training.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "f7241c26",
   "metadata": {},
   "source": [
    "**Imports**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "c1cfd05a",
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "from torch import nn\n",
    "\n",
    "import pandas as pd\n",
    "from torch import optim\n",
    "from torch.utils.data import DataLoader, Dataset\n",
    "from torchvision import datasets, transforms\n",
    "from sklearn.model_selection import train_test_split\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np \n",
    "import random \n",
    "import timeit\n",
    "from tqdm.auto import tqdm"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9a727ce9",
   "metadata": {},
   "source": [
    "**Hyperparameters**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "babd6040",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Device: mps!\n"
     ]
    }
   ],
   "source": [
    "random_seed = 42\n",
    "\n",
    "batch_size = 512\n",
    "\n",
    "num_classes = 46\n",
    "patch_size = 4\n",
    "img_size = 32\n",
    "in_channels = 1\n",
    "num_heads = 8\n",
    "dropout = 0.001\n",
    "hidden_dim = 1024\n",
    "adam_weight_decay = 0\n",
    "adam_betas = (0.9, 0.999)\n",
    "activation = \"gelu\"\n",
    "num_encoders = 4\n",
    "embed_dim = (patch_size ** 2) * in_channels # 16\n",
    "num_patches = (img_size // patch_size) ** 2 # 64\n",
    "\n",
    "random.seed(random_seed)\n",
    "np.random.seed(random_seed)\n",
    "torch.manual_seed(random_seed)\n",
    "torch.mps.manual_seed(random_seed)\n",
    "torch.backends.cudnn.deterministic = True\n",
    "torch.backends.cudnn.benchmark = False\n",
    "\n",
    "\n",
    "device = \"mps\" if torch.backends.mps.is_available() else \"cpu\"\n",
    "\n",
    "print(f\"Device: {device}!\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "a175c32d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([512, 65, 16])\n"
     ]
    }
   ],
   "source": [
    "from patch_embedding import PatchEmbedding\n",
    "\n",
    "model = PatchEmbedding(embed_dim, patch_size, num_patches, dropout, in_channels).to(device)\n",
    "\n",
    "x = torch.randn(512, 1, img_size, img_size).to(device)\n",
    "\n",
    "print(model(x).shape)\n",
    "\n",
    "# expected: torch.Size([512, num_patches + 1, embed_dim]) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "29cc899e",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Library/Frameworks/Python.framework/Versions/3.9/lib/python3.9/site-packages/torch/nn/modules/transformer.py:282: UserWarning: enable_nested_tensor is True, but self.use_nested_tensor is False because encoder_layer.norm_first was True\n",
      "  warnings.warn(f\"enable_nested_tensor is True, but self.use_nested_tensor is False because {why_not_sparsity_fast_path}\")\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([512, 46])\n"
     ]
    }
   ],
   "source": [
    "from vision_transformer import VisionTransformerKAN\n",
    "    \n",
    "model = VisionTransformerKAN(num_patches, img_size, num_classes, patch_size, embed_dim, num_encoders, num_heads, hidden_dim, dropout, activation, in_channels).to(device)\n",
    "\n",
    "x = torch.randn(512, 1, img_size, img_size).to(device)\n",
    "\n",
    "print(model(x).shape)\n",
    "\n",
    "# expected: torch.Size([512, num_classes])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "79c499d3",
   "metadata": {},
   "source": [
    "**Loading Data**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "9f0fbb09",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_dir = \"data/train\"\n",
    "test_dir = \"data/test\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "13e7600f",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_transform = transforms.Compose([\n",
    "    transforms.Grayscale(),\n",
    "    transforms.RandomRotation(15), # Random Rotation by 15 degs\n",
    "    transforms.ToTensor(), # this also converts all pixel values from 0 to 255 to be between 0.0 and 1.0 \n",
    "    transforms.Normalize([0.5], [0.5])\n",
    "])\n",
    "test_transform = transforms.Compose([\n",
    "    transforms.Grayscale(),\n",
    "    transforms.ToTensor(),\n",
    "    transforms.Normalize([0.5], [0.5])\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "53038067",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train data:\n",
      "Dataset ImageFolder\n",
      "    Number of datapoints: 78200\n",
      "    Root location: data/train\n",
      "    StandardTransform\n",
      "Transform: Compose(\n",
      "               Grayscale(num_output_channels=1)\n",
      "               RandomRotation(degrees=[-15.0, 15.0], interpolation=nearest, expand=False, fill=0)\n",
      "               ToTensor()\n",
      "               Normalize(mean=[0.5], std=[0.5])\n",
      "           )\n",
      "Test data:\n",
      "Dataset ImageFolder\n",
      "    Number of datapoints: 13800\n",
      "    Root location: data/test\n",
      "    StandardTransform\n",
      "Transform: Compose(\n",
      "               Grayscale(num_output_channels=1)\n",
      "               ToTensor()\n",
      "               Normalize(mean=[0.5], std=[0.5])\n",
      "           )\n"
     ]
    }
   ],
   "source": [
    "train_data = datasets.ImageFolder(root=train_dir, # target folder of images\n",
    "                                  transform=train_transform, # transforms to perform on data (images)\n",
    "                                  target_transform=None) # transforms to perform on labels (if necessary)\n",
    "\n",
    "test_data = datasets.ImageFolder(root=test_dir, transform = test_transform)\n",
    "\n",
    "print(f\"Train data:\\n{train_data}\\nTest data:\\n{test_data}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "e19c7c54",
   "metadata": {},
   "outputs": [],
   "source": [
    "class2idx = train_data.class_to_idx\n",
    "idx2class = {k:v for v, k in class2idx.items()}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "b1d2b093",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "character_19_dha\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "rand_int = np.random.randint(0, high = len(train_data))\n",
    "\n",
    "print(idx2class[train_data[rand_int][1]])\n",
    "_ = plt.imshow(train_data[rand_int][0].permute(1, 2, 0), cmap = \"gray\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "a8600c2d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(<torch.utils.data.dataloader.DataLoader at 0x4729c5100>,\n",
       " <torch.utils.data.dataloader.DataLoader at 0x4729c9520>)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_dataloader = DataLoader(dataset=train_data, \n",
    "                              batch_size=batch_size, \n",
    "                              shuffle=True) \n",
    "\n",
    "test_dataloader = DataLoader(dataset=test_data, \n",
    "                             batch_size=batch_size,  \n",
    "                             shuffle=False) \n",
    "\n",
    "train_dataloader, test_dataloader"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "642f652e",
   "metadata": {},
   "source": [
    "**Training**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "2505cbb4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a97e42bc887647918e9aad925b41abf5",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/50 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Library/Frameworks/Python.framework/Versions/3.9/lib/python3.9/site-packages/torch/optim/lr_scheduler.py:149: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.\n",
      "  warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1: Train Loss: 3.6388 | Val Loss: 3.2302 | Train Acc: 0.0488 | Val Acc: 0.1042\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 2: Train Loss: 3.0489 | Val Loss: 2.7634 | Train Acc: 0.1451 | Val Acc: 0.2080\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 3: Train Loss: 2.6511 | Val Loss: 2.3493 | Train Acc: 0.2440 | Val Acc: 0.3128\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 4: Train Loss: 2.2254 | Val Loss: 1.8520 | Train Acc: 0.3543 | Val Acc: 0.4496\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 5: Train Loss: 1.8239 | Val Loss: 1.5546 | Train Acc: 0.4625 | Val Acc: 0.5321\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 6: Train Loss: 1.5199 | Val Loss: 1.1870 | Train Acc: 0.5454 | Val Acc: 0.6415\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 7: Train Loss: 1.3096 | Val Loss: 1.0286 | Train Acc: 0.6051 | Val Acc: 0.6871\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 8: Train Loss: 1.1242 | Val Loss: 0.9040 | Train Acc: 0.6601 | Val Acc: 0.7270\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 9: Train Loss: 0.9786 | Val Loss: 0.7406 | Train Acc: 0.7026 | Val Acc: 0.7747\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 10: Train Loss: 0.8644 | Val Loss: 0.6793 | Train Acc: 0.7374 | Val Acc: 0.7904\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 11: Train Loss: 0.7856 | Val Loss: 0.6461 | Train Acc: 0.7597 | Val Acc: 0.8016\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 12: Train Loss: 0.7110 | Val Loss: 0.5016 | Train Acc: 0.7832 | Val Acc: 0.8489\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 13: Train Loss: 0.6529 | Val Loss: 0.4800 | Train Acc: 0.8009 | Val Acc: 0.8560\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 14: Train Loss: 0.6125 | Val Loss: 0.4544 | Train Acc: 0.8123 | Val Acc: 0.8580\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 15: Train Loss: 0.5510 | Val Loss: 0.4336 | Train Acc: 0.8308 | Val Acc: 0.8669\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 16: Train Loss: 0.5261 | Val Loss: 0.3815 | Train Acc: 0.8374 | Val Acc: 0.8841\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 17: Train Loss: 0.4961 | Val Loss: 0.3603 | Train Acc: 0.8454 | Val Acc: 0.8891\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 18: Train Loss: 0.4589 | Val Loss: 0.3697 | Train Acc: 0.8582 | Val Acc: 0.8867\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 19: Train Loss: 0.4385 | Val Loss: 0.3346 | Train Acc: 0.8645 | Val Acc: 0.8988\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 20: Train Loss: 0.4164 | Val Loss: 0.3235 | Train Acc: 0.8699 | Val Acc: 0.9019\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 21: Train Loss: 0.3979 | Val Loss: 0.3033 | Train Acc: 0.8764 | Val Acc: 0.9059\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 22: Train Loss: 0.3698 | Val Loss: 0.2976 | Train Acc: 0.8850 | Val Acc: 0.9109\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 23: Train Loss: 0.3639 | Val Loss: 0.3224 | Train Acc: 0.8856 | Val Acc: 0.8995\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 24: Train Loss: 0.3543 | Val Loss: 0.2967 | Train Acc: 0.8893 | Val Acc: 0.9099\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 25: Train Loss: 0.3417 | Val Loss: 0.2578 | Train Acc: 0.8931 | Val Acc: 0.9200\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 26: Train Loss: 0.3215 | Val Loss: 0.2529 | Train Acc: 0.8999 | Val Acc: 0.9230\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 27: Train Loss: 0.3068 | Val Loss: 0.2243 | Train Acc: 0.9031 | Val Acc: 0.9333\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 28: Train Loss: 0.3027 | Val Loss: 0.2357 | Train Acc: 0.9047 | Val Acc: 0.9288\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 29: Train Loss: 0.2863 | Val Loss: 0.2169 | Train Acc: 0.9109 | Val Acc: 0.9340\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 30: Train Loss: 0.2842 | Val Loss: 0.2163 | Train Acc: 0.9107 | Val Acc: 0.9346\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 31: Train Loss: 0.2805 | Val Loss: 0.2083 | Train Acc: 0.9115 | Val Acc: 0.9340\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 32: Train Loss: 0.2701 | Val Loss: 0.1970 | Train Acc: 0.9150 | Val Acc: 0.9415\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 