXbFihY3Adu3a1a7TKKzmzc2aNcs++5p7J126dLFBH43M64te2XqN2Cuj5wdsdFI4bdo0y+5fddVVbuTIka5BgwYRe74Ajk3f7f644Onz365dO/fpp5/aIK7m2xcoUMCydJdeemmCzI5O4Lds2WLHEVGA9/fff9vcLWXvN2/ebHOuNHCE6KTzt8TVVX7+pd4/BXA6tvv3XIOCU6ZMsYE/vbf6/lAQp8F+nTvq+O8HBF999dXgd4uuu+eee6y3g+bwAedMIM75WnetO6NOeGpr3KVLF6unb9eunXXO06LDqrP2re1lzpw5tsaJb4GsZREuuuiiBGuWqP5aHbXUQjfc3/WtkgFExzzb5557zrpmSvv27W0e7V133WXd0XRZ8+fU6nrs2LHB22gujZY58HNy1QJf8+40v27w4MHWRU3LJNxwww3BhWsBxM5xwX9ufUfcAgUKBLvfyj333GPz8hPPn9P5hJ97r4Wu1WX79ttvT3DfzLeLzvWOJ0+ebOvT+nVL/T6hy+qYqW6YWgZBndK1NIbm3onOFfWd4c/vdPzXd4bWQtUyWS1btkzQkTnxOojAuRT3QZ7oA63W9WqHHq75ydy5c21C9Z49exIclK+99lq7nagNsk4ItaB1KE2Y9bcDEH305f3YY4/ZouP6Ip4yZYp9xitWrGhrXBYuXDiwbNmy4PZ+Ev0ff/wRvE4DQ8WLFw9s3LjRLqtdtr7odRKghirPPvtsRJ4bgFM/LvTr188+vxrc0cm9vssbNmwYuOSSSxIEZ/Pnz7fgT+ud+ZN7rYFWqFAhO3/wNCicNm3a4HlG4nU2EVlqjPXbb7/Zv7X0jQbqnnnmGbvs3+/+/fvbuZ6WPvBLYhUrVsyuFy2VpUG/hQsXBu9XiQIlEbSPZM+ePfDNN99E4NkByTTIUxYuXbp0gUWLFiXocOU/1EuXLrU17GbNmmWX/SiMPtzZsmULHtSfeOIJ66wXOsLnHatzFoBzTyOoOonTorNVqlSxhcu9tWvX2oi7vpgVrPlRWnnllVcCVatWDbz00kvB63bu3GkDQRq1DR2ZVbc0ALFFn39l5jTQo8+7gjj/WVZmTifride3U+WPgjodN7S2nQZ99G+/5q3P5qlDrwaKEF3U5fLqq68OVmepY6oWsFf2zR//tQaqAj8Fg35NU+0nWhN14MCBwfsqWLCgddgMHeBXZZc6qS5ZsuScPzcgWQd5WqxUI3NJ0ciORu5VshVq5MiRgSJFigQ2bdpkl3/88Udb8BhA9NPnVp/fxOVToSZNmmQLmCu75+nLXyO5yvwruPNuvvlmK+/RiRyA6JZ4GQTxA7YK0kqXLh34/fffj7rdn3/+acGfX6ja30aBgEr8FMCpPE9TPsJRpmfUqFFn4RnhdIQrm50+fbqdG/r3S9N6tBTG0KFDbdmbXLlyWXm/z/55L7zwgpVzEtAh2iWLxiuaOKvudtu2bQv7ezVI0WRYdUYaPny4LWKubdVQRdfnz5/ftlPThauvvvocP3oAp0KfW3W6VRt0NU0RNVpQB9wqVaq4L774wjrnahK9FjhWtzTJmjWrXa+uemrC4r388svWXEXLqgCIbn4ZBHXFXrRokTXX8J0S9d2uhmpqquL5Jiw6HnTs2NG9/vrrdh6g2+i26sCrlvhqrLZw4UJryBRuSaQ777zTmrAgurop+6ZbapiiBiqi7wF1TdV1ouY6asqlZQ20TJaa6KjLsjqq6vvAd1XWkhmZM2e2ZXaAaJYsgrzrrrvOffvtt/YhDUcf/sqVK9uaN+qKVa9ePVs6IUOGDNZpC0Bs0nIo+tLWukZaMkGd7qZPn24ncWqHrSUPtEyCTgQV6HnqjKnlEhTk+eVWcubMGey+CyC6qVNm9erVrTO2jgM33HCDDeSKBmr8kkh+XVy/fIJoOQQteaLlEBKvY6fzgtCgkCWRooOO04888kjwfdXlxOsPaj1UrXmqrqk6vmvZHA3cq8umX/JG3TKVFFD3VAX8oq6p6qCppTF8YLdp0yZXt27dc/48gZORLII8HbBFI3P6sCYe4dHaJXfddZdr0aKFHQAGDhxoixbrBM8vZAkg9pQoUcLWuxo8eLAtkaBMnBY11hd5mjRpbBuN2GbKlMl+70/clAXUbZ577rngdixYDEQXfY/79vah/v33X8vY67O/fPly98EHH7gyZcrY97zWs1Tre2XyVeWTNm3a4O20rZY/0oCOllRSoODX100sNChE5Ok9f+ONN2wwT3Tc/v333235G2XkfDZWC5UrgNfxXapVq+YuueSSYJCnwF4ZXg30KWOnRdA1uDdz5kxLGGjfEN5/xIRAMtGzZ09rvqJGDKIJs6q11wRbzb95+eWXj6rZVjOVxK2WAcQWzbPQHBvNw/Gf68Sf9UcffdTm22miPYDol7jZmT7nfg7tm2++aa3vPc25VRdNNU776aefbMmEK6+80ppqfPzxxzYPV/+vV6+eNVzybfYRW1599VXrbqq5lmqUon+rQU7u3LmtN4Oa6qnjaffu3W2ZDE8NU0qWLGkdN32TFi2307p1a1sia/z48RF8VsCpS6H/uGRAT1OLUWqkRyVbKuPQAqda8PLhhx+2xdD9iL3fnpF7IPZpBLd58+Y2N2/ixInBkViVcSpTly1bNtewYUO3YMGCYNYfQPTzc2xfeOEFm1ulMmzNp9XnfPbs2ZaB0e/k/vvvt2od//lfu3atZet+/vlnK8v77bff7PKAAQNcxowZbZv/35yOkswoPraHlstqDqUWsVfWNnfu3PZ+aj7dV199ZfMpH3roIdejRw/L2NavX9/16tXL9hmdCypjp8ztjBkzgveny6HnhUCsSTZBnuipfvnll1aSqeYqF154oaXjAcQ3fe41mb5r167WTKl37942BzdLliw2sV4lmwBii07iNU9KJ+squVPZtUqtx48fbwO3mnv32GOP2SCOGmWIyjQ1517zsVTipyBvw4YNNsDDCX30B3I+uA9XLqnr1SBLvRRUkqtGKWqcowF7zddTw5yhQ4dag5VOnTq5efPmuZ9++sluq34Mo0ePtsGBkiVLnrPnB5xNySrIS4pq+n0nLgDx+Rlv3LixfanrxO7iiy+2ubdk7oDopZN2neCH+27W3NrbbrvNjRgxwuZZhdLcezVWqlOnjg3meMreac6Vsjian5cY5wLRSw1yNH8ytMpKwfyKFStcjRo13C233GK/V8Cu913dU8eNGxfMxq1Zs8au01w8dUCdO3eu7T/Dhg2zxjy6nW6fN2/eSD9V4IxJFo1XEvNxrW997Ed6AMQnfcZbtmxpk+k1kq9sPgEeED3U+Egn22qS4fnBV2Xb9KOOiJ5O2nXiHxrg+UYsKsFWlm/+/PkWAKj5xoMPPuiuuOIKCxwvv/zysI+Bc4Hoo/dQHVKff/55u6z9QY3xypYta6W6Cs50bFc1xg8//GBLYqk0U1k8DegpwNN+ocot7U9//PFHcEksZXi1jeh2BHiIN8kyyPMBHa2PgeRDQZ2WUtAIP4DooYyLMmvnnXdego7WGpBRkKbgTGXW6m6odveiufUK2NQNU/Rvfbf7tfBUrqn7LVKkiJ3wf/fdd7bGnf6tck1El8TrDYqCeq1ppwoMzZlWQCdTp0615W80IKAlcZTVVadUBYIK/DX3Uud3qtbQ/qL9QvuJ9g1fipkvXz4r7VTZJhCvkmWQBwAAosP7779vWfaRI0e6dOnS2XVa006Bmpqkad6dtlHzJM2t1TIICgbVWEVzasXP0dJ6uFrsfPfu3Tag8+qrr1ogoPtTkCh+qRREj3Dz7rQmobJrmj+pDO2HH35owaCCvquuuiqYpVOjFTXXUgMVZf6UlWvdurXNs1OGT4GiSjMV8GvQwKOCC/GOIA8AAERs6oTKJ1WuOWbMGMvQ7d271xak1vy5Z5991jokKmujhai1nU7yVX533333uXfeeccar2iOlX7URVvNM0LXz1NDFv0tH9yxxll0UfZN3S01l84vWK/3SwGdAjgFZgrIPvvsM7tO22thc9GcO9Hi5SrF3LJli13WvDtl7ooXL2770hNPPOE+/fRTK+UFkov/q2sAAAA4w461HJGu10n7L7/8Yhm4QYMGWWZOSxioOVLFihWtM26XLl2smYqCubFjx9pSSMrkqJmSAj+VYSpb9+eff1rZp07o1Tk38d8iuItOet9UOqkgTfvBtGnTXPny5YPNcrQ8hpa5UKZXGV2fpXv88cdtgXNRZlcBnt/XLrroIivXVLMV3R5IjsjkAQCAszLH6nglcd98840Fgpp3lzVrVmuaocvKvinzppb3OknXnKzOnTtbxubzzz+3y379O