33: Train Loss: 0.2626 | Val Loss: 0.2014 | Train Acc: 0.9161 | Val Acc: 0.9388\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 34: Train Loss: 0.2540 | Val Loss: 0.2227 | Train Acc: 0.9205 | Val Acc: 0.9310\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 35: Train Loss: 0.2427 | Val Loss: 0.2009 | Train Acc: 0.9231 | Val Acc: 0.9391\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 36: Train Loss: 0.2480 | Val Loss: 0.2128 | Train Acc: 0.9203 | Val Acc: 0.9353\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 37: Train Loss: 0.2372 | Val Loss: 0.1963 | Train Acc: 0.9241 | Val Acc: 0.9388\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 38: Train Loss: 0.2284 | Val Loss: 0.1981 | Train Acc: 0.9272 | Val Acc: 0.9381\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 39: Train Loss: 0.2396 | Val Loss: 0.2127 | Train Acc: 0.9241 | Val Acc: 0.9336\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 40: Train Loss: 0.2197 | Val Loss: 0.1840 | Train Acc: 0.9306 | Val Acc: 0.9438\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 41: Train Loss: 0.2205 | Val Loss: 0.1949 | Train Acc: 0.9300 | Val Acc: 0.9433\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 42: Train Loss: 0.2074 | Val Loss: 0.1663 | Train Acc: 0.9342 | Val Acc: 0.9485\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 43: Train Loss: 0.2062 | Val Loss: 0.1876 | Train Acc: 0.9337 | Val Acc: 0.9426\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 44: Train Loss: 0.2065 | Val Loss: 0.1933 | Train Acc: 0.9335 | Val Acc: 0.9395\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 45: Train Loss: 0.1992 | Val Loss: 0.2259 | Train Acc: 0.9363 | Val Acc: 0.9311\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 46: Train Loss: 0.1910 | Val Loss: 0.1822 | Train Acc: 0.9387 | Val Acc: 0.9436\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 47: Train Loss: 0.1889 | Val Loss: 0.1831 | Train Acc: 0.9394 | Val Acc: 0.9442\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 48: Train Loss: 0.1897 | Val Loss: 0.1601 | Train Acc: 0.9390 | Val Acc: 0.9511\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 49: Train Loss: 0.1936 | Val Loss: 0.1684 | Train Acc: 0.9378 | Val Acc: 0.9470\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 50: Train Loss: 0.1773 | Val Loss: 0.1495 | Train Acc: 0.9424 | Val Acc: 0.9538\r",
      "Training Time: 4804.64s\n"
     ]
    }
   ],
   "source": [
    "learning_rate = 5e-3\n",
    "num_epochs = 50\n",
    "\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "\n",
    "optimizer = optim.Adam(model.parameters(), betas = adam_betas, lr = learning_rate, weight_decay = adam_weight_decay)\n",
    "\n",
    "scheduler = optim.lr_scheduler.StepLR(optimizer = optimizer, step_size = 5, gamma = 0.95)\n",
    "\n",
    "start = timeit.default_timer()\n",
    "\n",
    "history = {\"train_loss\": [],\n",
    "          \"val_loss\": [],\n",
    "          \"train_acc\": [],\n",
    "          \"val_acc\": []}\n",
    "\n",
    "for epoch in tqdm(range(num_epochs), position = 0, leave = True):\n",
    "    model.train()\n",
    "    \n",
    "    train_labels = []\n",
    "    train_preds = []\n",
    "    \n",
    "    train_running_loss = 0\n",
    "    \n",
    "    for idx, img_label in enumerate(tqdm(train_dataloader, position = 0, leave = False)):\n",
    "        img = img_label[0].float().to(device)\n",
    "        label = img_label[1].type(torch.uint8).to(device)\n",
    "        \n",
    "        y_pred = model(img)\n",
    "        y_pred_label = torch.argmax(y_pred, dim = 1)\n",
    "        \n",
    "        train_labels.extend(label.cpu().detach())\n",
    "        train_preds.extend(y_pred_label.cpu().detach())\n",
    "        \n",
    "        loss = criterion(y_pred, label)\n",
    "        \n",
    "        optimizer.zero_grad()\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        \n",
    "        train_running_loss += loss.item()\n",
    "        \n",
    "    train_loss = train_running_loss / (idx + 1)\n",
    "    scheduler.step(loss)\n",
    "    \n",
    "\n",
    "    model.eval()\n",
    "\n",
    "    val_labels = []\n",
    "    val_preds = []\n",
    "    val_running_loss = 0\n",
    "\n",
    "    with torch.no_grad():\n",
    "        for idx, img_label in enumerate(tqdm(test_dataloader, position = 0, leave = False)):\n",
    "            img = img_label[0].float().to(device)\n",
    "            label = img_label[1].type(torch.uint8).to(device)\n",
    "\n",
    "            y_pred = model(img)\n",
    "            y_pred_label = torch.argmax(y_pred, dim = 1)\n",
    "\n",
    "            val_labels.extend(label.cpu().detach())\n",
    "            val_preds.extend(y_pred_label.cpu().detach())\n",
    "\n",
    "            loss = criterion(y_pred, label)\n",
    "\n",
    "            val_running_loss += loss.item()\n",
    "\n",
    "        val_loss = val_running_loss / (idx + 1)\n",
    "\n",
    "        train_acc = sum(1 for x, y in zip(train_preds, train_labels) if x == y) / len(train_labels)\n",
    "        val_acc = sum(1 for x, y in zip(val_preds, val_labels) if x == y) / len(val_labels)\n",
    "\n",
    "        print(f\"Epoch {epoch + 1}: Train Loss: {train_loss:.4f} | Val Loss: {val_loss:.4f} | Train Acc: {train_acc:.4f} | Val Acc: {val_acc:.4f}\", end=\"\\r\")\n",
    "\n",
    "        history[\"train_loss\"].append(train_loss)\n",
    "        history[\"val_loss\"].append(val_loss)\n",
    "        history[\"train_acc\"].append(train_acc)\n",
    "        history[\"val_acc\"].append(val_acc)\n",
    "            \n",
    "        \n",
    "stop = timeit.default_timer()\n",
    "print()\n",
    "print(f\"Training Time: {stop-start:.2f}s\")\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "383d2ce8",
   "metadata": {},
   "outputs": [],
   "source": [
    "torch.save(model.state_dict(), f\"models/kan-hindi-character-recognition-{num_epochs}epochs\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "d2d21ced",
   "metadata": {},
   "outputs": [],
   "source": [
    "torch.save({\n",
    "            'epoch': epoch,\n",
    "            'model_state_dict': model.state_dict(),\n",
    "            'optimizer_state_dict': optimizer.state_dict(),\n",
    "            'loss': loss,\n",
    "            }, f\"models/kan-hindi-character-recognition-{num_epochs}epochs-to-train.pt\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "52cfc758",
   "metadata": {},
   "source": [
    "Code to load checkpoint for further training\n",
    "```\n",
    "checkpoint = torch.load(\"models/hindi-character-recognition-100epochs-to-train.pt\")\n",
    "\n",
    "model.load_state_dict(checkpoint['model_state_dict'])\n",
    "optimizer.load_state_dict(checkpoint['optimizer_state_dict'])\n",
    "epoch = checkpoint['epoch']\n",
    "loss = checkpoint['loss']\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "f5260332",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Axes(0.125,0.53;0.775x0.35)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(2)\n",
    "history = pd.DataFrame(history)\n",
    "\n",
    "fig.suptitle('Training History')\n",
    "print(ax[0])\n",
    "\n",
    "ax[0].plot(range(1, len(history)+1), history[\"train_loss\"], color = 'b', label = \"Training Loss\")\n",
    "ax[0].plot(range(1, len(history)+1), history[\"val_loss\"], color = 'r', label = \"Validation Loss\")\n",
    "\n",
    "ax[1].plot(range(1, len(history)+1), history[\"train_acc\"], color = 'b', label = \"Training Accuracy\")\n",
    "ax[1].plot(range(1, len(history)+1), history[\"val_acc\"], color = 'r', label = \"Validation Accuracy\")\n",
    "\n",
    "ax[0].legend()\n",
    "ax[1].legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "b5d5ac1d",
   "metadata": {},
   "outputs": [],
   "source": [
    "history.to_csv(\"history-kan-100.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "29510b6e",
   "metadata": {},
   "source": [
    "### Resume from Checkpoint"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "7f9cb4a2",
   "metadata": {},
   "outputs": [],
   "source": [
    "checkpoint = torch.load(\"models/kan-hindi-character-recognition-70epochs-to-train.pt\")\n",
    "\n",
    "model.load_state_dict(checkpoint['model_state_dict'])\n",
    "optimizer.load_state_dict(checkpoint['optimizer_state_dict'])\n",
    "prev_epoch = checkpoint['epoch'] + 1\n",
    "loss = checkpoint['loss']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "4b965c57",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f23640d40b8341faa8cf20175806ecea",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/10 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Library/Frameworks/Python.framework/Versions/3.9/lib/python3.9/site-packages/torch/optim/lr_scheduler.py:149: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.\n",
      "  warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 71: Train Loss: 0.1202 | Val Loss: 0.1227 | Train Acc: 0.9605 | Val Acc: 0.9636\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 72: Train Loss: 0.1164 | Val Loss: 0.1364 | Train Acc: 0.9624 | Val Acc: 0.9594\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 73: Train Loss: 0.1127 | Val Loss: 0.1088 | Train Acc: 0.9632 | Val Acc: 0.9669\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 74: Train Loss: 0.1097 | Val Loss: 0.1074 | Train Acc: 0.9636 | Val Acc: 0.9692\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 75: Train Loss: 0.1230 | Val Loss: 0.1192 | Train Acc: 0.9600 | Val Acc: 0.9651\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 76: Train Loss: 0.1144 | Val Loss: 0.1100 | Train Acc: 0.9624 | Val Acc: 0.9676\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 77: Train Loss: 0.1064 | Val Loss: 0.1272 | Train Acc: 0.9654 | Val Acc: 0.9621\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 78: Train Loss: 0.1068 | Val Loss: 0.1048 | Train Acc: 0.9650 | Val Acc: 0.9693\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 79: Train Loss: 0.1099 | Val Loss: 0.1061 | Train Acc: 0.9642 | Val Acc: 0.9686\r"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/153 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/27 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 80: Train Loss: 0.1039 | Val Loss: 0.1147 | Train Acc: 0.9649 | Val Acc: 0.9670\r\n",
      "Training Time: 1047.24s\n"
     ]
    }
   ],
   "source": [
    "num_epochs = 10\n",
    "\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "\n",
    "scheduler = optim.lr_scheduler.StepLR(optimizer = optimizer, step_size = 1, gamma = 0.9)\n",
    "\n",
    "start = timeit.default_timer()\n",
    "\n",
    "history = {\"train_loss\": [],\n",
    "          \"val_loss\": [],\n",
    "          \"train_acc\": [],\n",
    "          \"val_acc\": []}\n",
    "\n",
    "for epoch in tqdm(range(num_epochs), position = 0, leave = True):\n",
    "    model.train()\n",
    "    \n",
    "    train_labels = []\n",
    "    train_preds = []\n",
    "    \n",
    "    train_running_loss = 0\n",
    "    \n",
    "    for idx, img_label in enumerate(tqdm(train_dataloader, position = 0, leave = False)):\n",
    "        img = img_label[0].float().to(device)\n",
    "        label = img_label[1].type(torch.uint8).to(device)\n",
    "        \n",
    "        y_pred = model(img)\n",
    "        y_pred_label = torch.argmax(y_pred, dim = 1)\n",
    "        \n",
    "        train_labels.extend(label.cpu().detach())\n",
    "        train_preds.extend(y_pred_label.cpu().detach())\n",
    "        \n",
    "        loss = criterion(y_pred, label)\n",
    "        \n",
    "        optimizer.zero_grad()\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        \n",
    "        train_running_loss += loss.item()\n",
    "        \n",
    "    train_loss = train_running_loss / (idx + 1)\n",
    "    scheduler.step(loss)\n",
    "    \n",
    "\n",
    "    model.eval()\n",
    "\n",
    "    val_labels = []\n",
    "    val_preds = []\n",
    "    val_running_loss = 0\n",
    "\n",
    "    with torch.no_grad():\n",
    "        for idx, img_label in enumerate(tqdm(test_dataloader, position = 0, leave = False)):\n",
    "            img = img_label[0].float().to(device)\n",
    "            label = img_label[1].type(torch.uint8).to(device)\n",
    "\n",
    "            y_pred = model(img)\n",
    "            y_pred_label = torch.argmax(y_pred, dim = 1)\n",
    "\n",
    "            val_labels.extend(label.cpu().detach())\n",
    "            val_preds.extend(y_pred_label.cpu().detach())\n",
    "\n",
    "            loss = criterion(y_pred, label)\n",
    "\n",
    "            val_running_loss += loss.item()\n",
    "\n",
    "        val_loss = val_running_loss / (idx + 1)\n",
    "\n",
    "        train_acc = sum(1 for x, y in zip(train_preds, train_labels) if x == y) / len(train_labels)\n",
    "        val_acc = sum(1 for x, y in zip(val_preds, val_labels) if x == y) / len(val_labels)\n",
    "\n",
    "        print(f\"Epoch {prev_epoch + epoch + 1}: Train Loss: {train_loss:.4f} | Val Loss: {val_loss:.4f} | Train Acc: {train_acc:.4f} | Val Acc: {val_acc:.4f}\", end=\"\\r\")\n",
    "\n",
    "        history[\"train_loss\"].append(train_loss)\n",
    "        history[\"val_loss\"].append(val_loss)\n",
    "        history[\"train_acc\"].append(train_acc)\n",
    "        history[\"val_acc\"].append(val_acc)\n",
    "            \n",
    "        \n",
    "stop = timeit.default_timer()\n",
    "print()\n",
    "print(f\"Training Time: {stop-start:.2f}s\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "c477a5c1",
   "metadata": {},
   "outputs": [],
   "source": [
    "torch.save({\n",
    "            'epoch': prev_epoch+epoch,\n",
    "            'model_state_dict': model.state_dict(),\n",
    "            'optimizer_state_dict': optimizer.state_dict(),\n",
    "            'loss': loss,\n",
    "            }, f\"models/kan-hindi-character-recognition-{prev_epoch+num_epochs}epochs-to-train.pt\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "2391c343",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Axes(0.125,0.53;0.775x0.35)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjUAAAHNCAYAAADmPIpxAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8g+/7EAAAACXBIWXMAAA9hAAAPYQGoP6dpAAC630lEQVR4nOzdd3iT1RcH8G+6W7qYpWW07L0EigypIspQrKCAbESmTNkoSxABQYYMGbJEhihbZMneu6wCLavMlt0C3cn5/XF+SVqa7qRvkp7P8+Rpxpv3PWnavCf3nnuviogIQgghhBAWzkbpAIQQQgghjEGSGiGEEEJYBUlqhBBCCGEVJKkRQgghhFWQpEYIIYQQVkGSGiGEEEJYBUlqhBBCCGEVJKkRQgghhFWQpEYIIYQQVkGSGiGETteuXeHn55el544fPx4qlcq4AZnI7du3oVKpsHz5cqVDEUIYkSQ1QlgAlUqVocv+/fuVDlURXbt2haura6qPq1Qq9OvXL9vHmT9/viRCQpgxO6UDEEKkb+XKlclu//7779i9e3eK+ytUqJCt4yxevBgajSZLzx09ejRGjhyZrePnFF9fX8TExMDe3j5Tz5s/fz4KFCiArl27miYwIUS2SFIjhAXo2LFjstvHjx/H7t27U9z/pujoaLi4uGT4OJk9ySdlZ2cHOzvL+EhRqVRwcnJSOgwAQGxsLBwcHGBjIw3nQmSX/BcJYSXeffddVK5cGWfOnEHDhg3h4uKCb7/9FgCwefNmfPTRR/Dx8YGjoyNKlSqFiRMnQq1WJ9vHmzU12tqT6dOnY9GiRShVqhQcHR1Ru3ZtnDp1KtlzDdXUaLt9Nm3ahMqVK8PR0RGVKlXCjh07UsS/f/9+1KpVC05OTihVqhQWLlxosjodQzU14eHh+PLLL1G0aFE4OjrC29sbgYGBuH37NgDAz88Ply9fxoEDB3Tdfe+++67u+Tdv3kTr1q2RL18+uLi44O2338a2bdtSvEaVSoW1a9di9OjRKFKkCFxcXBAUFASVSoWZM2emiPXo0aNQqVRYs2aN0X8PQlgby/haJYTIkKdPn6JZs2b44osv0LFjR3h5eQEAli9fDldXVwwePBiurq7Yu3cvxo4di6ioKEybNi3d/a5evRovX75Er169oFKp8NNPP6FVq1a4efNmuq07hw8fxoYNG/D111/Dzc0Nv/zyCz777DPcuXMH+fPnBwCcO3cOTZs2hbe3N77//nuo1WpMmDABBQsWzNTrf/LkSaa2T+qzzz7D5cuX0b9/f/j5+eHRo0fYvXs37ty5Az8/P8yaNQv9+/eHq6srvvvuOwDQ/X4jIiJQr149REdHY8CAAcifPz9WrFiBTz75BH///TdatmyZ7FgTJ06Eg4MDhg4diri4OJQvXx7169fHqlWr8M033yTbdtWqVXBzc0NgYGCWX5sQuQYJISxO37596c1/34CAAAJACxYsSLF9dHR0ivt69epFLi4uFBsbq7uvS5cu5Ovrq7t969YtAkD58+enZ8+e6e7fvHkzAaCtW7fq7hs3blyKmACQg4MDXb9+XXff+fPnCQDNmTNHd1+LFi3IxcWF7t+/r7svNDSU7OzsUuzTkC5duhCANC99+/ZN8bqWLVtGRETPnz8nADRt2rQ0j1OpUiUKCAhIcf+gQYMIAB06dEh338uXL6lEiRLk5+dHarWaiIj27dtHAKhkyZIp3pOFCxcSALpy5Yruvvj4eCpQoAB16dIl3d+BEIJIup+EsCKOjo748ssvU9zv7Oysu/7y5Us8efIE77zzDqKjo3H16tV099u2bVvkzZtXd/udd94BwF0u6WncuDFKlSqlu121alW4u7vrnqtWq/Hff//h008/hY+Pj2670qVLo1mzZunuX8vJyQm7d+82eEmPs7MzHBwcsH//fjx//jzDx9T6999/4e/vjwYNGujuc3V1Rc+ePXH79m0EBwcn275Lly7J3hMAaNOmDZycnLBq1SrdfTt37sSTJ0/SrZ0SQjDpfhLCihQpUgQODg4p7r98+TJGjx6NvXv3IioqKtljkZGR6e63ePHiyW5rE5yMJABvPlf7fO1zHz16hJiYGJQuXTrFdobuS42trS0aN26c4e2TcnR0xNSpUzFkyBB4eXnh7bffxscff4zOnTujcOHC6T4/LCwMderUSXG/djRaWFgYKleurLu/RIkSKbb19PREixYtsHr1akycOBEAdz0VKVIEjRo1ytLrEiK3kZYaIazIm9/+AeDFixcICAjA+fPnMWHCBGzduhW7d+/G1KlTASBDQ7htbW0N3k9EJn1uTho0aBBCQkIwefJkODk5YcyYMahQoQLOnTtn9GMZep8AoHPnzrh58yaOHj2Kly9fYsuWLWjXrp2MjBIig6SlRggrt3//fjx9+hQbNmxAw4YNdfffunVLwaj0ChUqBCcnJ1y/fj3FY4buM6VSpUphyJAhGDJkCEJDQ1G9enX8/PPP+OOPPwAg1ZFYvr6+uHbtWor7tV17vr6+GTp+06ZNUbBgQaxatQp16tRBdHQ0OnXqlMVXI0TuI+m/EFZO21KStGUkPj4e8+fPVyqkZLTdRps2bcKDBw9091+/fh3bt2/PkRiio6MRGxub7L5SpUrBzc0NcXFxuvvy5MmDFy9epHh+8+bNcfLkSRw7dkx33+vXr7Fo0SL4+fmhYsWKGYrDzs4O7dq1w7p167B8+XJUqVIFVatWzdqLEiIXkpYaIaxcvXr1kDdvXnTp0gUDBgyASqXCypUrzar7Z/z48di1axfq16+PPn36QK1WY+7cuahcuTKCgoJMfvyQkBC8//77aNOmDSpWrAg7Ozts3LgRERER+OKLL3Tb1axZE7/++it++OEHlC5dGoUKFUKjRo0wcuRIrFmzBs2aNcOAAQOQL18+rFixArdu3cL69esz1X3UuXNn/PLLL9i3b5+ui1AIkTGS1Ahh5fLnz49//vkHQ4YMwejRo5E3b1507NgR77//Ppo0aaJ0eAA4Wdi+fTuGDh2KMWPGoFixYpgwYQKuXLmSodFZ2VWsWDG0a9cOe/bswcqVK2FnZ4fy5ctj3bp1+Oyzz3TbjR07FmFhYfjpp5/w8uVLBAQEoFGjRvDy8sLRo0cxYsQIzJkzB7GxsahatSq2bt2Kjz76KFOx1KxZE5UqVcKVK1fQoUMHY79UIayaiszp65oQQiTx6aef4vLlywgNDVU6lBxVo0YN5MuXD3v27FE6FCEsitTUCCHMQkxMTLLboaGh+Pfff5MtRZAbnD59GkFBQejcubPSoQhhcaSlRghhFry9vdG1a1eULFkSYWFh+PXXXxEXF4dz586hTJkySodncpcuXcKZM2fw888/48mTJ7h586bZLLophKWQmhohhFlo2rQp1qxZg/DwcDg6OqJu3br48ccfc0VCAwB///03JkyYgHLlymHNmjWS0AiRBdJSI4QQQgirIDU1QgghhLAKktQIIYQQwipIUiOEEEIIqyBJjRBCCCGsgiQ1QgghhLAKktQIIYQQwipIUiOEEEIIqyBJjRBCCCGsgiQ1QgghhLAKktQIIYQQwipIUiOEEEIIqyBJjRBCCCGsgiQ1QgghhLAKktQIIYQQwipIUiOEEEIIqyBJjRBCCCGsgiQ1QgghhLAKktQIIYQQwipIUiOEEEIIqyBJjRBCCCGsgiQ1QgghhLAKktQIIYQQwipIUiOEEEIIqyBJjRBCCCGsgiQ1QgghhLAKktQIIYQQwipIUiOEEEIIqyBJjRBCCCGsgp3SAeQUjUaDBw8ewM3NDSqVSulwhBBCCJEBRISXL1/Cx8cHNjZpt8XkmqTmwYMHKFasmNJhCCGEECIL7t69i6JFi6a5Ta5Jatzc3ADwL8Xd3V3haIQQQgiREVFRUShWrJjuPJ6WXJPUaLuc3N3dJakRQgghLExGSkekUFgIIYQQVkGSGiGEEEJYBUlqhBBCCGEVck1NjciAuXOBW7eAn34CbG2VjkYIYWaICImJiVCr1UqHIqyMra0t7Ozssj3liiQ1gq1ZA/Tvz9cbNgQCA5WNRwhhVuLj4/Hw4UNER0crHYqwUi4uLvD29oaDg0OW9yFJjQAuXwa6d9ffXrJEkhohhI5Go8GtW7dga2sLHx8fODg4yCSmwmiICPHx8Xj8+DFu3bqFMmXKpDvJXmokqcntoqKAVq2A6GjgrbeAs2eBf/8FHj4EvL2Vjk4IYQbi4+Oh0WhQrFgxuLi4KB2OsELOzs6wt7dHWFgY4uPj4eTklKX9SKFwbkYEfPklEBICFC0K7NgB1KsHqNXA778rHZ0Qwsxk9duzEBlhjL8v+QvNzWbMADZsAOztgb//BgoWBL76ih9bupSTHiGEEMJCSFKTWx08CIwYwddnzQLq1OHrbdoArq7cenP4sGLhCSGEOfLz88OsWbMyvP3+/fuhUqnw4sULk8Uk9CSpyY0ePuTkRa0GOnYE+vTRP+bqCrRty9eXLFEmPiGEyCaVSpXmZfz48Vna76lTp9CzZ88Mb1+vXj08fPgQHh4eWTpeRknyxCSpyW0SEjihiYgAKlcGFiwA3hzFoO2C+usvLiQWQggL8/DhQ91l1qxZcHd3T3bf0KFDddtq59/JiIIFC2aqWNrBwQGFCxeW0WI5RJKa3GbECO5Wcnfnepo8eVJu8/bbQIUKPCJq7dqcj1EIIbKpcOHCuouHhwdUKpXu9tWrV+Hm5obt27ejZs2acHR0xOHDh3Hjxg0EBgbCy8sLrq6uqF27Nv77779k+32z+0mlUuG3335Dy5Yt4eLigjJlymDLli26x99sQVm+fDk8PT2xc+dOVKhQAa6urmjatCkePnyoe05iYiIGDBgAT09P5M+fHyNGjECXLl3w6aefZvn38fz5c3Tu3Bl58+aFi4sLmjVrhtDQUN3jYWFhaNGiBfLmzYs8efKgUqVK+Pfff3XP7dChAwoWLAhnZ2eUKVMGy5Yty3IspiRJTW7y11/AzJl8fcUKoEwZw9upVPrWGumCEkIYQAS8fp3zF2OOXxg5ciSmTJmCK1euoGrVqnj16hWaN2+OPXv24Ny5c2jatClatGiBO3fupLmf77//Hm3atMGFCxfQvHlzdOjQAc+ePUt1++joaEyfPh0rV67EwYMHcefOnWQtR1OnTsWqVauwbNkyHDlyBFFRUdi0aVO2XmvXrl1x+vRpbNmyBceOHQMRoXnz5khISAAA9O3bF3FxcTh48CAuXryIqVOnwtXVFQAwZswYBAcHY/v27bhy5Qp+/fVXFChQIFvxmAzlEpGRkQSAIiMjlQ5FGcHBRK6uRADR8OHpbx8RQWRnx9tfvGj6+IQQZismJoaCg4MpJiZGd9+rV/zxkNOXV68yH/+yZcvIw8NDd3vfvn0EgDZt2pTucytVqkRz5szR3fb19aWZM2fqbgOg0aNHJ/m9vCIAtH379mTHev78uS4WAHT9+nXdc+bNm0deXl66215eXjRt2jTd7cTERCpevDgFBgamGuebx0kqJCSEANCRI0d09z158oScnZ1p3bp1RERUpUoVGj9+vMF9t2jRgr788stUj20shv7OiDJ3/paWmtzg1Svgs8/457vvApMmpf+cQoWAFi34urTWCCGsUK1atZLdfvXqFYYOHYoKFSrA09MTrq6uuHLlSrotNVWrVtVdz5MnD9zd3fHo0aNUt3dxcUGpUqV0t729vXXbR0ZGIiIiAv7+/rrHbW1tUbNmzUy9tqSuXLkCOzs71NGOcgWQP39+lCtXDleuXAEADBgwAD/88APq16+PcePG4cKFC7pt+/Tpg7Vr16J69eoYPnw4jh49muVYTE2SGmtHxEsgXLkC+PhwjYxdBieS1nZBrVwJxMWZLkYhhMVxceHvSTl9MeaExnneqCkcOnQoNm7ciB9//BGHDh1CUFAQqlSpgvj4+DT3Y29vn+y2SqWCRqPJ1Pak8Lxg3bt3x82bN9GpUydcvHgRtWrVwpw5cwAAzZo1Q1hYGL755hs8ePAA77//frLuMnMiSY21++UX4M8/OZH56y/Ayyvjz23ShBOhp0+BJIVvQgihUvE4g5y+mHIQ0ZEjR9C1a1e0bNkSVapUQeHChXH79m3THdAADw8PeHl54dSpU7r71Go1zp49m+V9VqhQAYmJiThx4oTuvqdPn+LatWuoWLGi7r5ixYqhd+/e2LBhA4YMGYLFixfrHitYsCC6dOmCP/74A7NmzcKiRYuyHI8pydpP1uzIEUCbTf/8My+BkBl2dkDXrsCPP3IXVOvWRg9RCCHMRZkyZbBhwwa0aNECKpUKY8aMSbPFxVT69++PyZMno3Tp0ihfvjzmzJmD58+fZ2hY+MWLF+Hm5qa7rVKpUK1aNQQGBqJHjx5YuHAh3NzcMHLkSBQpUgSB/1+8eNCgQWjWrBnKli2L58+fY9++fahQoQIAYOzYsahZsyYqVaqEuLg4/PPPP7rHzI0kNdYqIoKTkMREnkyvf/+s7adbN05qdu0C7twBihc3bpxCCGEmZsyYgW7duqFevXooUKAARowYgSgF5uoaMWIEwsPD0blzZ9ja2qJnz55o0qQJbG1t031uw4YNk922tbVFYmIili1bhoEDB+Ljjz9GfHw8GjZsiH///VfXFaZWq9G3b1/cu3cP7u7uaNq0KWb+f7Ssg4MDRo0ahdu3b8PZ2RnvvPMO1prpdB8qUrojL4dERUXBw8MDkZGRcHd3Vzoc00pMBD74ANi/H6hYEThxgmcKzqr33uN9ff89MHassaIUQliI2NhY3Lp1CyVKlMjy6ski6zQaDSpUqIA2bdpg4sSJSodjMqn9nWXm/C01Ndbo2285CXF1Bdavz15CA+gLhpctAxRoihVCiNwkLCwMixcvRkhICC5evIg+ffrg1q1baN++vdKhmT1JaqzNhg3AtGl8fdkyoHz57O/zs88ADw/g9m1g797s708IIUSqbGxssHz5ctSuXRv169fHxYsX8d9//5ltHYs5kZoaaxISwoW9ADB4MPD558bZr7Mz0L498OuvXDDcuLFx9iuEECKFYsWK4ciRI0qHYZGkpcZavH7NLSovXwLvvANMmWLc/Wu7oDZuBNKY/lsIIYRQiiQ11oAI6NkTuHQJKFyY56V5Y3KnbHvrLaBaNZ6Eb/Vq4+5bCCGEMAJJaqzB/PmcaNjaAuvWAd7exj+GLHIphBDCzElSY+mOHwe++Yav//QTdz2ZSocOgKMjEBQEZGN2SyGEEMIUJKmxZI8fczFwQgL/1CY3ppIvH9CyJV+X1hohhBBmRpIaS6VWA+3aAffvA+XKcZJhykVRtLRdUKtWATExpj+eEEIIkUEmS2rmzZsHPz8/ODk5oU6dOjh58mSq216+fBmfffYZ/Pz8oFKpMGvWrBTb/Prrr6hatSrc3d3h7u6OunXrYvv27aYK3/yNGQPs2cNL1q5fD+TULMmNGgF+fkBkJM+JI4QQVuzdd9/FoEGDdLf9/PwMnqOSUqlU2LRpU7aPbaz95CYmSWr+/PNPDB48GOPGjcPZs2dRrVo1NGnSBI8ePTK4fXR0NEqWLIkpU6agcOHCBrcpWrQopkyZgjNnzuD06dNo1KgRAgMDcfnyZVO8BPO2ZQsweTJfX7IEqFQp545tYwN8+aX+2EIIYYZatGiBpk2bGnzs0KFDUKlUuHDhQqb3e+rUKfTs2TO74SUzfvx4VK9ePcX9Dx8+RLNmzYx6rDctX74cnp6eJj1GTjJJUjNjxgz06NEDX375JSpWrIgFCxbAxcUFS5cuNbh97dq1MW3aNHzxxRdwdHQ0uE2LFi3QvHlzlClTBmXLlsWkSZPg6uqK48ePm+IlmK/r14HOnfn6gAHAF1/kfAxdu3JX1759wI0bOX98IYRIx1dffYXdu3fj3r17KR5btmwZatWqhapVq2Z6vwULFoSLi4sxQkxX4cKFUz0nCsOMntTEx8fjzJkzaJxk1lkbGxs0btwYx44dM8ox1Go11q5di9evX6Nu3bpG2adFiI7mCfYiI4G6dfXLIeS04sWBDz/k68uWKRODEEKk4eOPP0bBggWxfPnyZPe/evUKf/31F7766is8ffoU7dq1Q5EiReDi4oIqVapgzZo1ae73ze6n0NBQNGzYEE5OTqhYsSJ2796d4jkjRoxA2bJl4eLigpIlS2LMmDFISEgAwC0l33//Pc6fPw+VSgWVSqWL+c3up4sXL6JRo0ZwdnZG/vz50bNnT7x69Ur3eNeuXfHpp59i+vTp8Pb2Rv78+dG3b1/dsbLizp07CAwMhKurK9zd3dGmTRtEREToHj9//jzee+89uLm5wd3dHTVr1sTp06cB8BpWLVq0QN68eZEnTx5UqlQJ//77b5ZjyQijL5Pw5MkTqNVqeHl5Jbvfy8sLV69ezda+L168iLp16yI2Nhaurq7YuHEjKlasaHDbuLg4xMXF6W4rsXy8UREBX38NXLgAFCoE/PUX4OCgXDxffQXs3AksX86rd9vaKheLECLnEfEXrZzm4pKhQRF2dnbo3Lkzli9fju+++w6q/z/nr7/+glqtRrt27fDq1SvUrFkTI0aMgLu7O7Zt24ZOnTqhVKlS8Pf3T/cYGo0GrVq1gpeXF06cOIHIyMhk9Tdabm5uWL58OXx8fHDx4kX06NEDbm5uGD58ONq2bYtLly5hx44d+O+//wAAHh4eKfbx+vVrNGnSBHXr1sWpU6fw6NEjdO/eHf369UuWuO3btw/e3t7Yt28frl+/jrZt26J69ero0aNHuq/H0OvTJjQHDhxAYmIi+vbti7Zt22L//v0AgA4dOqBGjRr49ddfYWtri6CgINj/f/LXvn37Ij4+HgcPHkSePHkQHBwM1+wusJweMrL79+8TADp69Giy+4cNG0b+/v7pPt/X15dmzpxp8LG4uDgKDQ2l06dP08iRI6lAgQJ0+fJlg9uOGzeOAKS4REZGZvo1mYUFC4gAIhsbor17lY6GKDaWKH9+jmnbNqWjEUKYUExMDAUHB1NMTIz+zlev+P8/py+vXmU47itXrhAA2rdvn+6+d955hzp27Jjqcz766CMaMmSI7nZAQAANHDhQdzvpOWrnzp1kZ2dH9+/f1z2+fft2AkAbN25M9RjTpk2jmjVr6m6PGzeOqlWrlmK7pPtZtGgR5c2bl14lef3btm0jGxsbCg8PJyKiLl26kK+vLyUmJuq2ad26NbVt2zbVWJYtW0YeHh4GH9u1axfZ2trSnTt3dPddvnyZANDJkyeJiMjNzY2WL19u8PlVqlSh8ePHp3rsNxn8OyOiyMjIDJ+/jd79VKBAAdja2iZrngKAiIiIVIuAM8rBwQGlS5dGzZo1MXnyZFSrVg2zZ882uO2oUaMQGRmpu9y9ezdbx1bUqVNcPwMAP/4IvPeesvEAPAlfx458XQqGhRBmqHz58qhXr56unvP69es4dOgQvvr/1BRqtRoTJ05ElSpVkC9fPri6umLnzp24c+dOhvZ/5coVFCtWDD4+Prr7DJVE/Pnnn6hfvz4KFy4MV1dXjB49OsPHSHqsatWqIU+ePLr76tevD41Gg2vXrunuq1SpEmyTtJx7e3unOkgnI8csVqwYihUrpruvYsWK8PT0xJUrVwAAgwcPRvfu3dG4cWNMmTIFN5LUWQ4YMAA//PAD6tevj3HjxmWpMDuzjJ7UODg4oGbNmtizZ4/uPo1Ggz179hi9/kWj0STrYkrK0dFRN/xbe7FIT57wxHrx8cCnnwLDhysdkZ52zpotW4As/tMIISyUiwvw6lXOXzJZpPvVV19h/fr1ePnyJZYtW4ZSpUohICAAADBt2jTMnj0bI0aMwL59+xAUFIQmTZogPj7eaL+mY8eOoUOHDmjevDn++ecfnDt3Dt99951Rj5GU/Rvr/qlUKmg0GpMcC+CRW5cvX8ZHH32EvXv3omLFiti4cSMAoHv37rh58yY6deqEixcvolatWpgzZ47JYgFMNPpp8ODBWLx4MVasWIErV66gT58+eP36Nb78/1Dgzp07Y9SoUbrt4+PjERQUhKCgIMTHx+P+/fsICgrC9evXdduMGjUKBw8exO3bt3Hx4kWMGjUK+/fvR4cOHUzxEsyDWs1LE9y5A5QuzfUrOTHBXkZVqQLUrg0kJgIrVyodjRAiJ6lUQJ48OX/J5GdgmzZtYGNjg9WrV+P3339Ht27ddPU1R44cQWBgIDp27Ihq1aqhZMmSCAkJyfC+K1SogLt37+Lhw4e6+94ckXv06FH4+vriu+++Q61atVCmTBmEhYUl28bBwQFqtTrdY50/fx6vX7/W3XfkyBHY2NigXLlyGY45M7SvL2lPR3BwMF68eJGsnrVs2bL45ptvsGvXLrRq1QrLkgwgKVasGHr37o0NGzZgyJAhWLx4sUli1TJ6oTAAtG3bFo8fP8bYsWMRHh6O6tWrY8eOHbri4Tt37sDGRp9PPXjwADVq1NDdnj59OqZPn46AgABdMdKjR4/QuXNnPHz4EB4eHqhatSp27tyJDz74wBQvwTxMmADs2gU4O/MEewaKxxT31VfcPbZkCTB4sHklXUKIXM/V1RVt27bFqFGjEBUVha5du+oeK1OmDP7++28cPXoUefPmxYwZMxAREZHqAJQ3NW7cGGXLlkWXLl0wbdo0REVF4bvvvku2TZkyZXDnzh2sXbsWtWvXxrZt23QtGVp+fn64desWgoKCULRoUbi5uaUYyt2hQweMGzcOXbp0wfjx4/H48WP0798fnTp1SjEwJ7PUajWCgoKS3efo6IjGjRujSpUq6NChA2bNmoXExER8/fXXCAgIQK1atRATE4Nhw4bh888/R4kSJXDv3j2cOnUKn332GQBg0KBBaNasGcqWLYvnz59j3759qFChQrZiTVeGK3gsXGYKjczCtm36wriVK5WOJnUvXhA5O3OcbxSHCyGsQ2oFnJbi6NGjBICaN2+e7P6nT59SYGAgubq6UqFChWj06NHUuXNnCgwM1G2TVqEwEdG1a9eoQYMG5ODgQGXLlqUdO3akKBQeNmwY5c+fn1xdXalt27Y0c+bMZMW5sbGx9Nlnn5GnpycBoGXLlhERpdjPhQsX6L333iMnJyfKly8f9ejRg16+fKl7vEuXLsliJyIaOHAgBQQEpPq7WbZsmcFBNaVKlSIiorCwMPrkk08oT5485ObmRq1bt9YVJsfFxdEXX3xBxYoVIwcHB/Lx8aF+/frp/k769etHpUqVIkdHRypYsCB16tSJnjx5kmosxigUVv3/F2f1oqKi4OHhgcjISPOvr7l1C6hZE3j+nIdxz5undERp69IF+P13brX57TeloxFCGFlsbCxu3bqFEiVKwMnJSelwhJVK7e8sM+dvWdDS3MTG8gR7z58D/v7AjBlKR5Q+bcHwn39yIZ8QQgihAElqzE2/fsC5c0D+/MDff/PQaXP3zjtAmTKc0Kxbp3Q0QgghcilJaszJkiV8UamANWuAJHMDmDWVCujWja+nsr6XEEIIYWqS1JiLs2eBvn35+sSJgKWN6urShZdKOHIEyOZyGEIIIURWSFJjDp494zqauDjg44+BJHP4WAxvb6B5c74urTVCmLf4eB5bKYSVkaRGaRoN0KkTcPs2ULIkjyKysdC3RVswvGIFkI1VYYUQJrR1K+DqyvNgZVIuGSwrFGKMvy8LPXtakUmTgH//BZyceIK9vHmVjijrmjcHvLx4yYRt25SORgjxpqgooHdv/tIxdSrw+HGGnqadej9aiVW5Ra6h/ft6c6mHzDDJjMIig3buBMaN4+u//gpUr65oONlmb8+1NT/9xAXPn36qdERCiKRGjwYePODrMTHA7NnADz+k+zRbW1t4enrqFkZ0cXHRLTUgRHYREaKjo/Ho0SN4enomW5Azs2TyPaWEhQFvvcX1ND16AIsWKR2RcVy7BpQvz11od+8CSVavFUIo6PRpnvuKiKeOmDuXl14JC8vQEixEhPDwcLx48cL0sYpcydPTE4ULF06RMGfm/C0tNUqIi+OVt58945mDf/lF6YiMp1w5oEED4PBhrq2xxKJnIaxNYiLQqxcnNO3bcwvN3r1AcDAwf36G/k9VKhW8vb1RqFAhJEjNnDAye3v7bLXQaElLjRJ69wYWLgTy5QPOnAH8/JSNx9iWLwe+/JJXFg8JkUUuhVDa7NnAoEGApydPueDlBfzxBw9SKFiQByq4uCgcpBCGyTIJ5mzFCk5oVCpg1SrrS2gAoHVrwM0NuH4dOHhQ6WiEyN3u3eNaGgCYMoUTGgD44gv+/Hn8mGvghLACktTkpPPnuZUG4ALhpk2VjcdU8uThD0xAPiyFUNrAgbyESd26XL+nZWcHjBjB13/6ieeuEcLCSVKTU1684An2YmOBZs2AMWOUjsi0tMsm/P03EBmpbCxC5Fb//ANs2MCzfS9YkHIOrK5deeLMe/e4O0oICydJTU7QaIDOnYEbNwBfX2DlSsudYC+j6tQBKlbkYaNr1igdjRC5z+vX+qVXBg8GqlZNuY2TEzBkCF+fMgVQq3MuPiFMwMrPrGZi6lSexdPBgSfYy59f6YhMT6XSzzAsXVBC5Lzvvwfu3OEvUtr5sAzp1Ysn/QwN5c8nISyYJDWmtmePvkhv7lwewp1bdOrEE/KdPg1cuKB0NELkHhcuADNm8PW5c7nOLTWurlx3AwA//ihrQgmLJkmNKd27xwWzGg0Pce7eXemIclbBgsAnn/B1aa0RImdoNNz6olYDrVrxIrnp6d+fk5vz53nZFiEslCQ1phIfz0Obnzzh5Q/mzcud87Vou6D++IMnHRRCmNbixcDx45ykzJ6dsefkywf06cPXJ02S1hphsSSpMZXBg/mDxdOT+6mdnZWOSBkffggULcqzJ2/apHQ0Qli3iAhg5Ei+/sMP/L+XUd98Azg6AseOyfxSwmJJUmMKq1ZxywzAI51KllQ2HiXZ2vKwUQBYulTRUISweoMH8/QRb73F6ztlhre3fiqGSZOMHpoQOUGSGmO7dAno2ZOvjx6dsf5sa/fll/xz925ePE8IYXy7dwOrV/N0EQsX8heKzBo+nJ+3ezdw6pTxYxTCxCSpMabISC7Mi44GPvgAGD9e6YjMQ8mSQKNG3E+/fLnS0QhhfWJjga+/5ut9+wK1amVtP35+QIcOfH3yZKOEJkROkqTGWIi4RSI0FChWjL8xGWHFUauhLRhetoxHZwghjOfHH3mtNR8frqXJjpEjeVDDxo28ircQFkSSGmOZPp0/BOzteWmAAgWUjsi8tGzJRdNhYTx3j8g19u8HPv+c60+FCVy9yrMBAzzaKZ1VjNNVoQL/vwLSWiMsjiQ1xrB/v37EwezZgL+/ouGYJWdnfbO2zFmTa1y9ylMVrV8PvPuuvPVGR8SL5CYkAM2b8/pyxvDtt/xzzRrg5k3j7FOIHCBJTXY9eAC0bctdKp066VfhFilpu6A2bgSePlU2FmFyUVH8hf/lS26ki4/n+Sf79uVzsDCC338HDhzgLw3GnAurZk2gSROewG/aNOPsU4gcIElNdhFxIWyVKrwKbm6cYC+jatTgS3w8D3sXVku7huvVqzxVypUrwMSJ/Nj8+UDjxsCjR8rGaPGePgWGDuXr48Zxka8xaVtrli7lL29CWABJarKrSBH+prRjB+DionQ05i/pIpcya6nVmjQJ2LyZ53LbsAEoXJhnONiyBXBz47ndatUCzpxROlILNnw4z1heuTLPT2NsDRsCDRrwlxDtOlJCmDlJaozBwYFHHYj0tW/PZ7oLF+SMZqX++Ue/KPSvvwK1a+sfa9ECOHECKFMGuHuXz5nSaJcFhw7pJ7NcuJAHKJiCtrVmwQLpMhYWQZIakbPy5tUXM0rVqNUJCeF6cCKeNkU772JSFSoAJ09yXWtsLNCxI/eiJCbmfLwWKT6eF6wEgB49gHr1THespk157brXr4E5c0x3HCGMRJIakfO0XVCrV/NEhcIqREUBn37KPxs0AGbOTH1bT0/uiho1im///DMnOc+e5USkFm76dC5SKlhQP5TbVFQqfWvNL79w1bcQZkySGpHz3n0XKFGCz37r1ysdjTACjQbo0oXPtT4+wF9/ca9sWmxtec64P//kcrTdu7mr6tKlnInZIt24oa+4njGDV9c2tVatgHLlgOfPuRtKCDMmSY3IeTY2+n4J6YKyCpMn8yLsDg76wuCMatMGOHqUB+/cvAm8/TbvQ7yBiMfDx8YC77+vn/fJ1Gxt9fNw/fwzH18IMyVJjVBG167ctH3gAE/vLizWtm3AmDF8ff58oE6dzO+jWjXg9GleIuz1ay67GjtWVtRIZt06YOdOzhznz8/Z6SM6dACKFwciInipEyHMlMmSmnnz5sHPzw9OTk6oU6cOTp48meq2ly9fxmeffQY/Pz+oVCrMmjUrxTaTJ09G7dq14ebmhkKFCuHTTz/FtWvXTBW+MLVixXhyL0A/ikNYnNBQfWFw7976cqmsyJ+fz9mDBvHtiRP1NTq53osX+l/Mt98CZcvm7PHt7YFhw/j61Kkye6IwWyZJav78808MHjwY48aNw9mzZ1GtWjU0adIEj1KZbSs6OholS5bElClTUDiVdusDBw6gb9++OH78OHbv3o2EhAR8+OGHeP36tSlegsgJ2jPg8uUy9MUCvXzJSUdkJA/AmT07+/u0s+MC4xUreOT/1q3c8hMSkv19W7TvvgPCwzmZ0XYF5bSvvgIKFeL129asUSYGIdJDJuDv7099+/bV3Var1eTj40OTJ09O97m+vr40c+bMdLd79OgRAaADBw5kKKbIyEgCQJGRkRnaPjNmzCD65x+i+Hij79q6xcURFShABBBt3ap0NCITNBqizz7jt87bm+jBA+Mf4+RJoiJF+BgeHkTbthn/GBbhxAkilYp/EXv3KhvLlCkcR/nyRGq1srGIXCMz52+jt9TEx8fjzJkzaNy4se4+GxsbNG7cGMeMuExvZGQkACBfTlT/p+H5c/7i9PHHPOqjb1/gyBGZLDdDHBx4vSxAuqAszJQpPHDN3p5/ensb/xi1a3OdTf363Br08cd83Fz1v5WYyHPSEPH/ynvvKRtPnz6Ahwevf7Fpk7KxCGGA0ZOaJ0+eQK1Ww8vLK9n9Xl5eCA8PN8oxNBoNBg0ahPr166Ny5coGt4mLi0NUVFSyiynExXEtQaFCPGP5/Pk8R0fJktz1ffmySQ5rPbRdUFu3chGiMHvbt3NvCMBrKNata7pjFS4M7N0L9OzJ5/VRo4AvvuBi4lzhl1+AoCCetHL6dKWjAdzdgf79+fqkSbkswxSWwCJHP/Xt2xeXLl3C2rVrU91m8uTJ8PDw0F2KFStmklgKF+Zagvv3efmnzp0BV1fg9m0e5lq5Mk/I+dNPPC28eEOlSlw0kZgIrFypdDQiHdev80oXRJxo9Ohh+mM6OPBKAAsWcM3NunXcenP7tumPrag7d3gIGMAfIIUKKRuP1sCBPLHQ2bPArl1KRyNEMkZPagoUKABbW1tEvPGtOyIiItUi4Mzo168f/vnnH+zbtw9FixZNdbtRo0YhMjJSd7lr4ozCzo4H86xYwQ0Of/4JfPIJN8+fPw+MGMEjIgMCgEWLZObUZGSRS4vw6hXQsiUPxKlblxsRclKvXsC+fXxuP3+eF8TcuzdnY8hRAwZwk1T9+kC3bkpHo1eggH6Zhh9/VDYWId5g9KTGwcEBNWvWxJ49e3T3aTQa7NmzB3Wz0U5NROjXrx82btyIvXv3okSJEmlu7+joCHd392SXnOLiwhOKbd7MAxYWLuQFbwFenbhXL27hCQzkb525fqWAtm35l3b1KmDEuithPEQ8X+KlS/y3+/ffPDopQ37/HahZ0yg1GA0acJ1NzZq8vuKHH3JLqdXlwps388XOjj9AbMysUX3IEG5CO3gQOHxY6WiE0DNFpfLatWvJ0dGRli9fTsHBwdSzZ0/y9PSk8PBwIiLq1KkTjRw5Urd9XFwcnTt3js6dO0fe3t40dOhQOnfuHIWGhuq26dOnD3l4eND+/fvp4cOHukt0dHSGYjLl6KeMunOHaOpUomrVeACB9uLqStS5M9GOHUQJCYqFp6yuXfmX0a2b0pEIA7SDXuztiY4cycQTN2wgsrHhJ6tURBkY2ZgR0dFEnTrp/4e6dCGKiTHKrpX38iVRsWL8wpJ8Tpqdnj05xmbNlI5EWLnMnL9NktQQEc2ZM4eKFy9ODg4O5O/vT8ePH9c9FhAQQF26dNHdvnXrFgFIcQkICNAHauBxALRs2bIMxWMOSU1Sly4RjRpF5OubPMHx8iIaMIDo+HEeNptrHDrEv4A8eYiiopSORiSxY4d+RPGCBZl44sGDRI6O/MQKFfR/5AMGECUmZjsujYanU9DmTP7+RPfuZXu3yhs8mF9QiRJEr18rHU3qrl/X//LPnlU6GmHFzCKpMTfmltRoaTREhw8Tff01Uf78yROcUqWIxowhunpV6ShzgEZDVK4cv/DfflM6GvF/168T5c3Lb0v37plItC9eJPL05CcGBnIT5E8/6f+4P/mE6NUro8S4ezdRvnz6LwWZakkyN+fOEdna8ov591+lo0lf+/Yca+vWSkcirJgkNQaYa1KTVHw8TzDWvj2Ri0vyBKdmTaKffya6f1/pKE1o6lR+sXXrKh2JIM45qlTht6ROHaLY2Aw+8c4d/ax59etzX5HWunX61ptatYgePjRKrDdu6GO1tydatMgou81ZiYnc3GRJScLFi/quxStXlI5GWClJagywhKQmqVeviFatImreXP/FTfvZ0agR0ZIlRM+fKx2lkT18qH+xwcFKR5OraTREbdroWz8y3K3z9Km+q6liRb79piNH9M2Svr5Ely8bJeaXL4k+/1z/v9KnD09abTHmzePA3d0t69tLYCDH3bWr0pEIKyVJjQGWltQk9egRf97Vr5+89cbRkahVK6L1662oSFL7ATlkiNKR5GraniI7Oy53ypDoaP0faZEi3GKTmtBQotKl9WsgGGn6f42G6Icf9DVA77xD9P/xCebtwQNOZgCiOXOUjiZzjh/X/7Hcvq10NMIKSVJjgCUnNUndvEk0aRJ/CU6a4Hh48MCh//4zSg2mcrZs4RdUsKCFfc22Hrt26es/58/P4JMSErhOBuBamkuX0n/O48f6JMjenmjFimzFndTWrfocoWhRotOnjbZr02jbVt8lZ4n/wO+/z/H366d0JMIKSVJjgLUkNVoaDVFQENHw4fyhnTTB8fHhARSnT1vgCKqEBF4hEeAmKJGjbt7UF91265bBvx+NRj+819GRRz1lVEyMvp8LIPr+e6P90V65oq89d3IiWrnSKLs1vh07OEgbG6IzZ5SOJmv27NH/oi2iaUxYEklqDLC2pCYptZrowAE+r2hHqmgv5crxeSLJlD/mb+RIDr55c6UjyVVevSKqWlU/PDrDXZrjx+tPyhs2ZP7AajXRiBHJJ50xUivdixdEH32k3/XgwWY2F1R0NFHJkhzcwIFKR5N1Gg3R22/z6xgxQulohJWRpMYAa05qkoqNJdq8mb/8OjklT3D8/Ylmz7aAL1IhIfqTpFVMPGL+NBqidu34116oENHduxl84oIF+j+wX3/NXhALFugLxRs1MlolfGIi0Xff6cNs3JjoyROj7Dr7vv1WX4Nk6fMzabuO3dyInj1TOhphRSSpMSC3JDVJRUVxmUKTJvoaCW2u8OGH/JjZ/joaNuRgf/hB6UhyhenT9bWeBw5k8EkbN+r/sMaONU4g//7LU2xrR08ZsfD0r7/0UyWUKEF04YLRdp01ly9zLZG1dLWq1fpx9RMnKh2NsCKS1BiQG5OapMLDuZWmTp3krTdOTtyqs2mTmdXlLl/OAZYsyR+WwmR279bnJhkeeHPokL4pMFOz8mXAuXNcGAYQFS5MdOqU0XZ9/jwnNABPXv3330bbdeao1Tw0CyD6+GMLLH5LxerV/Jry5+cx9kIYgSQ1BuT2pCap0FCiCRP0RZTaS968RD16EO3fbwZ5xKtX3IwNGG24r0jp1i39lDFdu2bw3Hrpkn624BYtTFOkcveuvsDHxYW7NozkyRP9YB2Au6Zy/O99yRL9a7OmYdCJifqh+jNmKB2NsBKS1BggSU1KGg0Pthg8WD/gSHspUoRo6FBe0kWxL5HaETUdOigUgHV7/ZqoenX9SOIMFQbfvasfble3rmnXJoqM5H5SbZ/p3LlG23VCgn6JJW1jyYsXRtt92h490g8x++mnHDpoDlq8WD8MM8PTUAuROklqDJCkJm2JiTzHzVdf8Zw3SROc8uV5BFVISA4HdeKEvo/M6qZPVpZGw7midkqgtObJ03n2jKhSJX5ShQqGZws2tvh47t5KOnzJiM0qv/+uX7WhXLkcWmetSxc+YJUq/PqsTWysfpmMhQuVjkZYAUlqDJCkJuNiY7kGtHXrlCOoatcmmjmTJ0A1OY2GqHLlTM4CJzJixgz+tdracndjuqKjiRo00H8DDwszeYw6Gg3Rjz/q/whbtUq+nlQ2nTqlb3xydyf65x+j7Tqlffv4QCoV0dGjJjyQwmbN0tfEmdUYemGJJKkxQJKarImM1I+gSroGlY0N1yWYfA2qmTP5gDVrmvAgucuePfr3cvbsDDwhMZHo00/1U1crNWxo9WoiBweO4+23iSIijLbr8HB9zqZS8azdRu92jY3VF7L16mXknZuZV6+IChTg17pqldLRCAsnSY0BktRkX3g4j46pWzd5642DA1HLljxk1ohfoNnjx/phr0FBRt557nP7tv5c07lzBk7cGg2fgLWzBWd4vLeJHDyon2GyZEmj9hfFxRH17q3/u/78cyMP4JkwQT8RUG6Yx+WHH/j1Vq5sBiMPhCWTpMYASWqMK7U1qNzcuGRg504jtjq3bs0779/fSDvMnaKjid56i3+Vb72VwQT0++/1zRfmMpfK1av6WXjz5s3csgwZsHChPo+uUoX/1rMtJERfvJNbWi6eP9cvwLVpk9LRCAsmSY0BktSYhkbDc3+MHElUvHjyBKdQIc5Djh3LZlO+dm2cvHmtaDnynKXREHXsyL/GAgUyWBKzaJH+zTS3mqaICP2kSw4O3DVlRIcPE3l58e7z5eMi+izTaHgaY+10xtYyJ01GjBqln848N71uYVSS1BggSY3pqdV8Mvj6a30Xh/ZSogTPB3L5chZ2nJhIVKwY72jNGqPHnRto6zZtbTM47c/mzfoZ+UaPNnl8WRIdzUXD2j+yH3806onz7l0ujNf+3mbOzOLuV63Sd99Z1CJsRhARoR9tkK3MUFiEpUuJ/vjD6LuVpMYASWpyVnw8z3jfsSPP3Jo0walWjWjq1EwOoBk7Vv9NV2TKvn36wuCZMzPwhCNH9CeiDC/VrRC1mmjIEP0fV/fuRh0mHRPDtUfa3XfunMm6sWfPuMkS4Jqa3Kh/f379jRopHYkwlfh4on799Mm7kedGkKTGAElqlPPqFdHatUSffKKvVdBe3nmH10F8/Didndy6xXUdAF8XGRIWxvPQAJxgppufXL6sL8T9+GPLGY47d66+ZenDD426qJlGwy1d2sSwWjWia9cy+GRtkXW5crl3IrqwMF5UDOC+aGFdIiL0a/UBROPHG70wXJIaAySpMQ9Pn3Kpxrvv6nMUgD/zmjfnlstUR5xo6xKMtXiilYuO5pHwAFGNGhmY/PfuXX0339tvm3a2YFPYulW/YmWVKplYajxj/vtPnyC6umagjOfoUf0f+L59Ro3F4nTrxr+HFi2UjkQY0+nT+s8MNzfutjYBSWoMkKTG/Ny9y6tDa0fkaC/OzkRffMHL/SRbZHPNGt6gWDGusxGp0mj03Sb582egcevZM/1Eh+XK8QJJluj0aV4EUztJ4LlzRt39/ftEAQH6v9WePVPpjoqP169Y3aWLUWOwSNeu6b/FnD+vdDTCGFas0I/oK1uW6MoVkx1KkhoDJKkxb1eucAOMdi087SVvXj5x7N9PpH4do+8a2bFD6ZDN2pw5+kkS9+xJZ+OYGP2K0d7elr/A4u3b+uUcXF25uMuIEhK4dlp7jq5a1UAJwU8/6YdOpdu3mku0acO/k3btlI5EZEdCAtGgQfoP6Y8+MvnCaZLUGCBJjWXQaIhOnuT/Ge0Xbu2lSBGiQzW46FDTurXSoZqtAwf0JQw//5zOxomJ+hFE7u7W8y36+XMuTNUOXTLBGkS7diXvjtJNP3P7tr4bbMkSox/XYp07p8+0c9soMGvx+LH+/0o7MjIHJlaUpMYASWosj3aRzW7d9ItsVkUQEUBxsKdpIx7LZ+Mb7tzRn2jbt0+nMFijIerTRz/Xi7XVfcTFEXXtqv8AHjHC6B/A9+9zfZj2ED26ayix2cf6KnhzHjmmhI8+0o9SE5bl7FkiX199Fp+Dk3FKUmOAJDWWLSaGaMMGnrr+tIqrXwdipm5er1mzcmiRTTMWE6OfV6VatQzU+U6cqJ8t+K+/ciLEnKfR6JcnALgLxMgTOCYmctepSkXUEuu5JdHePouTMlm5I0f4fbC3N3ohtzCh1au52BHgGoFLl3L08JLUGCBJjfWI/nk+EUA3XSuTjUqjO1/Z2PAAKZMvsmmGNBp9o0S+fBmY2n/xYv2Jfs6cHIlRUb//rp9PoH59k9S57N0cRfdtihAB9JP9t6aYg8w6aCutBw5UOhKRnoQEoqFD9Z8VTZsqsm6ZJDUGSFJjRZ4/100O9+TfEwYX2bSz426BKVO4K9/aewHmzdMndrt3p7Pxli36OV2+/TZH4jMLe/fq+zHLlDF+XcfAgUQA3XcqSU6IJoDoq68sb2S8ye3apR/m+OiR0tGI1Dx9SvTBB/oP1ZEjFRt1KkmNAZLUWJkOHfRjav/vxg3Di2wCXHTcpQuPCrfU0cqpOXhQXxg8bVo6Gx89qm9G/vJL68/23nT5sr4uIH9+7g4xhjNndIli4rYdNG6cfnRU5comHe1qeTQaolq1+Jfz3XdKRyMMOX+e17YBuOh93TpFw5GkxgBJaqzM3r36CZ8MfBUODeVJZj/+WD8QRXtRqbgOZ+xYPsdb8pQ3d+/qZ+H/4ot0cpTgYO6b0g7DtJTZgo3t4UP9SdXRMfsf2ImJ+v21bau7+7//9Iti5snDPWDi/zZs4F+Mh4fJhwOLTPrzT/2HZokSZjEiUpIaAySpsTJqNVHJkvyPt2JFmpvGxvIJZtgw/Xxob86F06YN1+Lcu5dD8RtBbCwnZ9q5Ul69SmPje/f0y6jXqZPOxrnAq1dEgYH6P4Kffsp6q9Uvv+iHxL9Rrf7wYfIRsN26SXcUEfH/r7ZJ9ccflY5GEHFyPmKE/o/1gw+4C8oMSFJjgCQ1VuiHH/ifr2HDTD3t3j1OYFq3JvL0TJnkVKnCCdB//5nvcj0aDddraJOyGzfS2Pj5c302V7asTAanlZhINGCA/o3v3TvzrVf37nFrIcCFTakc5vvv9d1RlSrJwCgiIlq5kn8hBQtKpqe0Z8+ImjTR/y8MG2ZWLbmS1BggSY0VundPX/AaEpKlXSQkcBfU2LHc6pF0PSptd/LHH3NX1vXrRo4/G379VV8YvGtXGhvGxOgXmytcWBYDNWTWLP0b37x5GouPGfD556SbVyCdfsy9e/UTSrq4pNvAaP0SEoj8/PgX8ssvSkeTe126RFSqlL54O91FzXKeJDUGSFJjpZo311fmG8Hjx