83JUmCwdetWW/tSAR6Nw2OH5scpA6tlDzTXTh0wFbhLiRIl3MyZM13Tpk3td++++66V3SpTq6VwFATK5MmTLbt72WWX2WWVempdPAI8JGcEeQAA4MyeXKRMaSV1alGvHwVivlwyNADTvLkJEyZYZubHH3+0kjwfGGoR60WLFlnTJJVxqhOigkJlfFSmqbl5ou2V+VGmTn/jWNlDRCc1S1HppbK39erVs0XK1UVT8/EUBGr+nYI7le1u377dFrx/77333PXXX28LoGvNY83rVFAI4P+wTh4AADgj/Ny34cOHuwEDBlg3Q2Xlvv32W8vQqN194u0VkKmpSrNmzSxwU4mmaO6Vgjs1zVCmR/PxRA00dL8q2UR80H6grOzq1asteFMp5qZNmyzw13v9+uuv27IZHTp0sFJelW4q2Ff5pjpoavkN7SsA/ocgDwAAnDFah1LldZobp4XHVVqnbJy6Ho4ePTrJuXE6uVc5ppY7KFeunF2nOXgqxVSQqHI83d6X5CG+6D3W+ob60by7KVOmWOBfoEABm7epxcp79OjhZsyYYfvB1VdfHemHDEQ1Gq8AAICTmm93rHVmlV3R7y+99NL/O9FIndpa4SsDd6zmJ+qyqeycWt6rkYqog2aTJk2s8Yq6bJ7oY0DsueSSSyxLpyBOpZean6klNTSvTgGeKGOnfYQADzg+jpAAAOCE+eBKDVC0bIGfa+f/r4BO8/E0b+r333+3+VJau07dDz/44AO3Z8+esAtgq7FGq1atbL6dbuepIYsCvNBlEAjw4o/eU2Xx1G1TDVbk9ttvt86Zfh6numZqzh6A4yPIAwAAJ2zZsmXW8ELZlMcee8waYYhvdqImGsq4qPNh6dKl3YoVK9yIESPs5F1r3+k2dgKSKBOn29euXduNGjXK5cyZ86ggkGUQ4l+1atVc/vz5bRH7//77L3g9jXSAk0eQBwAAjuKzZqH++ecfa4CirJva1i9ZssQ6Z9oJRcqUFpipJPORRx5xV155patQoYKbM2eOzbHSibsyM4sXL3abN28O+4rrBF+ZP61/RrYu+VFpb/Pmza3rqhrvADh1zMkDAABHUbCmtvZqX6/mKToBz5w5s82Jqlu3rgVwWsRazVF0nRas9oGZtldpptY+05plvovm7t27rRvi8QI4lkFIvrTP3HLLLZF+GEDMI5MHAADCdsnUfDiVYyrAU/dDdTmsXr26BXjSv39/N3v2bFvTzFM2T9srEzNr1iy3bt06C/B+++0399NPP7kbbrjhuO3uKc9LvnjvgTODJRQAAEDYbFqVKlXc+eefb2uWbdu2zRYu940vfIdLBXzqiqiMnoJClXkqC6ggUXPsdHvNzfvkk0+shFNz7tRpEwBw9hDkAQCQzCmg8/PpvPXr11vwptLLTp062bp3GTNmDP7eB3Na305z7T777DPL8oX+7ptvvrEMoO5L2/hlFQAAZxdz8gAASCbCzXXzGTkf4GlNumzZstki1M8995x75plnbJ0yBXh+bp747TV/StuqsUrx4sXd8uXL7f6uu+46V7lyZfsJ/Vt6DMdaLw8AcPqYkwcAQDKe76SATPPntLxBiRIlbPFxZe527Nhh65Z16NDBPf/88+7ff/8NBniegj558MEH3ejRoy1T17BhQyvtTCxxMAkAOHsI8gAASCa0nMH999/vtm7dGrxOHS+1tp0WNldJ5n333WcNUrSUgbJ699xzj8uSJYsbNmyYba+Om57+PXDgQPspWLCge+ihh+z+7r777qP+NksiAMC5Q5AHAECcSbyQuEok5ccff3RffPGFBXveypUrbe26d955x7J4KrPUnDp1zdT1WhOvXbt21jBFGT8tqaDf79u3z+3fv9/KM7t16+Y2bNhgQZ4ydT7DBwCIDII8AADijLJmW7ZscfPnz09QpqmsnDpbKoDTwubiu17u2bPH3XbbbS5fvnzuvPPOs0XPFfDJvffeawGeFqpW+WbZsmUtUNTcvTfffNPKNUXBnQLKxGWdAIBziyAPAIA48/PPP7v8+fPbIuXdu3cPBnRazqBmzZq23p3WsJNSpUq5iRMnWoMULVyuTN/777/vrrjiCvfnn3+6jRs3Winma6+9Zhk83XbkyJG2aLWn60XBHeucAUDksYQCAABxwJdk+iBL69f98ccflpXTOnUqpaxVq5b79ddfXZs2bVzRokWtBFNz6LQenrJ4gwcPDt6fbjtu3Dh3ySWX2O/kv//+s0AQABDdyOQBABDjwZ0yaQru9OPnw2menP49YMAAW/5AXS9nzJhhwd2NN97oVqxY4ebNm+dy5szpWrVqZYuZ9+vXzy1ZssQyeU2bNrVsX6FChYJ/SwGe/3sAgOhFJg8AgDigLNvkyZOttLJevXp2nQI4zaHr2bOn69Onjy1cXrduXZtbp+suuugiWx5BHn/8cWu+osYqv/32m2vWrJldlylTpgg/MwDAySLIAwAghmm+nQK4F1980RqiKIDTfDkFeyq/fOqpp6xLpjJ4X331lWvUqJHNt9PyCOnSpbNgT5fVkVMdM9euXevKlSsXLPv069sBAGIHR20AAGKASiT9vLtQL7/8svv888+tS+aCBQssyFOAJ5qHp2UOXn/9dQvgFMypZFMZvk8//dR99NFHbvr06batAjmVdWoOni/71N8jwAOA2EOPYwAAopgCLQVdWn9Otm/f7rJnz26dLNU0Rdk5BWaXX365dcP02TfdLnPmzLao+UsvveQaN27sSpQoYdvpR0slLFy40N1xxx1h/y7LIABA7KJcEwCAKAnkjmXmzJk2R06ZtSxZsrjevXvbEgkdO3Z033zzjQWBWiJh7969VnKpkk3Nt9P6d7lz57ayTQV8Png7ePCgzb8DAMQfgjwAACIk8Xw3rUGneXK+3NIHf1qjrm/fvu722293DRo0sE6Y69atc/fdd59d98ILL7gNGzZYIxUFbgr0OnfubAuWa108zcP78ccfLeuXI0eOo8pAfZYQABAfKNcEACBCfICneXFPPPGEZd3Spk3rbrrpJmumojlyyrhNnTrV3X333RboicostYB5gQIFrNxS2bzQoFBz9EqWLBkM3iZMmGD3nTjAEwI8AIg/NF4BACBCvvzySwvGtDi5MnJqiqLmKK+++qobM2aMbaMMnZY00Jw6dcvMnz+/GzlypBsxYoRl8HyZ5/fff2+ZQN22S5cu1mlTi5yLlkHImzevZQ4BAPGPTB4AABGibpgqrVT5pYI8UQZPwd7XX39tQVnx4sVtmYTSpUu7ypUr29w6besXJl+9erWVaWqNPGXwtmzZ4u6//373yCOPBP+ODwTplAkAyQOZPAAAIkRz5cqXL+8++OCD4HWfffaZ27Ztm2X4FJQpkGvRooWVWmqOnRYpV4AnKuOcMmWKO3DggGUDhw0bZpk/H+CRuQOA5IkgDwCACFGWrmrVqm7NmjW2pp0WIW/durXLlSuXdcFUBk9ZuFatWlkg17ZtWwvqVqxYYSWZjz76qDvvvPMsEFQ2r0aNGna/WuNOyNwBQPJEd00AACJIAZvKK5cuXeoGDhxopZhaC69nz55WytmuXTvXpEkTN2/ePMvUaX7ezp07XZEiRax0s3r16rx/AIAEmJMHAEAEKXun9e7URfPKK6+0xir6eeWVV9ykSZMsyNNyCQ8++KCrVauWNVdRB85ChQqd1Dp7AIDkg3JNAAAi7MYbb3Tp06e3zpieSjZVwtm9e3drzKK5er7EUwGe5ttpjTshwAMAhKJcEwCAKNC1a1f33Xffueeee86WP9AcPGXs9u/fb3PrtMg5AAAngkweAABRQAug//XXX278+PF2WQGepEuXzgI8n7UDAOB4mJMHAEAUqFmzpi2poHl54aRKleqcPyYAQGyiXBMAAAAA4gjlmgAARBEWMAcAnC4yeQAAAAAQR8jkAQAAAEAcIcgDAAAAgDhCkAcAAAAAcYQgDwAAAADiCEEeAAAAAMQRgjwAAAAAiCMEeQAAAAAQRwjyAAAAACCOEOQBAAAAQBwhyAMAAAAAFz/+H7KuzSvmY1ocAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 900x450 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def class_balanced_metrics(run_name: str, label: str) -> dict:\n",
    "    cm = aggregate_confusion(run_name)\n",
    "    tn, fp, fn, tp = cm.ravel()\n",
    "\n",