1xM3KWLvh4i6aV0aaK+fXntRKV6cA4f1o9Onjo1jQ0TE4k++0zfNRIUlGMxWpyNG/UF1DVq8Kx66dm2TT9jcQbXmAoP16/LCvAw/FzdE6jNzosVe2OhN5Ej1q/ngi+AC+iNvFaasUhSY4AkNVZqPU92Rt7eRm8uVav5f3zKFB4erh1hpL04OBC9/z6POLpwIWcGEt27p0+22rRJ45gaDWdf2kD37jV9cJbuxAl91XXRovympub1a30rw+DBmTpMYiLPB6htZKxYMRd3R8XE8P8uIEtK5CS1mkeeaT/MGjUy625pSWoMkKTGSsXF6dcF2LLFpIeKiiLatIlXFtCez5JefHy4EHTdOtPMTxUbS/T22/q6nzS/4U+apB/q9eefxg/GWt28SVS+vH5kXWp9e9qCymLFMtddlcS+fcm7o5Yty3LUlm36dP3cQZY8FNFSvHihX64CIPrmG7OqnzHELJKauXPnkq+vLzk6OpK/vz+dOHEi1W0vXbpErVq1Il9fXwJAM2fOTLHNgQMH6OOPPyZvb28CQBs3bsxUPJLUWLEhQ/ifMzAwxw6p0RBdu0Y0ezZRs2b6ngvtxcaGJwT8/ntuADDGZ3WPHrxvT8906nuWLtUHIrUKmffsmX7WWzu7lC0IFy/qm+02bcrWocLDk89v1qVLLuyOevmSq90BScBNLTiYBwsAPIHpypVKR5Qhiic1a9euJQcHB1q6dCldvnyZevToQZ6enhQREWFw+5MnT9LQoUNpzZo1VLhwYYNJzb///kvfffcdbdiwQZIakVxwsL624eFDRUKIieEv9YMHG578L39+onbtiJYvz1qICxfqG162b09jw61b+fdgxDqjXCk2lqhjR/0b+N13nMmq1UT16hk1iVareSCftjuqQoUcX1pHeePH84uvVi33TQiZUzZv1o/UK1aM6PRppSPKMMWTGn9/f+rbt6/utlqtJh8fH5o8eXK6z/X19TWY1CQlSY1IQbtOQpqVsznnzh1eXumzz7hG980kp3p1zjn270+/PvLoUX1hcJr/QseO6ZuMunSRk0N2aTREY8bo37T27XmdLICLK+/cMerh9u/Xl5c4O3ODW655C58+5ZWfAaJ//lE6GuuiVhONG6f/Ow4IIEqlgcFcKZrUxMXFka2tbYqko3PnzvTJJ5+k+3xJakSW/Pabfh4WMzsTxMcTHTrEX/a1E88mvbi68pf+X39NuRDl/fv6E93nn6fx0q5c0c8W3KwZH1QYx9KlKavEf/7ZJIeKiCD68EP9YTp1ynLJjuUZNoxfdN26Zvc/bLEiI4k++UT/B9W/v0V+NmTm/G0DI3vy5AnUajW8vLyS3e/l5YXw8HBjHy5VcXFxiIqKSnYRVqxNGyBPHiAkBDhyROlokrG3Bxo0AH74ATh1CoiIAP74A+jYEShYEHj1Cti8GejTByhZEihXDhg4EPj3X+Dzz4GHD4FKlYBlywCVysABHjwAmjYFnj0DatcG/vqLDyqM48svge3bAXd3vl29OjBggEkOVagQH2rSJMDGBli5kt/SS5dMcjjz8s03gKMjcOwYcPCg0tGY3N27wKJFwKefAnnzAnXrGvllh4QAdeoAW7bw73XZMuCXX6z+s8HoSY25mDx5Mjw8PHSXYsWKKR2SMCU3N6BtW76+ZImysaSjUCGgQwc+YYWHA2fO8EnsnXcAW1v+LPrlF+Cjj/jz3dMT2LQJcHU1sLPISKBZMyAsDChTBti2jZM7YVyNG/ObMXgw8PffgJ2dyQ5lYwN8+y2wbx/g4wNcvQr4+/OfNZHJDqs8b2+gWze+PmmSsrGYQEICcOAAMGIEUKUKULw40KsXf6F58QI4fhwICABatgRCQ7N5sG3bOBu+ehUoUoSzpa5djfAqLICxm4nMpfspNjaWIiMjdZe7d+9K95O1O3JEPz7WQt/nFy94rb+ePXmpJienNAqDY2L0o3QKF07ZdyUs3qNHyWev79jRyrujbt3SF7qfPKl0NNl2/z4PnjNUW2djwzXnEycSHTzIq3Roi8Xt7IgGDszC0ktqNe9QO0N2gwaKDZ4wJrMoFO7Xr5/utlqtpiJFikihsDAtjUY/x8iiRUpHk20aTRrd34mJ+in63dzMdiZQkX1qNReIa8/15cqlPS+gxevcmV9oy5ZKR5JpCQk84/e33/LE1G/WzxUowInp6tVET56kfP6lS1wSp93e05PLtzK0Bl1UFFGrVvon9+ljNbM0K57UrF27lhwdHWn58uUUHBxMPXv2JE9PTwoPDyciok6dOtHIJMNN4+Li6Ny5c3Tu3Dny9vamoUOH0rlz5yg0NFS3zcuXL3XbAKAZM2bQuXPnKCwsLEMxSVKTS0ybxv/QdepYb7GhRkPUrx+/Tnt7oj17lI5I5IBDh4iKFNFPMbJ4sZX+iQcH61saLGCq5YgIXsfriy/00+0kvdSuzYOPMjNf1a5d+jVoAaKSJYn++iuN9zs0lFdKBXgG8cWLjfXyzILiSQ0R0Zw5c6h48eLk4OBA/v7+dPz4cd1jAQEB1KVLF93tW7duEYAUl4CAAN02+/btM7hN0v2kRZKaXCIiQj9Sxd2dvy61bs3jpxcv5mlc79zhr7+W6scf9Z92a9cqHY3IQY8fJ/8m36GDlXZHaVscOnZUOpIUEhOJjh/nRKV27ZSL4ObNywnO779nb+R0YiIP6tTOOg0Q1a/PyVEy27dzkw7AQyWPHs3OyzNLmTl/q4isuvRMJyoqCh4eHoiMjIS7dhSDsE6DBwMzZ6a9jaMjDzUqVYovpUvrr/v5AQ4OORJqpi1fzqNxAGDWLB4mJXIVjQaYNg347jtArebRcuvWAVWrKh2ZEZ05A9Sqpa+cL1lS0XCePgV27eIRiTt2AE+eJH+8Rg2geXOu2a9Tx7h15K9eAT/9BEyfDsTE8H3t2gGTfyT4rp3KVeVEPHzq77+5utzKZOb8LUmNsE4xMcCtW8D168CNG3zRXr99G0hMTP25NjY8NCFpspM06VFqdNG//wKffMJnsuHDgalTlYlDmIUjR4AvvgDu3QOcnHjEXPfuqQz7t0RNmwI7dwK9ewO//pqjh9ZogKAg/pf791/gxAm+T8vdHfjwQ05imjbNmTzi3j1g9Gjg998BZ3qN5Tbd0Fqzjh/s0QOYM4e/rFkhSWoMkKRG6CQmAnfupEx2tNe1X4dSU7iw4Rae0qWBfPlMc1Y5cQJo1AiIjgY6d+YWG6s5e4msevIE6NKFT7wA0L49sGABz3Bg8Q4e5DHODg78RcTb26SHe/EC2L1b3xrz5rRqVapwEtO8OVCvnnLTvVzachPO7T9FqdcXEQ97fOs6B6V+6oUePUw604CiJKkxQJIakSFE/GlmKNm5cYMnuEuLh4fhZKdUKf46Z5OFqaFCQvhT9OlToEkTYOtWq59AS2ScRgP8/DMwahQ34pUty91R1aopHVk2EfHkTUeOAEOGcP+LkXd/8SInMdu382HUav3jefLw9ETabiWzmOps926ej+v5c8R6euFLt/VYe7c+AKBCBe6WbN7c+r7vSFJjgCQ1wiieP0+9hefBg7Sf6+Skr+N5M+nx9TWcqDx8yAnN7dtcY7BvXyqz8Inc7uhR7o66e5d7IX75hXslLPoEt307n6Xz5OEJJvPnz9buXr4E9uzRdyvdv5/88fLl+XDNm/Ms4GbTm0PEmeuIEZzF+vsDGzYgoVARLFwIjB/P33kA4P33eVOLT2qTkKTGAElqhMlFR6ddx5P0a+CbbG25jidpslOyJDBhAnD+PN9/5AhPRyxEKp4+5e6obdv49hdfAAsX6ld4sDhEwFtvcYHLuHF89s7k069e1Scxhw7xzL5azs7cq6ttjSlRwqjRG0d0NBdLrVnDt7/8Epg/n78k/d+LFzwJ8y+/APHxnMh++SUwcaJ11A1LUmOAJDVCUQkJqdfx3LiRdh2Plxd/DVd4BIiwDBoNMGMGd0clJvLqGevW8ZJVFumvv3htt7x5ubUmnYKh16+5QVObyISFJX+8VClegqRZMy7ZcXY2YezZdfs2r5sQFMQFM7NmAV9/nWrz261bwMiR/H4DgIsLjykYOtSyV0+RpMYASWqE2SLibiZDXVpxccCKFTxmVIhMOHaMyy+03VGzZvFaQxbXHaVW84qu167x2OZhw1JsEhqqT2IOHOB/Gy1HR05etN1KZcrkYOzZsXcvJ3NPn/LKt3//DTRsmKGnapcpO36cb/v48IK6nTtzo7ClkaTGAElqhBC5zbNnvI7h1q18u21bXhnaXD4CExO5kfLNS3R08ts+u5ajwZIv8drNC3OH3MarRCfExPB6rvv3c/6flK+vPol57z0La6UgAmbP5uYVtRqoWRPYuDHTlcpE3GIzciQ3+ADcWvfzz9zlZkkkqTFAkhohRG5ExHNRjhjBSUTp0nyyM9T4l5BgOKlI65LRbQ1tl9Z0UUnZIQHXURq+uIM+mI8F6JPscXt7HiilTWTKl7fAFimAfym9egErV/LtTp24KCobfWSxsTyFzaRJnAQCwMcf80ip8uWNEHMOkKTGAElqhBC52fHj3FJz5w5P/VKiRMokI61adlNzcuJzd2qXVg/noueF/nji6ouJnULh6GoPZ2dufXj/ffNpfcqyO3eAVq14NmVbW25SGTDAaNnZkyfA99/zPIZqNR+iVy+uvS5Y0CiHMBlJagyQpEYIkds9e8ajYrZsSX/bpAmFi0vaCcebl8xu7+SUgSmcYmJ4CZNHj7jOrHNnY/xKlPf6NQ/L6twZePyYh63/9Rf3m5nAtWtcPKz9G3B35yU3BgxINqDKrEhSY4AkNUIIwd1RZ85wd1BqSYajo5l230ydykUi5csDly9nbTLLnETETSR37vAwrLCwlNeTLiRVvTrXz/j5mTy0fft4TsNz5/i2ry8wZQq35pnbey9JjQGS1AghhIWLiuL5nCIjgfXrubtGSQkJPINfWklLesuuADxMvW1bLhB2cTF93P+n0XD5znff6ScifPtt7vmqVy/HwkiXJDUGSFIjhBBWYMwYHp/81lvA6dOmbVZ49SrthOX+/eQrXaamcGFuCvH15aTszeseHoo2j0RHcyIzdSr3hgFA69bccmMO02NJUmOAJDVCCGEFnjzhRCA6mleebNIka/sh4hqW1BKWsLD013oDuOq6WLHUk5ZixcxovYW0PXzIOePSpfzrcXAA+vfn1cE9PZWLS5IaAySpEUIIKzF4MI9Tb9iQZ9szJD6eW1JSS1ru3OHxzunx9NQnKoZaWby8zL+2J5MuXOBpcnbv5tv58/MqFb17K7OWriQ1BkhSI4QQVuL+fR6TnpDAk7DY2qZMXh484OaGtKhUgLd32l1DufR8QcQNYUOGAFeu8H1ly/Kkzp98krO9ZZLUGCBJjRBCWJGePYHFi9PextHRcKKivV60KPexiFQlJgK//QaMHcu9dQAvOzFjBpc15QRJagyQpEYIIazI3bu82KNanXorS6FC5jc+2UJFRXHh8IwZvLaWSsUTHk+axLmhaY8tSU0KktQIIYQQ2XPnDvDtt8CqVXzb2Zm7qIYPT3cB9SzLzPnbuqqbhBBCCGEyxYsDf/wBnDgBNGjA0/D88AOvfr54sbJLbQCS1AghhBAik/z9gYMHeQ7EUqWAiAguc6peneu0lSJJjRBCCCEyTaXiSZ2Dg3mEfd68PCDNx0e5mCSpEUIIIUSWOTgAgwYB168D69YpM5eNlp1yhxZCCCGEtciXjy9KkpYaIYQQQlgFSWqEEEIIYRUkqRFCCCGEVZCkRgghhBBWIdcUCmsnTo6KilI4EiGEEEJklPa8nZEFEHJNUvPy5UsAQLFixRSORAghhBCZ9fLlS3h4eKS5Ta5Z+0mj0eDBgwdwc3ODShY4MygqKgrFihXD3bt3ZX0sMyDvh/mR98S8yPthXkz1fhARXr58CR8fH9jYpF01k2taamxsbFDU1EuJWgl3d3f5gDAj8n6YH3lPzIu8H+bFFO9Hei00WlIoLIQQQgirIEmNEEIIIayCJDVCx9HREePGjYOjo6PSoQjI+2GO5D0xL/J+mBdzeD9yTaGwEEIIIaybtNQIIYQQwipIUiOEEEIIqyBJjRBCCCGsgiQ1QgghhLAKktTkMn5+flCpVCkuffv2xbNnz9C/f3+UK1cOzs7OKF68OAYMGIDIyEilw7Zqab0nSRERmjVrBpVKhU2bNikTbC6Qkffj2LFjaNSoEfLkyQN3d3c0bNgQMTExCkZtvdJ7P8LDw9GpUycULlwYefLkwVtvvYX169crHLX1UqvVGDNmDEqUKAFnZ2eUKlUKEydOTLYuExFh7Nix8Pb2hrOzMxo3bozQ0NAciS/XzCgs2KlTp6BWq3W3L126hA8++ACtW7fGgwcP8ODBA0yfPh0VK1ZEWFgYevfujQcPHuDvv/9WMGrrltZ7ktSsWbNkiY8ckN77cezYMTRt2hSjRo3CnDlzYGdnh/Pnz6c7fbvImvTej86dO+PFixfYsmULChQogNWrV6NNmzY4ffo0atSooVTYVmvq1Kn49ddfsWLFClSqVAmnT5/Gl19+CQ8PDwwYMAAA8NNPP+GXX37BihUrUKJECYwZMwZNmjRBcHAwnJycTBsgiVxt4MCBVKpUKdJoNAYfX7duHTk4OFBCQkIOR5Z7GXpPzp07R0WKFKGHDx8SANq4caNyAeYyb74fderUodGjRyscVe715vuRJ08e+v3335Ntky9fPlq8eLES4Vm9jz76iLp165bsvlatWlGHDh2IiEij0VDhwoVp2rRpusdfvHhBjo6OtGbNGpPHJ18tcrH4+Hj88ccf6NatW6otAJGRkXB3d4ednTTq5QRD70l0dDTat2+PefPmoXDhwgpHmLu8+X48evQIJ06cQKFChVCvXj14eXkhICAAhw8fVjrUXMHQ/0e9evXw559/4tmzZ9BoNFi7di1iY2Px7rvvKhuslapXrx727NmDkJAQAMD58+dx+PBhNGvWDABw69YthIeHo3HjxrrneHh4oE6dOjh27JjJ45MzVS62adMmvHjxAl27djX4+JMnTzBx4kT07NkzZwPLxQy9J9988w3q1auHwMBA5QLLpd58P27evAkAGD9+PKZPn47q1avj999/x/vvv49Lly6hTJkyCkZr/Qz9f6xbtw5t27ZF/vz5YWdnBxcXF2zcuBGlS5dWLlArNnLkSERFRaF8+fKwtbWFWq3GpEmT0KFDBwBc4wQAXl5eyZ7n5eWle8yUJKnJxZYsWYJmzZrBx8cnxWNRUVH46KOPULFiRYwfPz7ng8ul3nxPtmzZgr179+LcuXMKR5Y7vfl+aDQaAECvXr3w5ZdfAgBq1KiBPXv2YOnSpZg8ebJiseYGhj6zxowZgxcvXuC///5DgQIFsGnTJrRp0waHDh1ClSpVFIzWOq1btw6rVq3C6tWrUalSJQQFBWHQoEHw8fFBly5dlA5Pampyq9u3b5ONjQ1t2rQpxWNRUVFUt25dev/99ykmJkaB6HInQ+/JwIEDSaVSka2tre4CgGxsbCggIEC5YHMBQ+/HzZs3CQCtXLky2bZt2rSh9u3b53SIuYqh9+P69esEgC5dupRs2/fff5969eqV0yHmCkWLFqW5c+cmu2/ixIlUrlw5IiK6ceMGAaBz584l26Zhw4Y0YMAAk8cnNTW51LJly1CoUCF89NFHye6PiorChx9+CAcHB2zZssX0lepCx9B7MnLkSFy4cAFBQUG6CwDMnDkTy5YtUyjS3MHQ++Hn5wcfHx9cu3Yt2bYhISHw9fXN6RBzFUPvR3R0NACkGHlma2ura1UTxhUdHZ3m77tEiRIoXLgw9uzZo3s8KioKJ06cQN26dU0foMnTJmF21Go1FS9enEaMGJHs/sjISKpTpw5VqVKFrl+/Tg8fPtRdEhMTFYo2d0jtPTEEMvrJ5NJ6P2bOnEnu7u70119/UWhoKI0ePZqcnJzo+vXrCkSaO6T2fsTHx1Pp0qXpnXfeoRMnTtD169dp+vTppFKpaNu2bQpFa926dOlCRYoUoX/++Ydu3bpFGzZsoAIFCtDw4cN120yZMoU8PT1p8+bNdOHCBQoMDKQSJUrkSMu/JDW50M6dOwkAXbt2Ldn9+/btIwAGL7du3VIm2FwitffEEElqTC+992Py5MlUtGhRcnFxobp169KhQ4dyOMLcJa33IyQkhFq1akWFChUiFxcXqlq1aooh3sJ4oqKiaODAgVS8eHFycnKikiVL0nfffUdxcXG6bTQaDY0ZM4a8vLzI0dGR3n///Qx9thmDiijJNIBCCCGEEBZKamqEEEIIYRUkqRFCCCGEVZCkRgghhBBWQZIaIYQQQlgFSWqEEEIIYRUkqRFCCCGEVZCkRgghhBBWQZIaIYQQQlgFSWqEEEIIYRUkqRFCCCGEVbBTOoCcotFo8ODBA7i5uUGlUikdjhBCCCEygIjw8uVL+Pj4pFgh/E25Jql58OABihUrpnQYQgghhMiCu3fvomjRomluk2uSGjc3NwD8S3F3d1c4GiGEEEJkRFRUFIoVK6Y7j6cl1yQ12i4nd3d3SWqEEEIIC5OR0hEpFBZCCCGEVZCkRgghhBBWQZIaIYQQQlgFSWqEEEJYplevgIQEpaMQZkSSGiGEEJbn9Gkgf34gb17gww+BSZOAQ4eAuDilIxMKyjWjn4QQQliRSZOA+Hi+7N7NFwBwcgLefhto2BAICODrLi7KxipyjCQ1QpijuDhg7FjgwAFg8WKgShWlIxLCfISGAps38/XNm4E7d4CDB/n/5dEjYP9+vgCAvT1QuzYnOA0bAvXrAxmY70RYJhURkdJB5ISoqCh4eHggMjJS5qkR5u3yZaB9e+DCBb7t56dvahdCAH37AvPnA82bA9u26e8nAq5d0yc4Bw4A9+8nf66tLfDWW/qWnAYNuAtLmK3MnL8lqRHCXBDxB/XQoUBsLFCwIDebh4UBjRsD27cDdtK4KnK5p0+BYsWAmBhgzx6gUaPUtyUCbt3SJzgHD/LtpFQqoGpVTnACAoB33uH/PWE2JKkxQJIaYdYePQK6ddN/62zaFFi2DHj8GKhbF3j9GhgyBJg+Xdk4hVDaDz8AY8YA1asDZ89yUpIZd+/qE5wDB4CQkJTbVKyo764KCAC8vY0SusgaSWoMkKRGmK3t24GuXTmxcXQEfvoJ6NcP0K5G+/ffQOvWfH3VKu6aEiI3io3l7tiICGDlSqBjx+zv8+FDHjWlbc25fDnlNmXK6BOcgACgePHsH1dkmCQ1BkhSI8xOTAwwYgQwZw7frlwZWL3acFHwd98BP/7IIzuOHgVq1MjZWIUwB0uXAl99BRQpwt1I9vbGP8aTJ5zkaFtygoK4GyspX199gtOwIVCqVOZbjESGSVJjgCQ1wqxcvMgtLpcu8e0BA4ApUwBnZ8Pbq9VAixbcqlO8OBcOS7+/yE2IOOG/fJlbM4cNy5njvngBHDmib8k5c4b/H5Py8UneklO+vCQ5RiRJjQGS1AizoNFwy8yIETxs28uLa2eaNUv/uS9eAP7+PJz1vfeAXbukcFjkHjt28P+JqyvXxXh6KhPHq1fcWqqtyzlxIuWsxgUL6pOchg05GbORuW6zSpIaAySpEYp7+BD48ktg506+/fHHwJIlQKFCGd9HcDBQpw5/sA4aBMycaZJQhTA7H3wA/Pef+f3dx8QAx4/ru6uOHePan6Ty5uVRVdpEp3p1+UKSCZLUGCBJjVDU1q08uunJE66L+flnoE+frDVRb9wItGrF13//HejUybixCmFuzp/nRMDGBrhxg4uFzVVcHHcPa1tyjhzhLyFJubnxJIDa7qqaNQEHB2XitQCS1BggSY1QRHQ0zzvz6698u1o1LgauWDF7+x07Fpg4kROkw4f5Q1EIa9WlCyfwbdoAf/6pdDSZk5jIQ8+1LTmHDgGRkcm3cXEBAgOBb7/lAQMiGUlqDJCkRuS4oCAuBr5yhW8PGcLr1Tg6Zn/fGg1/CP7zD09Edvp05rqxhLAUDx5wy0xCAtev+PsrHVH2qNU8UCDphIBPn+ofb9UKGD1aRjgmkZnzt1QuCWFsGg13L/n7c0Lj7c1FvdOnGyehAbgZ/o8/gHLluGiydeuUxYpCWIM5c/hvu0EDy09oAF6moXp1YOBAYMMGnp/q1Cn+H1ap+L633uLRjidPKh2txZGkRghjevAAaNKEu5wSErg15cIFLnI0Ng8PYNMm7p8/eJBbgoSwJq9eAQsW8HVr/fu2sQFq1QLWreMpHtq35/v++YcHBTRtynU5IkMkqRHCWDZu5KGb//3H880sXMj3FShgumOWL88tNgB/o12+3HTHEiKnLVvGUxmULs0tF9auYkWeNfzKFa4jsrXl0ZINGgDvv8/dVSJNWUpq5s2bBz8/Pzg5OaFOnTo4mUYTWUJCAiZMmIBSpUrByckJ1apVw44dO1Jsd//+fXTs2BH58+eHs7MzqlSpgtOnT+sef/XqFfr164eiRYvC2dkZFStWxAJtBi+Ekl6/Bnr25L7wZ8+46fjsWb4vJybg+uQTYPx4vt67tzRZC+ugVgOzZvH1b77hE3xuUbYsf0EJCQG6d+eZk/fuBd59l4eF796dcpZjwSiT1q5dSw4ODrR06VK6fPky9ejRgzw9PSkiIsLg9sOHDycfHx/atm0b3bhxg+bPn09OTk509uxZ3TbPnj0jX19f6tq1K504cYJu3rxJO3fupOvXr+u26dGjB5UqVYr27dtHt27dooULF5KtrS1t3rw5Q3FHRkYSAIqMjMzsSxYidadPE5UtSwQQqVREw4cTxcXlfBxqNVFgIMdRpAjRw4c5H4MQxvT33/z3nC8f0evXSkejrLAwoq+/JnJw4N8JQFSnDtG2bUQajdLRmVxmzt+ZTmr8/f2pb9++uttqtZp8fHxo8uTJBrf39vamuXPnJruvVatW1KFDB93tESNGUIMGDdI8bqVKlWjChAnJ7nvrrbfou+++y1DcktQIo0pMJJoyhcjOTp9I7NmjbEyRkUTly3M8DRook1wJYSx16/LfcgY/43OFe/eIBg4kcnLSJzc1axJt3MhfbKxUZs7fmep+io+Px5kzZ9C4cWPdfTY2NmjcuDGOHTtm8DlxcXFwcnJKdp+zszMOHz6su71lyxbUqlULrVu3RqFChVCjRg0sXrw42XPq1auHLVu24P79+yAi7Nu3DyEhIfjwww8z8xKEyL67d4HGjYGRI3kOis8+42LgRo2UjcvdHdi8mX8ePsxN9kJYomPH+OLgwCvWC1akCHfJ3b7Na1/lycNrUbVsyUPA//qLR1/mZpnJlu7fv08A6OjRo8nuHzZsGPn7+xt8Trt27ahixYoUEhJCarWadu3aRc7OzuTg4KDbxtHRkRwdHWnUqFF09uxZWrhwITk5OdHy5ct128TGxlLnzp0JANnZ2ZGDgwOtWLEi1VhjY2MpMjJSd7l796601Ijs++svorx5+RtSnjxES5aYX/PvP/9wVxhA9NtvSkcjROZ99hn//X75pdKRmLfHj4m+/ZbIzU3fclOhAtGqVdyabCVM1v2UlaTm0aNHFBgYSDY2NmRra0tly5alr7/+mpycnHTb2NvbU926dZM9r3///vT222/rbk+bNo3Kli1LW7ZsofPnz9OcOXPI1dWVdu/ebfC448aNIwApLpLUiCyJiuIPWO0HR+3aRCEhSkeVuokTOU4HB6Jjx5SORoiMu35dn5RfvKh0NJbh2TOiceOIPD31n1FlyhAtW0YUH690dNlmsu6nAgUKwNbWFhEREcnuj4iIQOHChQ0+p2DBgti0aRNev36NsLAwXL16Fa6urihZsqRuG29vb1R8Y9r4ChUq4M6dOwCAmJgYfPvtt5gxYwZatGiBqlWrol+/fmjbti2mT59u8LijRo1CZGSk7nL37t3MvFQh9E6e5KbdZct4NNO33/K8EWXKKB1Z6r79lkdjxcfzz4cPlY5IiIyZNYtPy02ayJIBGZU3L4+AvH2bZy3Pnx8IDeUFdMuVAxYv5s+CXCBTSY2DgwNq1qyJPXv26O7TaDTYs2cP6tatm+ZznZycUKRIESQmJmL9+vUIDAzUPVa/fn1cu3Yt2fYhISHw9fUFwMPCExISYPPG0u22trbQpNJ/6OjoCHd392QXITJFreYPiHr1eBG9YsWA/fv5Pnt7paNLm40NDwmtWJETms8+44X2hDBnz54BS5fydWudbM+UPDz4C83t28BPP/HSKbdu8fQSpUsD8+alXEHc2mS2GWjt2rXk6OhIy5cvp+DgYOrZsyd5enpSeHg4ERF16tSJRo4cqdv++PHjtH79erpx4wYdPHiQGjVqRCVKlKDnz5/rtjl58iTZ2dnRpEmTKDQ0lFatWkUuLi70xx9/6LYJCAigSpUq0b59++jmzZu0bNkycnJyovnz52cobhn9JDLl9m2id97RN+W2bctNvJYmNFTfJN2zp9LRCJG2H3/kv9WqVc2vVs0SvX5NNGsWkbe3/rPM25to5kyLGiZv0iHdRERz5syh4sWLk4ODA/n7+9Px48d1jwUEBFCXLl10t/fv308VKlQgR0dHyp8/P3Xq1Inu37+fYp9bt26lypUrk6OjI5UvX54WLVqU7PGHDx9S165dycfHh5ycnKhcuXL0888/kyaDf/iS1IgMW7OGyMODPwBcXYlWrLDsD9jt2/U1CgsWKB2N5Xr9mmj3bquoUTBLcXH6k28ag0BEFsTEEM2bR1SsmD65KVSI6KefiF6+VDq6dGXm/C2rdAuhFRXFw0dXruTbb7/NSxCUKqVsXMYwZQowahR3m+3bB9Svr3REluXKFV5w8PJlnr5elqMwvhUrgK5dAR8f7jJxcFA6IusTH8+/58mT+XcMcP3NN9/wZ5+Hh7LxpUJW6RYis44d45VzV67kepSxY4FDh6wjoQGAESP0K3l//jlw/77SEVmO1auB2rU5oQH4pHD0qLIxWRsiXtkeAPr3l4TGVBwcgB49gGvXeOBDmTLA06fA6NGAnx8XGz9/rnSU2SJJjcjdEhOB778H3nmHv7n4+fGK199/D9jZKR2d8ahUXIBZpQoQHi6FwxkRGwv06gV06MDrezVqBLRpw4/168eF5MI4/vsPuHiRJ5Pr1UvpaKyfvT23igUH8wKaFSrwwqHffw/4+gLffQc8eaJ0lFkiSY3IvW7d4sXhxo/nE1THjkBQkPV2zbi6Aps28fDPEyeAvn1lUbzUXL8O1K0LLFrECeG4ccCuXcDcuYCnJ3DuHPDbb0pHaT20rTTduvHfp8gZdnZA+/bApUvAunVA1arAy5fAjz/yF7xhw4A3pnAxd5LUiNyHiGtlqlXjbid3d/62snKl2fYpG03JksDatdzFtmQJICvdp/T337zSelAQULAgsHMnJ762tnx7wgTe7ttvueleZM+lS/w7trEBBg1SOprcycaGu6fPnQM2buS//9evgenTgRIl+H158EDpKDNEkhqRu7x4wd0JnTrxN5L69YHz5/nbSm7x4YdcOAwAAwZw7ZDg7rgBA/jD/eVL7pIMCgI++CD5dn36cDfes2fAmDGKhGpVZszgny1bctItlGNjA3z6KXD6NLBtG1CnDhATA8yeze9N377A/yfFNVeS1Ijc49AhLgZes4a/dU+YwJPp+fkpHJgChg4FvviCa4o+/xy4d0/piJR1+zYnMXPm8O2RI4G9e3kkzpvs7PTbLVjA325F1oSHcyspIJPtmROVCmjenFuyd+3i/424OGD+fJ7Er0cP4OZNpaM0SJIaYf0SEvgb9bvvAmFh/I3j8GG+z5qKgTNDpeKakGrVgEePeCkFa59pNDVbtvAyGKdOAfny8TfUyZPT/tsICOCkkIhH60htUtbMncvDjOvW5YswLyoVt1QePMhfABs14s/T334DypblYuOQEKWjTEaSGmHdrl8HGjQAfvgB0Gh4jpGgIJ6DJrfLk4f7z/Pl4xN679656+SckMCFkIGB3C359tvc6tK8ecaeP20a4OLC64BpWxtExr1+Dfz6K1+XVhrzFxAA7NnDf+9Nm/LgihUreORUhw48ksoMSFIjrBMRT5BWowYvSOnpyQWyy5cDbm4KB2dGSpTgUQ82NvwBNXeu0hHljHv3uOVOuyDuN98ABw4AxYtnfB9Fi+praoYN48kbRcatWMF1SSVLch2HsAz16gHbt/MIyhYt+Mvi6tW8+Gjr1lyjqCBJaoT1ef4caNuWV6h99YqHbZ8/z/eJlN5/n1sdAD6579+vaDgmt2MH11YdPcqj3TZs4GLVrEz49s03PIFZeDgwcaLRQ7VaajUwcyZfHzSIa9yEZfH3567bs2e5+5qIRw7WqKFovY0skyCsz7vv8rduOzs+0QwbJh+a6SHiEWGrVgEFCgBnzmSu1cISJCby0OxJk/j2W28Bf/2V/RE327dzl5WdHU8gV758tkO1eps28WgnT0/g7l2eQ0lYtkuX+H8rJobfXyOSZRJE7nX8OCc0Dg7c9ztypCQ0GaFS8URzNWrwTKItW/KHk7V4+BBo3Fif0Hz9Nf99GGMIcbNm3AyfmMhDwnPH98Ts0U6217u3JDTWonJlHln699+KhiFJjbAus2fzz/btuXlUZJyLCxcOFyjATco9e1rHCXrvXu5uOnCAT6Br1wLz5gFOTsY7xsyZgKMjsHu30b+lWp0TJ3j0ob09jxwT1kXhEaWS1Ajrce8edycAwMCBysZiqXx9+Xdoa8uzLmuTREukVvNcRI0b87D1qlW5W80UtVWlSnE3J8B1NtHRxj+GtdC20rRrZ3geICGyQZIaYT3mz+cTWUAAfzMXWfPuu/oTz9Ch3NJhaR494m6hceO4tal7d+6aLFvWdMccNQooVoznQvrpJ9Mdx5LdugWsX8/XBw9WNhZhlSSpEdYhOhpYuJCvy/ox2TdgANC5MyeJbdrwjLuW4tAhrg3avZu71FasABYvBpydTXtcFxd9Mjh1Kp/ARXKzZ/MQ4MaNeeJHIYxMkhphHf74g+e8KFGCizZF9qhUvARAzZq8aGPLlubfpaLRcDLx3nu8+F6FCjypYOfOORfD55/zrKuxsdIS8aYXL3gRVYBbAIUwAUlqhOUj0td+DBggo52MxdmZC4cLFuRZmLt3N9/C4adPgU8+4dFuajUPTz91CqhYMWfjUKmAX37hv8FNm3j1acEWLeJ5oypX5kVVhTABSWqE5fvvP56i29WVJ9wTxlOsGA/RtLPj4ZraFZXNyfHj3N20bRuPaFq8mLuc8uRRJp5KlTi5BvhnfLwycZiT+HhO9gBuwVKplI1HWC1JaoTlmzWLf3brxjPECuNq2FD/Ox4+nGtVzAERD6V+5x2ewK1MGU5wundX/qQ5bhzg5cWL/VnyCDJjWbcOuH8fKFyYp1sQwkSylNTMmzcPfn5+cHJyQp06dXDy5MlUt01ISMCECRNQqlQpODk5oVq1atixY0eK7e7fv4+OHTsif/78cHZ2RpUqVXD69Olk21y5cgWffPIJPDw8kCdPHtSuXRt37tzJyksQ1uLaNeDff/kkJnNemM7XX3MrmEbDq1MrXQT74gXw2Wf8rT8xkYuZT582n+JTDw9gyhS+PmEC1/jkVkT6Aup+/Xg+HyFMhTJp7dq15ODgQEuXLqXLly9Tjx49yNPTkyIiIgxuP3z4cPLx8aFt27bRjRs3aP78+eTk5ERnz57VbfPs2TPy9fWlrl270okTJ+jmzZu0c+dOun79um6b69evU758+WjYsGF09uxZun79Om3evDnV474pMjKSAFBkZGRmX7IwZ337EgFELVooHYn1i4kh8vfn33fVqkSvXikTx+nTRCVKcBwODkTz5hFpNMrEkha1mujttznODh2UjkY5e/bw78DZmejJE6WjERYoM+fvTCc1/v7+1LdvX91ttVpNPj4+NHnyZIPbe3t709y5c5Pd16pVK+qQ5J98xIgR1KBBgzSP27ZtW+rYsWNmw9WRpMYKPXtG5OLCH5h79igdTe5w7x6Rlxf/ztu0ydlkQqPhBMbBgY9fogTRqVM5d/ysOH2aSKXieA8eVDoaZTRvzq//66+VjkRYqMycvzPV/RQfH48zZ86gcePGuvtsbGzQuHFjHDt2zOBz4uLi4PTGdOTOzs44fPiw7vaWLVtQq1YttG7dGoUKFUKNGjWwePFi3eMajQbbtm1D2bJl0aRJExQqVAh16tTBpjSmI4+Li0NUVFSyi7AyS5bwMOMqVXgYrzC9IkX0hcPr1ulX9za1ly95Btq+fbno9NNPeSmHWrVy5vhZVbMm0KMHX+/fn0dm5SZXrui7h7/5RuloRG6QmWzp/v37BICOHj2a7P5hw4aRv7+/wee0a9eOKlasSCEhIaRWq2nXrl3k7OxMDg4Oum0cHR3J0dGRRo0aRWfPnqWFCxeSk5MTLV++nIiIHj58SADIxcWFZsyYQefOnaPJkyeTSqWi/fv3GzzuuHHjCECKi7TUWImEBCJfX/4G+NtvSkeT+/z6K//uVSqiHTtMe6zz54nKluXj2dkRzZhhnt1NqXn8mChvXo5/3jylo8lZ3bvz6/70U6UjERbMZN1PWUlqHj16RIGBgWRjY0O2trZUtmxZ+vrrr8nJyUm3jb29PdWtWzfZ8/r3709vv/12suO2a9cu2TYtWrSgL774wuBxY2NjKTIyUne5e/euJDXW5O+/+cOyQAGi6Gilo8l9NBr9CcvTkyg01DTH+O03IicnPk6xYkTHjhn/ODlh7lx+DXnzcpKTG4SHEzk68us+dEjpaIQFM1n3U4ECBWBra4uIiIhk90dERKBw4cIGn1OwYEFs2rQJr1+/RlhYGK5evQpXV1eULFlSt423tzcqvjFJVoUKFXQjmwoUKAA7O7s0t3mTo6Mj3N3dk12EFdEOMe7d2/TT34uUVCpg7lzg7bd5JNKnn/LEasby+jXQtSsPz46NBZo3B86d4+NZol69eGTW8+fAd98pHU3OmD8fiIsD/P2B+vWVjkbkEplKahwcHFCzZk3s2bNHd59Go8GePXtQt27dNJ/r5OSEIkWKIDExEevXr0dgYKDusfr16+PatWvJtg8JCYGvr6/uuLVr105zG5GLnDkDHD4M2NsDffooHU3u5ejIixMWLgxcvsxJiDFmHA4O5hPh77/zzLyTJwNbtwL582d/30qxswPmzOHrixfz37A1i4nhpAYAhgxRft4gkXtkthlo7dq15OjoSMuXL6fg4GDq2bMneXp6Unh4OBERderUiUaOHKnb/vjx47R+/Xq6ceMGHTx4kBo1akQlSpSg58+f67Y5efIk2dnZ0aRJkyg0NJRWrVpFLi4u9Mcff+i22bBhA9nb29OiRYsoNDSU5syZQ7a2tnQog82aMvrJinTqJMNkzcnRo0T29vyeTJqUvX2tXKkf0ebtTXTggHFiNBft2/Nrq1uXh3xbqwUL+HX6+nL9mxDZYNIh3UREc+bMoeLFi5ODgwP5+/vT8ePHdY8FBARQly5ddLf3799PFSpUIEdHR8qfPz916tSJ7t+/n2KfW7dupcqVK5OjoyOVL1+eFi1alGKbJUuWUOnSpcnJyYmqVatGmzZtynDMktRYiQcP9CdQcx/Om5ssWqQvHN62LfPPj44m6tGD9wEQNW5MlME5qCzK/ftErq78Gv8/EMLqqNX6wu6ZM5WORliBzJy/VUTmukKdcUVFRcHDwwORkZFSX2PJxo4FJk7kPvok0wIIM9CnD6/s7eHBi0mWKZOx54WE8IzA589zN8W4ccDo0da7MOlPPwEjRvAyCteuWd/SHlu38uKiHh68fIWbm9IRCQuXmfO3rP0kLEdsLJ80AWDgQGVjESnNns3JZmQkFw6/fJn+c9at47lmzp8HChXidaXGjbPehAYABg0CypYFIiKA779XOhrj0y6J0LOnJDQix0lSIyzHmjXA48e8cnTLlkpHI97k4MAT8/n4cLFv5868VpQhcXE8GV3btpz8BATw6Kb338/ZmJXg4KBfsXrOHP5d5aB//+X5KqtVAz7/HPj2W2DZMm74jIjIZq336dPAgQNcGK1dqVyIHGSndABCZAiRfhh3//78oSnMT+HCwIYNvLL3pk3ApEnAmDHJt7l1S78AJcBn1e+/z13vaZMmQGAgsHkzn/x3786REUJ//MGD1LQTG1+4kHIbd3fuOXzzUrYskC9fOgfQttK0bQsULWrM0IXIEKmpEZZh/35eCsHFBbh3D8ibV+mIRFqWLgW++opP1Js3Ay1a8P2bNwNdunAXVb58fJZt1kzZWJVy6xZQoQK3Wv31FzebmNCsWfqVCjp25MXWQ0P1l5AQ4M6dtFtq8uUznOyUKQO4v7gDlCzJGdPZs0CNGiZ9PSL3yMz5Oxd9NRIWTdtK06WLJDSWoFs3PrHNm8dn0CNHuI9jxgx+vG5d4M8/uSsxtypRgguGJ0wABg/mCQZdXIx+GCJuLJs0iW8PGsQNKjYGig9iY4GbN5MnOtrr9+8Dz54BJ07w5U3znWejj1qNYO9G2LyjBsrc4GSndGkgTx6jvywhDJKWGmH+bvz/05GIF8grX17piERGJCRwjcyhQ1z4q+3zGDKEJ9Szt1c2PnMQHc2tNXfu8IiviRONunu1mtcAXbiQb0+aBIwalbWertevgevXk7fuaBOf2EeRuIticMdLNMc2bEfzZM/18dG36CS9lCoFvLHesRApZOb8LUmNMH+DBvHImmbNuMpRWI6ICB7ddO8e4OkJLF/OtSRCb8MG4LPPeIbm4GDuwjGCuDhuJPv7b05iFizgAUmmEDvpZziNHorIIhXwS49LCLluo0t6nj1L/XkqFVC8uOHurBIlJO8VTJIaAySpsVBRUVxw+PIlsGMHF1gKy3LtGrByJdfYlCihdDTmhwj48EPgv/94fpfNm7O9y5cveYDgnj082GrVKhOW7CQkcJPL3bu8BET37skefvYsZVeW9hIVlfpubW0BP7/kiY724utr3aP+RXKS1BggSY2Fmj2bW2oqVOD1hWQNGWGNrlwBqlYFEhO5NTIbxdOPH3N5zunTgKsrD0Iz6Uj5NWuA9u15nqGwsAz3JxEBjx4Z7s66fp175lJjb88NWmXLAm+9BbzzDlCnDr9ekfOI+H07cgR48gQYPty4+5ekxgBJaiyQWs2fWjdvctt5r15KRySE6QwdyhW8ZcoAFy9yd1Qm3bnDjT7XrgEFCnB+VLu2CWLVIuIDnDnDw/LHjjXabh88MFywfOMGd629ydaWE5wGDfhSvz5P2iyMLy6O3/IjR/SXJ0/4MWdnHtxozK5DSWoMkKTGAm3ZwvUXefNyTYYJRoYIYTaiooBy5YDwcGDKFB4ZlQlXrnBCc+8eDyrbtSsHauoPHADefZdbZ+7cAQoWNPEB+bvOvXuc6Fy9Chw/zrXod++m3LZsWX2S88473Esmjb2Z9/QpcPSoPoE5dSplYunkBPj7czI5ciTPd2QsktQYIEmNBWrUCNi3jz/cp0xROhohTG/lSp6JOU8ebm4pUiRDTztxgrucnj3jntqdO3NotPwnn/BaT7166ZcwUcidOzwr8uHDnORcupRyGy8vfYLToAHPqpyb5nzMCCJuDTt8WJ/EXLmScruCBfUtYvXrcyuZg4NpYpKkxgBJaizM+fNA9ercpnzrVu6ez0TkHkR8pjh6FGjXDli9Ot2n7NoFtGrFQ679/bnLKX/+HIj12jV9U9DVq9zKZEaeP+df46FDfII+dQqIj0++jasrT5mkbc2pUyf3zakTH88rlBw5ok9kHj1KuV358py8aBOZ0qVzrtVLkhoDJKmxMN268WRtbdrwJG1C5BbnzgE1a3KCs38/r4uVinXreNh2QgJ3Pa1fn4PFsr178wQ4LVpwV7GZi43lxEbbmnPkCNd+JGVnl7wup0GDHOlRy1HPnwPHjulbYU6c4N9NUg4OXCqlbYWpV49rtJQiSY0BktRYkEePePKKuDj+qlW3rtIRCZGz+vTh7pwqVXhmZgN9JL/+yhPrEfFSS7//brrm/xQeP+b/0djYdBMvc6VW84BKbXfVoUM8a/KbypVL3mVVsqTl1OUQAbdvJ+9Kunw55VIY+fNz4qJthalZ07wmRZSkxgBJaizIxIk8isLfn6sALeUTRAhjefqUq1yfPeOVvPv10z1ExP8i48bx7T59eJMcnbdlwgQOoGZNbv6wgv9RIn1djrbL6vLllNt5eydvyala1XzqchISuOc+aRLz8GHK7cqUSd6VVK6ceb+FktQYIEmNhYiP55m1wsN5xrD27ZWOSAhlLFjAGYunJw/1KVgQGg0wcCAwdy5vMnYsMH58Dp+QYmP5f/TRI675adcuBw+es54909eaaOtyEhKSb+PmlrIuJ6cGakZG8vc+bYwnTqSc38fennPPpF1JljbUXZIaAySpsRB//AF06sSLxdy6lYPt6UKYGbWaCxvOnQO6d0f8vMXo2pXnugOAX34B+vdXIK7ffgN69ODi/Rs3ctVaBjEx+rqcQ4e4d/zNWZHt7DiJ0HZZ1a9vvHoUbUuSthXmwoWUXUmensm7kmrX5rljLJkkNQZIUmMBiHidoLNneeW9b79VOiIhlHXkCNCgAUilwqC3T+CXY7VhZ8f1M4o0kGg0QOXKPMZ3+nRenDQXU6t56Li2u+rQIZ408E3ly+trcho04NVC0mtdS0zkORiTJjH37qXcrmRJfStMgwY8pN/QCuyWTJIaAySpsQCHD/N/vpMTz6SlZLm9EGYirm1nOK5biRPwRyOnY1i/0QZNmyoUzLZtwMcfc5/L3buAh4dCgZgnIl4pQpvkHD7Ma5S+yccnZV1OdDR3H2m7ko4fB169Sv48W1ugRo3k88N4e+fMa1NSps7flAVz584lX19fcnR0JH9/fzpx4kSq28bHx9P3339PJUuWJEdHR6patSpt3749xXb37t2jDh06UL58+cjJyYkqV65Mp06dMrjPXr16EQCaOXNmhmOOjIwkABQZGZnh54gc9vnnRABR9+5KRyKEWbh3jyig7AOKhBsRQKHfLlU2oPfe4//RwYOVjcOCPH5MtHkz0bBhRG+/TWRvz7/CpBdXVyIbm5T3u7sTNWlCNHEi0d69RK9eKf1qlJGZ83emk5q1a9eSg4MDLV26lC5fvkw9evQgT09PioiIMLj98OHDycfHh7Zt20Y3btyg+fPnk5OTE509e1a3zbNnz8jX15e6du1KJ06coJs3b9LOnTvp+vXrKfa3YcMGqlatGvn4+EhSY01u39b/V1+8qHQ0Qiju2jUiX1/+l/jefTpfKViQ6PlzZQI6e5ZjsLUlCgtTJgYr8Po10f79nKg0aULk5qZPYooXJ2rfnmjePKLz54kSE5WO1jyYNKnx9/envn376m6r1Wry8fGhyZMnG9ze29ub5s6dm+y+Vq1aUYcOHXS3R4wYQQ0aNEj32Pfu3aMiRYrQpUuXyNfXV5IaazJ0KP9XN26sdCRCKO70ac5fAKIyZYhuXYsjKl+e7xg4UJmgOnTg47drp8zxrVRiIn+Pu3tX6UjMV2bO35kqJ4qPj8eZM2fQuHFj3X02NjZo3Lgxjh07ZvA5cXFxcHpjFh9nZ2ccPnxYd3vLli2oVasWWrdujUKFCqFGjRpYvHhxsudoNBp06tQJw4YNQ6VKldKNNS4uDlFRUckuwky9egVo3++BA5WNRQiF7dsHvPcez2/31ltcX+FX1oGHOwE8ntvQwkamdO+efmbvXF4cbGy2tlx7XbSo0pFYh0wlNU+ePIFarYbXG4Pcvby8EB4ebvA5TZo0wYwZMxAaGgqNRoPdu3djw4YNeJhkRqCbN2/i119/RZkyZbBz50706dMHAwYMwIoVK3TbTJ06FXZ2dhgwYECGYp08eTI8PDx0l2KydpD5+v13nnChdGlelU+IXGrjRqBpU+DlS178et8+oFCh/z/4wQe8yJNazWO5c3KMxy+/8HCcgAAeryyEmTL5wK/Zs2ejTJkyKF++PBwcHNCvXz98+eWXsEky5kyj0eCtt97Cjz/+iBo1aqBnz57o0aMHFvx/1dczZ85g9uzZWL58OVQZnGVq1KhRiIyM1F3uGlqXXihPowFmz+brAwda31hEITJoyRLg8895/smWLYHt24EUAz1+/plHB+7fD/z1V84E9vIlsGgRX5dWGmHmMnUGKVCgAGxtbREREZHs/oiICBQuXNjgcwoWLIhNmzbh9evXCAsLw9WrV+Hq6oqSJUvqtvH29kbFihWTPa9ChQq4c+cOAODQoUN49OgRihcvDjs7O9jZ2SEsLAxDhgyBn5+fweM6OjrC3d092UWYoR07eLZUd3egSxeloxEixxEBU6cC3btzjv/VV7xQpcG1d/z8gJEj+fqQIbw0t6ktWcItqeXKAR99ZPrjCZENmUpqHBwcULNmTezZs0d3n0ajwZ49e1A3nUUHnZycUKRIESQmJmL9+vUIDAzUPVa/fn1cu3Yt2fYhISHw9fUFAHTq1AkXLlxAUFCQ7uLj44Nhw4Zh586dmXkJwtxoW2m6d+e5L4TIRTQaYNgwfZ4yciSXl6W5ltDw4Zzc3LsH/PijaQNMTARmzeLr33wjLanC/GW2Cnnt2rXk6OhIy5cvp+DgYOrZsyd5enpSeHg4ERF16tSJRo4cqdv++PHjtH79erpx4wYdPHiQGjVqRCVKlKDnSYYlnjx5kuzs7GjSpEkUGhpKq1atIhcXF/rjjz9SjUNGP1mBS5d4NIWNDdHNm0pHI0SOSkgg6tJFP5x3+vRMPHnjRn6SgwNRaKiJIiSiP//k4xQoQBQdbbrjCJGGzJy/M722aNu2bfH48WOMHTsW4eHhqF69Onbs2KErHr5z506yepnY2FiMHj0aN2/ehKurK5o3b46VK1fC09NTt03t2rWxceNGjBo1ChMmTECJEiUwa9YsdOjQIbs5W86IjQWePJHy9czSjub49FOeNzwXI+IW/rt3k1/Uav72LhO3WpeYGKBtW2DrVh79smRJJntfAwOBJk2AnTuBQYOAf/4xfpBEXMMDAF9/bfkLCIlcQZZJyK67d3lEQkwMz2vt6mq8fVuzp085CYyNBQ4cABo2VDoik4qOTpmwaC937vDPN6dE1woI4NIjgzUWwuK8eAF88glPpe/kxPUzLVpkYUfXrgFVqvCy0Vu38vIFxqRdtsTRkf9IdcOwhMhZmTl/Z7qlRrzBzo77tsPDucJv7dr0VyoTXDgQG8sLmbzzjtLRZEtCAnD/fsokJenl6dOM7StfPl78uFgxzvlWr+acr0MHPvnZ2pr2tQjTCg/nIdvnz3Nt/Nat2cjny5XjVppp0/hn48bGzXy1rTSdOklCIyyGtNQYw5EjPKlEYqKsXJsRCQnc3XT/PrBiBdC5s9IRpUqj4RNRWi0s4eEZmzLE1VWfsKR2yZMn+XP27+eTYFwc0LMnsGCB5MyW6uZN4MMPgRs3AC8vbn2rXj2bO335kpObhw+Nu7J9aCjvl4hXZKxQwTj7FSILZJVuA0y+Sve8eUC/fvxVevdunhJUGLZ2LdCuHX+yh4Vx87YCiIBnz1JvXbl7l/OuhIT09+XgwC0rxYunnrB4emYtIdmwAWjdmhOsMWOACRMyvw+hrAsXuAQmPJzz+d27gVKljLTzVauAjh0BFxfg6lX+Y8uuvn2B+fN5Msxt27K/PyGyQZIaA0ye1BABXbvy7LgFCwJnzhjnw8Ua1a3L9UfjxwPjxpnsMC9fpt66or3ExKS/HxsbwMeH387UkpaCBU072nXhQqB3b74+Zw7nz8IyHD7M5S6RkUDVqtxC4+1txAMQcR/W4cNAmzb65Qyy6ulT/qOOiQH27AEaNTJOnEJkkSQ1Bpg8qQH4Q6BePSAoCKhdGzh4UKo733T8OCc1Dg6cYbyx5EZWqdXA2bP8Dfi//4Bz57ggMyMKFUqeoLyZuHh7pzNvSA6ZMIFzQJWKG7vatFE6IpGef/7hVrbYWKBBA66hSTLw03jOn+eFojQaYO/e7LUUT5oEjB7NfWNnz0p/p1CcFAorxdmZ+wpq1QJOnQIGDNBPLy6YdrI9bfdTFhFxbcJ//3Eis3ev4STG0zNlq0rSpKVIEcvJO8eMAR494p7Ojh25qDjJ2rLCzPz+O9CtGyfcH33Ehd4uLiY6WLVq3JQ3fz6vC3XuHGBvn/n9xMXxgpkA1wZKQiMsjLTUmMKuXVzdScRJTY8epj2epbh3jwsKEhP5G2CNGpl6+pMn3BquTWTCwpI/7uHBX1A/+IAHVJUoYX0j7NVqzgf/+otf2/79sr6gOZo5Exg8mK937gz89lvWcoxMefYMKFuWu49mzcraivfLlnEmVqQIcOtWDgQtRPoydf420QSAZifHZxSeNEk/4+eJEzlzTHM3ahT/TgICMrR5dDTRrl1Ew4YR1aihn3lVe7G3511NnEh0/DjP0JobxMYSvf8+/w4KFiQKCVE6IqGl0ej/zAGib74hUqtzMICFC/nA7u5E/5/lPcM0GqJKlfj5U6eaJj4hsiAz529pqTEVjQb47DNg0yYeFnPmTO6e6yE6mvt7nj0DNm7kWYTfoFZzq7m2JebIEW4NT6pqVe5yadyYayPfHAKdW7x8ybMInD3LywAdPWrk4lORaWo10KcPT8EE8LJMI0fmcA+OWg34+/MfxpdfAkuXZvy5O3YAzZpxE+DduyYq/hEi86RQ2IAcT2r4oFwwHBLC/SK7dplHxakSFi0CevXiPqHQUN0scjdv6ot79+7lnCepokW5O6lxY+D9941WV2wVHj0C6tcHrl/nZO/AATkPKSUujidIXL+eR8EtWKBgr/OxYzxgAeDC/Dp1Mva8Dz7gf8SBA/WLWAphBiSpMUCRpAbgiavq1OE58IcO5dk/cxsioHJlIDgYryfOwL/lvtElMrduJd/U3V1fF9O4MZcISK1i6m7e5MQmPJxbrnbskCV6ctrLl9zwuHcvD+pbvZobaRXVtStPbFmrFnDiRPrzDZw/z6OdbGy4At/PLweCFCJjJKkxQLGkBuCvb59/ztf//DNXjcWNjQWCZ+/GWyM/xGsbVxTR3EMk9Ksz2tvzCG9tl1Lt2rm3MSurzp/nhCYqik+uf/0lv8Oc8vgx99icOcO9Nps3m8m0LuHhPCNwVBT3h3Xvnvb2XbrwcC1jzHMjhJFl5vxtwunChM5nnwEjRvD1bt2AS5eUjceENBruzp86lVtb8uYFHo6cBQBYovkSkfBA5crAN9/wRKXPnnG3yZgxnNzIyTjzqlUDtmzhiZk3beIFlXPHVxVlhYXx3DNnzgAFCgD79plJQgMAhQvz5JYAMGoU8Px56ts+eACsWcPXZYkXYeGkpSanJCbyMO89e4AyZYCTJ62mAOLWreTzxSRdvLEMQhCCctBAhX+mX4N/hzIoXFi5WK3Zxo3cIKjR8NxpEycqHZH1Cg7mdZzu3+d5j3bt4oYRs5KQwF1KwcE8BfWcOYa3GzUKmDKFM7RDh3I0RCEyQrqfDFA8qQG4rbpWLZ5Jt0UL/lptyrn1TeTZM05etInMzZvJH3dz47qYxo2BDsf7Id/qefx6t2xRJuBcZPFiXvgSAH75hedhE8Z1/DhPpvfsGVCxIrBzJxe0m6U9e/gf0caGhxZWrZr88VeveFTiixepjkoUQmmS1BhgFkkNAJw+zd+I4uL4q/To0crFkkGxsTxkWFvce+ZM8u4NOzvg7bf1xb21a/9/zq7nz/nTPjqan/j++4q9htzkhx+4O0+l4l6Ftm2Vjsh67NwJtGrFf9J16nAXav78SkeVjtatgb//5sKr/fuTV97PmcMzn5cuzYth/n9UohDmRJIaA8wmqQH0s3aqVPyp2KyZsvG8QaPh4lNtS8yhQ5zYJFWpEicwH3zAn5VubgZ2NH06MGwYj3y6cEGGMeUQIj5PzZ3LyeW2bfw+iawj4lawfv24V+fDD7n+3yJmrL5zByhfntemW72ap6QGeE6bsmW5qXXePC7GEsIMSVJjgFklNQDP0rVgAdfVnD4NlCqlaDhE3Du0Zg23WD95kvxxb+/k88X4+KSzw8RE/vYXFsZzxH/1lcliFymp1UD79rzeUJ48/AW9Vi2lo7JMUVE8xdLatXy7bVseKOTgkPOxqNVqJCQkZP6Jv/7K664VKgRs385/FLt2cfbr6clVzjIXgFCQvb09bFNpKZSkxgCzS2ri4nhK2OPHuZ/72DETrnaXtrNneTTSwYP6+1xdOTxtIlOhQiYbWrTD2AsU4G+K8oGZ4+LiuPZjzx5+G44c4S/mIuPOneNRztevc8/MpEnc+JjTpXBEhPDwcLzI6NLzKXfAo5wSE3kyqLx5edh3XBwvmmYlgxaEZfP09EThwoWheuNkI6t0WwJHR+7nfust7prp0QP4448c7aJ5+BD47jtg+XL+zHN2Bvr25VpBf/9srmWnnZG0Vy9JaBTi6Mi1n++9x3VQH37ItVHptrIJEHFD6jff8Hm/aFFuqalfX5l4tAlNoUKF4OLikuJDP0O0XzBUKk5iXr/mBKdsWVm4UiiKiBAdHY1Hjx4BALyzs+ZLVhaXmjt3Lvn6+pKjoyP5+/vTiTQWbIyPj6fvv/+eSpYsSY6OjlS1alXavn17iu3u3btHHTp0oHz58pGTkxNVrlyZTp06pdvH8OHDqXLlyuTi4kLe3t7UqVMnun//foZjzvEFLTNq/34iW1teRG727Bw5ZHQ0r7eZJ49+4b0OHYju3DHSAU6f5p3a2RFl4j0SphERQVSmDL8lVaoQPX+udETm7cULotat9f8bH39M9OSJcvEkJiZScHAwPcluEBoNr3566pT+cuuWUWIUwhiePHlCwcHBlJiYmOz+zJy/M92I+ueff2Lw4MEYN24czp49i2rVqqFJkya6DOtNo0ePxsKFCzFnzhwEBwejd+/eaNmyJc6dO6fb5vnz56hfvz7s7e2xfft2BAcH4+eff0bevHkBANHR0Th79izGjBmDs2fPYsOGDbh27Ro++eSTLCVyZiUggAtqAZ74yoTzRBDxZKHly3MLzevXPILj2DFuJCpWzEgHmj2bf7ZtK80CZqBQIR61U7gwcPEi8MknXDMqUjpzhhtPtbMy//wz15opOcJJW0Pjkt3uaZWK/8mTtvLIYmrCjGj/xrNUN6aV2UzK39+f+vbtq7utVqvJx8eHJk+ebHB7b29vmjt3brL7WrVqRR06dNDdHjFiBDVo0CBTcZw8eZIAUFhYWIa2N9uWGiL+BtWuHX8t9PIySevGyZNE9evrv30WK0a0ahUf2qgePCCyt+eDnDxp5J2L7Dh/nsjDg9+awECihASlIzIfGg3RL78QOTjw78fXl+j4caWjYjExMRQcHEwxMTHG2eHdu9xKExJinP0JYSSp/a2brKUmPj4eZ86cQePGjXX32djYoHHjxjh27JjB58TFxcHJySnZfc7Ozjh8+LDu9pYtW1CrVi20bt0ahQoVQo0aNbB48eI0Y4mMjIRKpYJnKgVucXFxiIqKSnYxWyoVjxetUgWIiOAC2/h4o+z6/n2gc2eukTlyhGuRJ0zgKSnatzdBCc+CBTzmtV49nrBGmI2qVfXLKWzeDPTuLcspADyd0mef8UCg+HiuKTt3LuOLW1ucIkWAkiVl0UphnTKTRd2/f58A0NGjR5PdP2zYMPL39zf4nHbt2lHFihUpJCSE1Go17dq1i5ydncnBwUG3jaOjIzk6OtKoUaPo7NmztHDhQnJycqLly5cb3GdMTAy99dZb1L59+1RjHTduHAFIcTHLlhqt69eJPD35q+LXX2drV69fE33/PZGLi751pnNnonv3jBSrITExRAUL8sHWrTPhgUR2bNxIZGPDb9O33yodjbJOnCDy8+Pfhb09l7UZvfUym4zeUmPBfH19aebMmRneft++fQSAnkshmUUwRkuNyZOaR48eUWBgINnY2JCtrS2VLVuWvv76a3JyctJtY29vT3Xr1k32vP79+9Pbb7+dYn/x8fHUokULqlGjRpovMDY2liIjI3WXu3fvmn9SQ0S0bRuRSsWfssuWZfrpajXRH38QFS2qT2bq18+hnqClS/V9W9K3YdYWL9b/feRQfbpZ0WiIfv6Za9kBohIlzLe31BKTGkNfKJNexo0bl6X9Pnr0iF6/fp3h7ePi4ujhw4ekycFMtVy5cuTg4EAPHz7MsWNaixzvfipQoABsbW0RERGR7P6IiAgUTmWVwoIFC2LTpk14/fo1wsLCcPXqVbi6uqJkyZK6bby9vVGxYsVkz6tQoQLu3LmT7L6EhAS0adMGYWFh2L17d5rj1R0dHeHu7p7sYhGaNwfGjePrvXvzJDIZdPw49/p07Ajcuwf4+nJh8KFDOdATRKQfxt2vnyy3bea6d+flFABg0CD9xHK5wbNnQGAg1+UnJnLX09mz0ltqTA8fPtRdZs2aBXd392T3DR06VLctESExMTFD+y1YsGCmCqYdHBwMzntiKocPH0ZMTAw+//xzrFixIkeOmZZsFdxaqEwlNQ4ODqhZsyb27Nmju0+j0WDPnj2oW7dums91cnJCkSJFkJiYiPXr1yMwMFD3WP369XHt2rVk24eEhMDX11d3W5vQhIaG4r///kN+s19wJRvGjAE+/pgnyGjVKvmy1wbcvQt06ADUrQucOMET5/34I3DlCk8cliP/zwcO8Hw7Li58xhRm79tvecFLIq672r1b6YhM79gxXrh661aeEXjePB7pJHPPGVfhwoV1Fw8PD6hUKt3tq1evws3NDdu3b0fNmjXh6OiIw4cP48aNGwgMDISXlxdcXV1Ru3Zt/Pfff8n26+fnh1naL08AVCoVfvvtN7Rs2RIuLi4oU6YMtiRZOHf//v1QqVS6SQuXL18OT09P7Ny5ExUqVICrqyuaNm2Khw8f6p6TmJiIAQMGwNPTE/nz58eIESPQpUsXfJqBxT6XLFmC9u3bo1OnTli6dGmKx+/du4d27dohX758yJMnD2rVqoUTJ07oHt+6dStq164NJycnFChQAC1btkz2Wjdt2pRsf56enli+fDkA4Pbt21CpVPjzzz8REBAAJycnrFq1Ck+fPkW7du1QpEgRuLi4oEqVKlizZk2y/Wg0Gvz0008oXbo0HB0dUbx4cUyaNAkA0KhRI/Tr1y/Z9o8fP4aDg0OyXMBsZLZ5aO3ateTo6EjLly+n4OBg6tmzJ3l6elJ4eDgREXXq1IlGjhyp2/748eO0fv16unHjBh08eJAaNWpEJUqUSNbHefLkSbKzs6NJkyZRaGgorVq1ilxcXOiPP/4gIu5y+uSTT6ho0aIUFBREDx8+1F3i4uIyFLdZj34y5PlzolKluG38gw+I3hi3T0T08iXRmDFEzs68mUpF1K0bD0DKcYGBHESfPgocXGSVWk30xRf81uXJY75dMNmlVhP99JO+u6l0aaKzZ5WOKmPebJLXaIhevVLmkpVenGXLlpGHh4futrbOpWrVqrRr1y66fv06PX36lIKCgmjBggV08eJFCgkJodGjR5OTk1OyEa5v1tQAoKJFi9Lq1aspNDSUBgwYQK6urvT06dNkx9Keb5YtW0b29vbUuHFjOnXqFJ05c4YqVKiQrD7zhx9+oHz58tGGDRvoypUr1Lt3b3J3d6fAwMA0X2dUVBTlyZOHLl26RImJieTl5UUHDx7UPf7y5UsqWbIkvfPOO3To0CEKDQ2lP//8U1fO8c8//5CtrS2NHTuWgoODKSgoiH788cdkr3Xjxo3Jjunh4UHL/l+mcOvWLQJAfn5+tH79erp58yY9ePCA7t27R9OmTaNz587RjRs36JdffiFbW9tk88sNHz6c8ubNS8uXL6fr16/ToUOHaPHixUREtGrVKsqbNy/Fxsbqtp8xYwb5+fkZvVsvx2tqtObMmUPFixcnBwcH8vf3p+NJxj4GBARQly5ddLf3799PFSpUIEdHR8qfP3+qk+Zt3bqVKleuTI6OjlS+fHlatGiR7jHtm2Xosm/fvgzFbHFJDRHRhQv6St9Ro3R3q9VEK1YQ+fjo6yIaNiQ6c0ahOK9f19cBXbmiUBAiq+LiiBo35revQAGiq1eVjsi4Hj8mat5c/7/yxRdElvQx8OYH/atX+teS05dXrzIff2pJzaZNm9J9bqVKlWjOnDm624aSmtGjR+tuv3r1igDoJng1lNQAoOvXr+ueM2/ePPLy8tLd9vLyomnTpuluJyYmUvHixdNNahYtWkTVq1fX3R44cGCyc+HChQvJzc1Nl3C9qW7dusmmOnlTRpOaWbNmpRknEdFHH31EQ4YMISJOxhwdHXVJzJtiYmIob9689Oeff+ruq1q1Ko0fPz7d42SWMZKaLBU+9OvXL0VzlNb+/fuT3Q4ICEBwcHC6+/z444/x8ccfG3zMz88PlBvHnlapAixZwqvqTp4M1KqFI16tMGgQr4EJACVK8Nx9LVsquAj23Ln8mde0Kc/sJyyKgwOwYQPQqBH/XTVpYj3LKRw6xP8+9+/zUPZffuEVSWTBeOXVemOF1VevXmH8+PHYtm0bHj58iMTERMTExKSorXxT1apVddfz5MkDd3f3VCeDBXiCt1JJFhD29vbWbR8ZGYmIiAj4+/vrHre1tUXNmjWh0WjSjGPp0qXo2LGj7nbHjh0REBCAOXPmwM3NDUFBQahRowby5ctn8PlBQUHo0aNHmsfIiDd/r2q1Gj/++CPWrVuH+/fvIz4+HnFxcbrapCtXriAuLg7vv/++wf05OTnputPatGmDs2fP4tKlS8m6+cyJVHOauy++AE6dAmbMQEzbLvgqsSKuoTzc3IDRo3lujTemAcpZUVGceAFccSoskpsb8O+/vLZRaCgnNgcP8rqHlkijAaZMAcaO5RXLy5bl2pkk5z+L5eICvHql3LGNJU+ePMluDx06FLt378b06dNRunRpODs74/PPP0d8OnN22b+xbpVKpUozATG0fXa/NAcHB+P48eM4efIkRowYobtfrVZj7dq16NGjB5zTWQMvvccNxWmoEPjN3+u0adMwe/ZszJo1C1WqVEGePHkwaNAg3e81veMCQPfu3VG9enXcu3cPy5YtQ6NGjZLVvJqTHF5rVmTWy5fAGIepOKB6F86Jr7ARLTGgaxRCQ4HhwxVOaABg2TIOsnx5XjFRWKyCBYFduwBvb+DSJctdTuHRI6BZM14KRK3m0YBnzlhHQgNwK1OePMpcTNnCdeTIEXTt2hUtW7ZElSpVULhwYdy+fdt0BzTAw8MDXl5eOHXqlO4+tVqNs+mMQl2yZAkaNmyI8+fPIygoSHcZPHgwlvz/S1/VqlURFBSEZ8+eGdxH1apV0yy8LViwYLKC5tDQUERHR6f7mo4cOYLAwEB07NgR1apVQ8mSJRESEqJ7vEyZMnB2dk7z2FWqVEGtWrWwePFirF69Gt26dUv3uEqRpMZMqdXA0qX8DfOHKXZoTX/isWMRVMBVzI7sCq9CZtAdp1Zzez4ADBwobfpWwM8P2LED8PAADh/mhsIMjrY1C/v38+imXbt4cfglS4Dff+cRgcK8lSlTBhs2bEBQUBDOnz+P9u3bp9vlYwr9+/fH5MmTsXnzZly7dg0DBw7E8+fPUx0WnpCQgJUrV6Jdu3aoXLlyskv37t1x4sQJXL58Ge3atUPhwoXx6aef4siRI7h58ybWr1+vm41/3LhxWLNmDcaNG4crV67g4sWLmDp1qu44jRo1wty5c3Hu3DmcPn0avXv3TtHqZEiZMmWwe/duHD16FFeuXEGvXr2STcvi5OSEESNGYPjw4fj9999x48YNHD9+XJeMaXXv3h1TpkwB/a+9e4+LOfv/AP6amXRXLtGFLkMhbYrkkq/Lym4Um7WUS0S5LOu6a8Uq7Ppallw2u2t3/VKy7ktp+dJaYpO0rArL6mLEqta1RqQ0nd8fh9HUlKKazLyfj8c8mPl85vN5N6dp3nM+73MOYwqjshobSmoaoZMn+ZwZgYFAXh5gawtsjmkNkxP7eAFEdDRQ7pddZQ4dAq5d49coxo9XdTSkjnTpwoc86+ryZRWmTWv8yynIZHz5D3d3IDcXsLfnV20DAijXflOsW7cOzZs3h5ubG4YNGwYPDw9069atweMICgrCmDFjMGHCBPTu3RuGhobw8PCotNzPc7Gxsbh3757SD3p7e3vY29sjPDwc2tra+PXXX9G6dWt4enrC0dERq1atgkgkAgAMGDAAe/fuRWxsLJydnTFw4ED88ccf8mOtXbsWlpaW6Nu3L8aOHYv58+fXaM6e4OBgdOvWDR4eHhgwYIA8sSovJCQEn3zyCZYsWQJ7e3v4+vpWqksaM2YMtLS0MGbMmCpfi0ahzsuXG6k3YfRTVhZjI0a8GGlgbMxnPVUYtf7DD3yjUMjYr7+qKlTu7bd5LAsWqDYOUi8OHHixnEK5wXeNTm4uY+7uL943Eye+2iidxupNnFFYnchkMtahQweFUVaaSCKRMKFQyP6sx2G2DT6jMKkfUikQFMS/Xe7fDwiFwPTpvGDz449554zclCm8C6esjA/raOBrznIXLgDx8YBIBHz0kWpiIPXqvfeAH3/k/1+5Evj6a9XGo8yxY/xy07FjvIh161Ze5lWhVpKQGsvOzsbmzZuRnp6OixcvYvr06ZBIJBg7dqyqQ1OJp0+fIi8vD8HBwejVq5dKes9qg5IaFZLJ+IeGrS2wejVfIfidd4C0NOC773jhZiUCAR9C3b07n2n4gw9UU835/BPugw8AK6uGPz9pEIGBwLOJRTF3LrBjh0rDkZPJ+Goi77zDF7Z/6y0+HH3CBFVHRt50QqEQkZGRcHV1RZ8+fXDx4kX89ttvsLe3V3VoKpGYmAhzc3OcPXsW33//varDeSkBY439anndkEqlMDY2RkFBQaNYB+r4cWDePN7hAfCC4HXr+NJPNaoBuHEDcHEB7t4F/P3519OGKh64fZsnMsXFQGIiX3CKqC3GeEITFsaX9Dp4kA/5VpWcHGDsWF57BvBVOb7+um6HGzcmT548gUQigVgsbty1DIS8pqp+12vz+U09NQ0sIwMYPpwXNF64wNec2bCBD6H18qpFXmJlxVerFAp5n3tDZtA//MATGldXvuAUUWsCAbB+Pb/a+XwByHL1iw0qLo5fbjp5ko9o2r4d2LxZfRMaQkjtUFLTQPLz+arADg7AgQO8FGXWLCAzk4+GrsHIvMoGDuQzjAH8IKdP12XIypWU8GtjAP/6TkNLNIJQCERG8qmIHj3iPYoV1qCtV6WlfAHOwYOBO3cAJyc+94yGljkQQqpASU09Ky0FNm0C7Oz45aWnT/nEYBcv8u78115sfP58YORIfuCRI/kY8Pq0Zw8/h4UFPx/RGNrawL59vIPu3j2e4Ny6Vf/n/ecf4O23ebEyAHz4IXDmDL9kSwgh5VFSU49+/ZV3lc+YwUtf7O2Bw4f5dPR1VnMmEPBZ+jp35hN0+PjwBKc+MMavQwD8h1IYlkU0gaEhn56oQwde1jV4MPDgQf2d73//4++hU6f4Ug67d/MvCVRaQghRhpKaevD338DQobyY8q+/gBYt+ICltDT+IVDnmjblY8GNjPgKfp9+Wg8nAS8KPn+ef6JMnVo/5yCN3vPlFCwseC3YsGFADWZrr5WnT/kyIF5evFeoWzf+q+fjU7fnIYSoF0pq6tD9+7y0xdGRf5vV0uJlJ5mZfCqXV6qbqamOHfl88AAfCrJ9e92f4/kwbj+/KsabE01hbc2XU2jWjOe6vr51t5zCjRtA//7AmjX8/syZvFzM1rZujk8IUV+U1NSBp0+BjRv5H92wMP7Hfdgw/i12/foGXOnY25sv3Q3wSfrS0uru2NnZvDcI4EuDE43n6PhiOYWDB3nn3etOEBEbyy83JSXx9ad+/pm/t3R06iRk8oYZMGAA5s6dK79vY2ODDRs2VPscgUCAmJiY1z53XR2HNCxKal7T1at8rZzZs3ltwVtv8a752FjeedLgli3j172KioD33+fdR3Xhm2/4LMbu7vzTjBAA//kPr3MRifhUSZ999mrHKSnhs2d7e/P3kasrkJLCh4+TN8+wYcMwuIpr7QkJCRAIBLjwfJKuWjh79iym1vGl72XLlsHZ2bnS47m5uRgyZEidnqsqRUVFaNGiBUxMTFBcXNwg51RXlNS8JktLoLAQMDHhBYwpKXyWU5URifi0r2IxIJHwS0Wvu9JtYSHwf//H/1/uWxMhgOJyCqtW8XmXakMiAfr2fVGDPncuLwwWi+syStKQAgMDcfToUfzzzz+VtkVERKB79+7o0qVLrY/bqlWrGi3iWBfMzMyg00BdhPv27YODgwM6deqk8t4hxhhK6+pasgpQUvOa9PX5otkZGXyoqZaWqiMCr0zev59fFzh8GPj889c7XlQUn2jH1pZPUEJIBQEBL4Zcz5tX85Ku6Giga1c+mV+zZkBMDE9uaGDdm23o0KFo1aoVIiMjFR4vLCzE3r17ERgYiHv37mHMmDFo06YN9PX14ejoiJ07d1Z73IqXnzIyMtCvXz/o6uqic+fOOHr0aKXnBAUFoUOHDtDX10e7du0QEhKCp89GiEZGRuLzzz9HWloaBAIBBAKBPOaKl58uXryIgQMHQk9PDy1btsTUqVNRWFgo3z5x4kQMHz4coaGhMDc3R8uWLfHRRx/Jz1Wd8PBw+Pn5wc/PD+Hh4ZW2//XXXxg6dCiMjIzQtGlT9O3bF1lZWfLtW7ZsgYODA3R0dGBubo6ZM2cCAK5fvw6BQIDU1FT5vvn5+RAIBDhx4gQA4MSJExAIBDh8+DBcXFygo6ODU6dOISsrC97e3jA1NYWhoSFcXV3x22+/KcRVXFyMoKAgWFpaQkdHB7a2tggPDwdjDLa2tggNDVXYPzU1FQKBAJmZmS99TV5VY/gIfuN1767qCJRwduZTrY4fD3zxBQ9y2LDaH6es7EWB8OzZfBY2QpQICuLrMG3YAEycyOdgqmq0X3ExH6S3cSO/36sXsGsXL0AmL8FY3Q83qyl9/RpNuKmlpYUJEyYgMjISixcvhuDZc/bu3QuZTIYxY8agsLAQLi4uCAoKgpGREQ4dOoTx48ejffv26NGjx0vPUVZWhhEjRsDU1BTJyckoKChQqL95rmnTpoiMjISFhQUuXryIKVOmoGnTpliwYAF8fX1x6dIlHDlyRP6BbWxsXOkYjx49goeHB3r37o2zZ8/i9u3bmDx5MmbOnKmQuMXHx8Pc3Bzx8fHIzMyEr68vnJ2dMWXKlCp/jqysLCQlJWH//v1gjGHevHnIzs6G9bM3w61bt9CvXz8MGDAAx48fh5GRERITE+W9KZs2bcLHH3+MVatWYciQISgoKEBiYuJLX7+KFi5ciNDQULRr1w7NmzfHzZs34enpiRUrVkBHRwdRUVEYNmwYrl69Cqtn6/1NmDABSUlJCAsLg5OTEyQSCe7evQuBQICAgABERERg/vz58nNERESgX79+sK3Pqv96WD28UarN0uVqZdYsxgDGjIwYS0+v/fP/978Xz5dK6z4+olZkMsbGjuW/Mvr6jJ05U3mfzEzGXFz4PgBj8+czVlLS8LG+KYqKitjly5dZUVERf6Cw8MWL19C3wsIax33lyhUGgMXHx8sf69u3L/Pz86vyOV5eXuyTTz6R3+/fvz+bM2eO/L61tTVbv349Y4yxuLg4pqWlxW7duiXffvjwYQaARUdHV3mONWvWMBcXF/n9pUuXMicnp0r7lT/Ojz/+yJo3b84Ky/38hw4dYkKhkOXl5THGGPP392fW1tastLRUvs+oUaOYr69vlbEwxthnn33Ghg8fLr/v7e3Nli5dKr+/aNEiJhaLWUkVbxILCwu2ePFipdskEgkDwFJSUuSPPXjwQKFd4uPjGQAWExNTbZyMMebg4MA2btzIGGPs6tWrDAA7evSo