    "    real_recall = tn / (tn + fp) if (tn + fp) else np.nan\n",
    "    fake_recall = tp / (tp + fn) if (tp + fn) else np.nan\n",
    "    balanced_accuracy = np.nanmean([real_recall, fake_recall])\n",
    "\n",
    "    real_precision = tn / (tn + fn) if (tn + fn) else np.nan\n",
    "    fake_precision = tp / (tp + fp) if (tp + fp) else np.nan\n",
    "    real_f1 = 2 * real_precision * real_recall / (real_precision + real_recall) if (real_precision + real_recall) else np.nan\n",
    "    fake_f1 = 2 * fake_precision * fake_recall / (fake_precision + fake_recall) if (fake_precision + fake_recall) else np.nan\n",
    "    macro_f1 = np.nanmean([real_f1, fake_f1])\n",
    "\n",
    "    return {\n",
    "        \"label\": label,\n",
    "        \"run\": run_name,\n",
    "        \"real_recall\": real_recall,\n",
    "        \"fake_recall\": fake_recall,\n",
    "        \"balanced_accuracy\": balanced_accuracy,\n",
    "        \"real_f1\": real_f1,\n",
    "        \"fake_f1\": fake_f1,\n",
    "        \"macro_f1\": macro_f1,\n",
    "        \"standard_accuracy\": summary_df.set_index(\"run\").loc[run_name, \"accuracy\"],\n",
    "        \"standard_f1\": summary_df.set_index(\"run\").loc[run_name, \"f1\"],\n",
    "    }\n",
    "\n",
    "\n",
    "balanced_df = pd.DataFrame([\n",
    "    class_balanced_metrics(run, label)\n",
    "    for run, label in RUN_LABELS.items()\n",
    "]).set_index(\"label\").reindex(order).reset_index()\n",
    "\n",
    "display(\n",
    "    balanced_df[[\n",
    "        \"label\", \"run\", \"real_recall\", \"fake_recall\", \"balanced_accuracy\",\n",
    "        \"real_f1\", \"fake_f1\", \"macro_f1\", \"standard_accuracy\", \"standard_f1\",\n",
    "    ]]\n",
    "    .style.format({\n",
    "        \"real_recall\": \"{:.4f}\",\n",
    "        \"fake_recall\": \"{:.4f}\",\n",
    "        \"balanced_accuracy\": \"{:.4f}\",\n",
    "        \"real_f1\": \"{:.4f}\",\n",
    "        \"fake_f1\": \"{:.4f}\",\n",
    "        \"macro_f1\": \"{:.4f}\",\n",
    "        \"standard_accuracy\": \"{:.4f}\",\n",
    "        \"standard_f1\": \"{:.4f}\",\n",
    "    })\n",
    ")\n",
    "\n",
    "fig, axes = plt.subplots(1, 2, figsize=(12, 4.5), sharey=True)\n",
    "plot_df = balanced_df.iloc[::-1]\n",
    "\n",
    "axes[0].barh(plot_df[\"label\"], plot_df[\"balanced_accuracy\"], color=\"#4C78A8\", alpha=0.9)\n",
    "axes[0].set_xlim(0.86, 1.0)\n",
    "axes[0].set_xlabel(\"Balanced accuracy\")\n",
    "axes[0].set_title(\"Equal class weight: accuracy\")\n",
    "for y, (_, row) in enumerate(plot_df.iterrows()):\n",
    "    axes[0].text(row[\"balanced_accuracy\"] + 0.002, y, f\"{row['balanced_accuracy']:.3f}\", va=\"center\", fontsize=8)\n",
    "\n",
    "axes[1].barh(plot_df[\"label\"], plot_df[\"macro_f1\"], color=\"#54A24B\", alpha=0.9)\n",
    "axes[1].set_xlim(0.86, 1.0)\n",
    "axes[1].set_xlabel(\"Macro F1\")\n",
    "axes[1].set_title(\"Equal class weight: F1\")\n",
    "for y, (_, row) in enumerate(plot_df.iterrows()):\n",
    "    axes[1].text(row[\"macro_f1\"] + 0.002, y, f\"{row['macro_f1']:.3f}\", va=\"center\", fontsize=8)\n",
    "\n",
    "fig.tight_layout()\n",
    "fig.savefig(FIGURES_DIR / \"06_phase3_balanced_metrics.png\", dpi=180, bbox_inches=\"tight\")\n",
    "plt.show()\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(9, 4.5))\n",
    "x = np.arange(len(balanced_df))\n",
    "width = 0.35\n",
    "ax.bar(x - width / 2, balanced_df[\"balanced_accuracy\"], width, label=\"balanced accuracy\", color=\"#4C78A8\", alpha=0.9)\n",
    "ax.bar(x + width / 2, balanced_df[\"macro_f1\"], width, label=\"macro F1\", color=\"#54A24B\", alpha=0.9)\n",
    "ax.set_xticks(x)\n",
    "ax.set_xticklabels(balanced_df[\"label\"], rotation=25, ha=\"right\")\n",
    "ax.set_ylim(0.86, 1.0)\n",
    "ax.set_ylabel(\"score\")\n",
    "ax.set_title(\"Balanced metrics by model family\")\n",
    "ax.legend()\n",
    "fig.tight_layout()\n",
    "fig.savefig(FIGURES_DIR / \"06_phase3_balanced_metrics_grouped.png\", dpi=180, bbox_inches=\"tight\")\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "12612891",
   "metadata": {},
   "source": [
    "The practical Phase 3 gain is fewer missed fakes while keeping real-image false alarms in view. This is the decisive part of the model-family comparison. ConvNeXt-Tiny has the strongest AUC, but ResNet50 has the best accuracy/F1 profile and a better confusion-matrix tradeoff: fewer false positives on real images than ConvNeXt-Tiny (`545` vs `697`) while keeping false negatives almost the same (`578` vs `567`).\n",
    "\n",
    "The balanced metrics reinforce that ResNet50 is a stronger final choice than the small AUC difference alone suggests. EfficientNet-B0 is competitive and has slightly fewer real false alarms, but it misses many more fakes (`895` false negatives), which is a worse detector tradeoff. For this task, ResNet50 is therefore the selected Phase 3 model: almost the same ranking quality as ConvNeXt-Tiny, but better operational behavior.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "33334c4d",
   "metadata": {},
   "source": [
    "## Conclusion\n",
    "\n",
    "Phase 3 confirms that architecture choice matters after preprocessing has been fixed. The selected Phase 2 ResNet18 facecrop run reaches AUC `0.9755`. ConvNeXt-Tiny reaches the highest mean AUC at `0.9868`, but the difference from ResNet50 is very small (`0.9857`). ResNet50 is the best overall model because it has the strongest logged accuracy (`0.9532`) and F1 (`0.9688`), and its confusion matrix gives the best practical balance between false alarms on real images and missed fake images.\n",
    "\n",
    "The report decision should therefore be: Phase 2 found the right input pipeline, and Phase 3 selects ResNet50 as the best classifier backbone for this task. ConvNeXt-Tiny is the AUC winner, but ResNet50 is the better detector when the full metric set is considered. The remaining caution is unchanged: source-wise and pairwise behavior must stay in the analysis because high global AUC alone does not prove source-agnostic generalization.\n",
    "\n",
    "Next: `07_data_scaling.ipynb` turns from architecture choice to data scaling. It asks whether the selected top model families, especially ResNet50, improve further when trained with more of the available data.\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": ".venv",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}