0n1v3brFRCIRS05OZowxVlJSwkxMTFhkZGSVx6/0u/5MbT6/X+lr97fffgsbGxvo6uqiZ8+e+KOahWCePn2KL774Au3bt4euri6cnJxw5MiRSvvdunULfn5+aNmyJfT09ODo6Ihz586VT76wZMkSmJubQ09PD4MGDUJGRsarhK9ZQkN5NadUyguHy3WX1sjzrt7Jk/l8OIRUQyjkBcMeHrwzwcuLz9v03N69fM6ZP//kPTkHD/Kh2/U63QFRiU6dOsHNzQ1btmwBAGRmZiIhIQGBgYEAAJlMhuXLl8PR0REtWrSAoaEh4uLicOPGjRod/8qVK7C0tISFhYX8sd5K1qLbvXs3+vTpAzMzMxgaGiI4OLjG5yh/LicnJxgYGMgf69OnD8rKynC13HohDg4OEIlE8vvm5ua4fft2lceVyWTYunUr/Pz85I/5+fkhMjISZc9qIVNTU9G3b180UfImuX37NnJycuDu7l6rn0eZ7hUuORQWFmL+/Pmwt7dHs2bNYGhoiCtXrshfu9TUVIhEIvTv31/p8SwsLODl5SVv/19++QXFxcUYNWrUa8danVonNbt378bHH3+MpUuX4vz583BycoKHh0eVDRccHIwffvgBGzduxOXLl/Hhhx/i/fffR0pKinyfBw8eoE+fPmjSpAkOHz6My5cvY+3atWhebiz06tWrERYWhu+//x7JyckwMDCAh4cHnjx58go/tgbR1uZLG5ib85kAAwNrPu72r7/4UC6hkE8WQkgNaGvzodg9erxYTiEri09C7ePD8+s+fYDUVJ70kFrS1+dfTlRxq2WRbmBgIPbt24eHDx8iIiIC7du3l38IrlmzBl9//TWCgoIQHx+P1NRUeHh4oKSkpM5eqqSkJIwbNw6enp44ePAgUlJSsHjx4jo9R3kVEw+BQCBPTpSJi4vDrVu34OvrCy0tLWhpaWH06NHIzs7GsWPHAAB6enpVPr+6bQAgfFYuwMr9za+qxqd8wgYA8+fPR3R0NL788kskJCQgNTUVjo6O8tfuZecGgMmTJ2PXrl0oKipCREQEfH19673Qu9ZJzbp16zBlyhRMmjQJnTt3xvfffw99fX15NlbRtm3b8Nlnn8HT0xPt2rXD9OnT4enpibVr18r3+eqrr2BpaYmIiAj06NEDYrEY7777Ltq3bw+AN8iGDRsQHBwMb29vdOnSBVFRUcjJyVF5pfgbwdycf8poafEEZ926mj0vLIz/6+1NQ1FIrTxfTqFjR+DmTf7vpk1826JFwIkTQNu2Kg3xzSUQAAYGqrnVcgFbHx8fCIVC7NixA1FRUQgICJDX1yQmJsLb2xt+fn5wcnJCu3btkJ6eXuNj29vb4+bNm8jNzZU/dubMGYV9Tp8+DWtrayxevBjdu3eHnZ0dsrOzFfbR1taGTCZ76bnS0tLw6NEj+WOJiYkQCoXo+Bpzd4SHh2P06NFITU1VuI0ePVpeMNylSxckJCQoTUaaNm0KGxsbeQJUUatnk6SWf43KFw1XJzExERMnTsT7778PR0dHmJmZ4fr16/Ltjo6OKCsrw8mTJ6s8hqenJwwMDLBp0yYcOXIEAQEBNTr366hVUlNSUoI///wTgwYNenEAoRCDBg1CUlKS0ucUFxdDt8JCLXp6ejh16pT8fmxsLLp3745Ro0ahdevW6Nq1KzZv3izfLpFIkJeXp3BeY2Nj9OzZs9rzSqVShZtGc3N7cSlpwQIgPr76/e/dA7Zt4/+nYdzkFZiYAHFxQJs2gEzG7x85Anz5ZSMZJUjqnaGhIXx9fbFo0SLk5uZi4sSJ8m12dnY4evQoTp8+jStXrmDatGn4999/a3zsQYMGoUOHDvD390daWhoSEhKwePFihX3s7Oxw48YN7Nq1C1lZWQgLC0N0dLTCPjY2NpBIJEhNTcXdu3eVzhMzbtw46Orqwt/fH5cuXUJ8fDxmzZqF8ePHw9TUtHYvyjN37tzBL7/8An9/f7z11lsKtwkTJiAmJgb379/HzJkzIZVKMXr0aJw7dw4ZGRnYtm2b/LLXsmXLsHbtWoSFhSEjIwPnz5/HxmcV+Hp6eujVqxdWrVqFK1eu4OTJkwh+PkHrS9jZ2WH//v1ITU1FWloaxo4dq9DrZGNjA39/fwQEBCAmJgYSiQQnTpzAnj175PuIRCJMnDgRixYtgp2dndLLg3WtVknN3bt3IZPJKjWiqakp8qpYHdrDwwPr1q1DRkYGysrKcPToUezfv18hc7x27Ro2bdoEOzs7xMXFYfr06Zg9eza2bt0KAPJj1+a8K1euhLGxsfxmaWlZmx9VPc2YAUyYwEc0+fryr9BV2byZT+Dn7MwnESHkFVhbA7//zhOZ1FRea0M0S2BgIB48eAAPDw+F+pfg4GB069YNHh4eGDBgAMzMzDB8+PAaH1coFCI6OhpFRUXo0aMHJk+ejBUrVijs895772HevHmYOXMmnJ2dcfr0aYSEhCjs88EHH2Dw4MF4++230apVK6XDyvX19REXF4f79+/D1dUVI0eOhLu7O7755pvavRjlREVFwcDAQGk9jLu7O/T09PDTTz+hZcuWOH78OAoLC9G/f3+4uLhg8+bN8ktd/v7+2LBhA7777js4ODhg6NChCvWmW7ZsQWlpKVxcXDB37lz897//rVF869atQ/PmzeHm5oZhw4bBw8MD3bp1U9hn06ZNGDlyJGbMmIFOnTphypQpCr1ZAG//kpISTJo0qbYv0at5aSlxhWpmAOz06dMKj3/66aesR48eSp9z+/Zt5u3tzYRCIROJRKxDhw5sxowZTFdXV75PkyZNWO/evRWeN2vWLNarVy/GGGOJiYkMAMvJyVHYZ9SoUczHx0fpeZ88ecIKCgrkt5s3b2rm6KeKHj9mzNmZj2RwdWWsQpU5Y4wPRWnThu9TTaU6IaT+VTUihJA3we+//86aNGkiHyVWnQYf/WRiYgKRSFSpi/Dff/+FmZmZ0ue0atUKMTExePToEbKzs/H333/D0NAQ7dq1k+9jbm6Ozp07KzzP3t5eXmX9/Ni1Oa+Ojg6MjIwUbgSAnh6fmK9FC+DsWeXrOO3fD9y6BbRuDYwe3fAxEkIIeaMVFxfjn3/+wbJlyzBq1KhXvkxXW7VKarS1teHi4qJQlFRWVoZjx4699FqZrq4u2rRpg9LSUuzbtw/e3t7ybX369FEYFgcA6enp8smHxGIxzMzMFM4rlUqRnJzcINfo1I5YDOzcyYv+Nm/mt/Ke195Mn04rCRJCCKm1nTt3wtraGvn5+Vi9enXDnbi2XUm7du1iOjo6LDIykl2+fJlNnTqVNWvWTN61NH78eLZw4UL5/mfOnGH79u1jWVlZ7Pfff2cDBw5kYrGYPXjwQL7PH3/8wbS0tNiKFStYRkYG2759O9PX12c//fSTfJ9Vq1axZs2asQMHDrALFy4wb29vJhaLa9wlq7GT71Xnyy/5JSZtbcaeTZDEkpJePJabq9r4CCF0+YlojLq4/FTrMQi+vr64c+cOlixZgry8PDg7O+PIkSPyrqUbN27Ix8YDwJMnTxAcHIxr167B0NAQnp6e2LZtG5o1aybfx9XVFdHR0Vi0aBG++OILiMVibNiwAePGjZPvs2DBAjx69AhTp05Ffn4+/vOf/+DIkSOVRlaRWli4kC+6ExPDl0P+888XSyKMGQNUcWmPEEIIaYwEjNV0JrY3m1QqhbGxMQoKCqi+pjypFHB1BdLTgZ49eWJTWgqcP89XGiSEqNSTJ08gkUggFovpSxxRa1X9rtfm85tWJ9R0RkZ8qWRDQyA5mSc0/fpRQkNII1PdzLSEqIO6+B2nKbAI0LkzEBkJjBzJ79Nke4Q0Gtra2hAKhcjJyUGrVq2gra0tn5WXEHXAGENJSQnu3LkDoVAIbW3tVz4WJTWE++ADIDwcuH4deO89VUdDCHlGKBRCLBYjNzcXOTk5qg6HkHqjr68PKysrhbrc2qKkhrzQAOtyEEJqT1tbG1ZWVigtLX3pOkWEvIlEIhG0tLReuxeSkhpCCHkDCAQCNGnSpNJK0ISQF6hQmBBCCCFqgZIaQgghhKgFSmoIIYQQohY0pqbm+RyDUqlUxZEQQgghpKaef27XZK5gjUlqHj58CACwtLRUcSSEEEIIqa2HDx/C2Ni42n00ZpmEsrIy5OTkoGnTpjRxVRWkUiksLS1x8+ZNWkqiEaD2aHyoTRoXao/Gpb7agzGGhw8fwsLC4qVz2GhMT41QKETbtm1VHcYbwcjIiP5ANCLUHo0PtUnjQu3RuNRHe7ysh+Y5KhQmhBBCiFqgpIYQQgghaoGSGiKno6ODpUuXQkdHR9WhEFB7NEbUJo0LtUfj0hjaQ2MKhQkhhBCi3qinhhBCCCFqgZIaQgghhKgFSmoIIYQQohYoqSGEEEKIWqCkRsPY2NhAIBBUun300Ue4f/8+Zs2ahY4dO0JPTw9WVlaYPXs2CgoKVB22WquuTcpjjGHIkCEQCASIiYlRTbAaoCbtkZSUhIEDB8LAwABGRkbo168fioqKVBi1+npZe+Tl5WH8+PEwMzODgYEBunXrhn379qk4avUlk8kQEhICsVgMPT09tG/fHsuXL1dYl4kxhiVLlsDc3Bx6enoYNGgQMjIyGiQ+jZlRmHBnz56FTCaT37906RLeeecdjBo1Cjk5OcjJyUFoaCg6d+6M7OxsfPjhh8jJycHPP/+swqjVW3VtUt6GDRtoiY8G8LL2SEpKwuDBg7Fo0SJs3LgRWlpaSEtLe+n07eTVvKw9JkyYgPz8fMTGxsLExAQ7duyAj48Pzp07h65du6oqbLX11VdfYdOmTdi6dSscHBxw7tw5TJo0CcbGxpg9ezYAYPXq1QgLC8PWrVshFosREhICDw8PXL58Gbq6uvUbICMabc6cOax9+/asrKxM6fY9e/YwbW1t9vTp0waOTHMpa5OUlBTWpk0blpubywCw6Oho1QWoYSq2R8+ePVlwcLCKo9JcFdvDwMCARUVFKezTokULtnnzZlWEp/a8vLxYQECAwmMjRoxg48aNY4wxVlZWxszMzNiaNWvk2/Pz85mOjg7buXNnvcdHXy00WElJCX766ScEBARU2QNQUFAAIyMjaGlRp15DUNYmjx8/xtixY/Htt9/CzMxMxRFqlortcfv2bSQnJ6N169Zwc3ODqakp+vfvj1OnTqk6VI2g7P3h5uaG3bt34/79+ygrK8OuXbvw5MkTDBgwQLXBqik3NzccO3YM6enpAIC0tDScOnUKQ4YMAQBIJBLk5eVh0KBB8ucYGxujZ8+eSEpKqvf46JNKg8XExCA/Px8TJ05Uuv3u3btYvnw5pk6d2rCBaTBlbTJv3jy4ubnB29tbdYFpqIrtce3aNQDAsmXLEBoaCmdnZ0RFRcHd3R2XLl2CnZ2dCqNVf8reH3v27IGvry9atmwJLS0t6OvrIzo6Gra2tqoLVI0tXLgQUqkUnTp1gkgkgkwmw4oVKzBu3DgAvMYJAExNTRWeZ2pqKt9Wnyip0WDh4eEYMmQILCwsKm2TSqXw8vJC586dsWzZsoYPTkNVbJPY2FgcP34cKSkpKo5MM1Vsj7KyMgDAtGnTMGnSJABA165dcezYMWzZsgUrV65UWayaQNnfrJCQEOTn5+O3336DiYkJYmJi4OPjg4SEBDg6OqowWvW0Z88ebN++HTt27ICDgwNSU1Mxd+5cWFhYwN/fX9XhUU2Nprp+/ToTCoUsJiam0japVMp69+7N3N3dWVFRkQqi00zK2mTOnDlMIBAwkUgkvwFgQqGQ9e/fX3XBagBl7XHt2jUGgG3btk1hXx8fHzZ27NiGDlGjKGuPzMxMBoBdunRJYV93d3c2bdq0hg5RI7Rt25Z98803Co8tX76cdezYkTHGWFZWFgPAUlJSFPbp168fmz17dr3HRzU1GioiIgKtW7eGl5eXwuNSqRTvvvsutLW1ERsbW/+V6kROWZssXLgQFy5cQGpqqvwGAOvXr0dERISKItUMytrDxsYGFhYWuHr1qsK+6enpsLa2bugQNYqy9nj8+DEAVBp5JhKJ5L1qpG49fvy42tdbLBbDzMwMx44dk2+XSqVITk5G79696z/Aek+bSKMjk8mYlZUVCwoKUni8oKCA9ezZkzk6OrLMzEyWm5srv5WWlqooWs1QVZsoAxr9VO+qa4/169czIyMjtnfvXpaRkcGCg4OZrq4uy8zMVEGkmqGq9igpKWG2trasb9++LDk5mWVmZrLQ0FAmEAjYoUOHVBStevP392dt2rRhBw8eZBKJhO3fv5+ZmJiwBQsWyPdZtWoVa9asGTtw4AC7cOEC8/b2ZmKxuEF6/imp0UBxcXEMALt69arC4/Hx8QyA0ptEIlFNsBqiqjZRhpKa+vey9li5ciVr27Yt09fXZ71792YJCQkNHKFmqa490tPT2YgRI1jr1q2Zvr4+69KlS6Uh3qTuSKVSNmfOHGZlZcV0dXVZu3bt2OLFi1lxcbF8n7KyMhYSEsJMTU2Zjo4Oc3d3r9HftrogYKzcNICEEEIIIW8oqqkhhBBCiFqgpIYQQgghaoGSGkIIIYSoBUpqCCGEEKIWKKkhhBBCiFqgpIYQQgghaoGSGkIIIYSoBUpqCCGEEKIWKKkhhBBCiFqgpIYQQgghaoGSGkIIIYSoBUpqCCGEEKIW/h+931hGNGX1wgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(2)\n",
    "history = pd.DataFrame(history)\n",
    "\n",
    "fig.suptitle('Training History')\n",
    "print(ax[0])\n",
    "\n",
    "ax[0].plot(range(prev_epoch+1, prev_epoch+len(history)+1), history[\"train_loss\"], color = 'b', label = \"Training Loss\")\n",
    "ax[0].plot(range(prev_epoch+1, prev_epoch+len(history)+1), history[\"val_loss\"], color = 'r', label = \"Validation Loss\")\n",
    "\n",
    "ax[1].plot(range(prev_epoch+1, prev_epoch+len(history)+1), history[\"train_acc\"], color = 'b', label = \"Training Accuracy\")\n",
    "ax[1].plot(range(prev_epoch+1, prev_epoch+len(history)+1), history[\"val_acc\"], color = 'r', label = \"Validation Accuracy\")\n",
    "\n",
    "ax[0].legend()\n",
    "ax[1].legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "a472d31b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Axes(0.125,0.53;0.775x0.35)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiMAAAHNCAYAAADMjHveAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8g+/7EAAAACXBIWXMAAA9hAAAPYQGoP6dpAACJ6ElEQVR4nOzdd3iUVdrA4d+kJ6SSQAqk0HszFAMiCChNBGyIIEFFVxcURD4BAdF1XXBxFburq7AqiuIKojQRAek91BBagFCSECC9z5zvj+NMMiSBBJJMynNf17lm5i3zPm8GMk9ONSilFEIIIYQQNmJn6wCEEEIIUbtJMiKEEEIIm5JkRAghhBA2JcmIEEIIIWxKkhEhhBBC2JQkI0IIIYSwKUlGhBBCCGFTkowIIYQQwqYkGRFCCCGETUkyIkQNMXbsWMLCwm7q3FdffRWDwVC+AVWQ06dPYzAYWLhwoa1DEUKUE0lGhKhgBoOhVGXDhg22DtUmxo4di7u7e4n7DQYDEyZMuOXrfPTRR5LACFFFOdg6ACFquq+++srq9ZdffsnatWuLbG/VqtUtXeezzz7DZDLd1LkzZ85k2rRpt3T9yhIaGkpWVhaOjo5lOu+jjz7Cz8+PsWPHVkxgQoibJsmIEBVs9OjRVq+3b9/O2rVri2y/VmZmJm5ubqW+Tlm/nAtzcHDAwaF6/DowGAy4uLjYOgwAsrOzcXJyws5OKpmFuBXyP0iIKqB37960bduWPXv2cOedd+Lm5sbLL78MwE8//cTgwYMJCgrC2dmZJk2a8Prrr2M0Gq3e49o+I+a+FW+99RaffvopTZo0wdnZmS5durBr1y6rc4vrM2JuHlm2bBlt27bF2dmZNm3asHr16iLxb9iwgc6dO+Pi4kKTJk3497//XWH9UIrrMxIfH8/jjz9Ow4YNcXZ2JjAwkKFDh3L69GkAwsLCOHz4MBs3brQ0i/Xu3dty/qlTp3jooYeoW7cubm5u3H777axYsaLIPRoMBhYvXszMmTNp0KABbm5uREVFYTAYeOedd4rEunXrVgwGA99++225/xyEqEmqx59CQtQCly9fZuDAgTzyyCOMHj0af39/ABYuXIi7uzuTJ0/G3d2d33//nVdeeYXU1FTmzZt3w/f95ptvSEtL4y9/+QsGg4F//vOf3H///Zw6deqGtSmbN2/mxx9/5K9//SseHh689957PPDAA5w9exZfX18A9u3bx4ABAwgMDOS1117DaDTyt7/9jXr16pXp/pOSksp0fGEPPPAAhw8f5rnnniMsLIzExETWrl3L2bNnCQsLY/78+Tz33HO4u7szY8YMAMvPNyEhge7du5OZmcnzzz+Pr68v//3vf7nvvvv44YcfGD58uNW1Xn/9dZycnJgyZQo5OTm0bNmSHj16sGjRIl544QWrYxctWoSHhwdDhw696XsTolZQQohKNX78eHXtf71evXopQH3yySdFjs/MzCyy7S9/+Ytyc3NT2dnZlm2RkZEqNDTU8jo2NlYBytfXV125csWy/aefflKA+vnnny3bZs+eXSQmQDk5OakTJ05Ytu3fv18B6v3337dsGzJkiHJzc1Pnz5+3bDt+/LhycHAo8p7FiYyMVMB1y/jx44vc14IFC5RSSl29elUBat68ede9Tps2bVSvXr2KbJ80aZIC1KZNmyzb0tLSVKNGjVRYWJgyGo1KKaXWr1+vANW4ceMin8m///1vBajo6GjLttzcXOXn56ciIyNv+DMQoraTZhohqghnZ2cef/zxIttdXV0tz9PS0khKSqJnz55kZmZy9OjRG77viBEj8PHxsbzu2bMnoJsmbqRfv340adLE8rp9+/Z4enpazjUajfz2228MGzaMoKAgy3FNmzZl4MCBN3x/MxcXF9auXVtsuRFXV1ecnJzYsGEDV69eLfU1zVauXEnXrl254447LNvc3d15+umnOX36NEeOHLE6PjIy0uozAXj44YdxcXFh0aJFlm1r1qwhKSnphn2DhBDSTCNEldGgQQOcnJyKbD98+DAzZ87k999/JzU11WpfSkrKDd83JCTE6rU5MSnNF/e155rPN5+bmJhIVlYWTZs2LXJccdtKYm9vT79+/Up9fGHOzs68+eabvPjii/j7+3P77bdz7733MmbMGAICAm54/pkzZ+jWrVuR7ebRTWfOnKFt27aW7Y0aNSpyrLe3N0OGDOGbb77h9ddfB3QTTYMGDejTp89N3ZcQtYnUjAhRRVz71zZAcnIyvXr1Yv/+/fztb3/j559/Zu3atbz55psApRrKa29vX+x2pVSFnluZJk2axLFjx5gzZw4uLi7MmjWLVq1asW/fvnK/VnGfE8CYMWM4deoUW7duJS0tjeXLlzNy5EgZaSNEKUjNiBBV2IYNG7h8+TI//vgjd955p2V7bGysDaMqUL9+fVxcXDhx4kSRfcVtq0hNmjThxRdf5MUXX+T48eN07NiRf/3rX3z99dcAJY7sCQ0NJSYmpsh2cxNYaGhoqa4/YMAA6tWrx6JFi+jWrRuZmZk89thjN3k3QtQukrILUYWZayYK10Tk5uby0Ucf2SokK+bmlWXLlnHhwgXL9hMnTrBq1apKiSEzM5Ps7GyrbU2aNMHDw4OcnBzLtjp16pCcnFzk/EGDBrFz5062bdtm2ZaRkcGnn35KWFgYrVu3LlUcDg4OjBw5ku+//56FCxfSrl072rdvf3M3JUQtIzUjQlRh3bt3x8fHh8jISJ5//nkMBgNfffVVlWomefXVV/n111/p0aMHzz77LEajkQ8++IC2bdsSFRVV4dc/duwYffv25eGHH6Z169Y4ODiwdOlSEhISeOSRRyzHhYeH8/HHH/P3v/+dpk2bUr9+ffr06cO0adP49ttvGThwIM8//zx169blv//9L7Gxsfzvf/8rUzPLmDFjeO+991i/fr2lKU0IcWOSjAhRhfn6+vLLL7/w4osvMnPmTHx8fBg9ejR9+/alf//+tg4P0F/yq1atYsqUKcyaNYvg4GD+9re/ER0dXarRPrcqODiYkSNHsm7dOr766iscHBxo2bIl33//PQ888IDluFdeeYUzZ87wz3/+k7S0NHr16kWfPn3w9/dn69atTJ06lffff5/s7Gzat2/Pzz//zODBg8sUS3h4OG3atCE6OppRo0aV960KUWMZVFX6E0sIUWMMGzaMw4cPc/z4cVuHUqk6depE3bp1Wbduna1DEaLakD4jQohblpWVZfX6+PHjrFy50mrK9dpg9+7dREVFMWbMGFuHIkS1IjUjQohbFhgYyNixY2ncuDFnzpzh448/Jicnh3379tGsWTNbh1fhDh06xJ49e/jXv/5FUlISp06dqjKL+QlRHUifESHELRswYADffvst8fHxODs7ExERwT/+8Y9akYgA/PDDD/ztb3+jRYsWfPvtt5KICFFGUjMihBBCCJuSPiNCCCGEsClJRoQQQghhU5KMCCGEEMKmJBkRQgghhE1JMiKEEEIIm5JkRAghhBA2JcmIEEIIIWxKkhEhhBBC2JQkI0IIIYSwKUlGhBBCCGFTkowIIYQQwqYkGRFCCCGETUkyIoQQQgibkmRECCGEEDYlyYgQQgghbEqSESGEEELYlCQjQgghhLApSUaEEEIIYVOSjAghhBDCpiQZEUIIIYRNSTIihBBCCJuSZEQIIYQQNiXJiBBCCCFsSpIRIYQQQtiUJCNCCCGEsClJRoQQQghhU5KMCCGEEMKmJBkRQgghhE052DqA0jCZTFy4cAEPDw8MBoOtwxFCCCFEKSilSEtLIygoCDu7kus/qkUycuHCBYKDg20dhhBCCCFuQlxcHA0bNixxf7VIRjw8PAB9M56enjaORgghhBClkZqaSnBwsOV7vCTVIhkxN814enpKMiKEEEJUMzfqYiEdWIUQQghhU5KMCCGEEMKmJBkRQgghhE1Viz4jFeXFF+H33+G776B5c1tHI4QQFUMpRX5+Pkaj0dahiBrG3t4eBweHW552o1YnIzt3QlQUbN0qyYgQombKzc3l4sWLZGZm2joUUUO5ubkRGBiIk5PTTb9HrU5GIiJg82bYtg3GjrV1NEIIUb5MJhOxsbHY29sTFBSEk5OTTBwpyo1SitzcXC5dukRsbCzNmjW77sRm11Ork5Hu3fXj1q22jUMIISpCbm4uJpOJ4OBg3NzcbB2OqIFcXV1xdHTkzJkz5Obm4uLiclPvU6s7sEZE6MfDhyElxbaxCCFERbnZv1aFKI3y+PdVq/+F+vtDo0aglO4/IoQQQojKV6uTEZCmGiGEqC3CwsKYP39+qY/fsGEDBoOB5OTkCotJaLU+GTE31WzbZts4hBBCaAaD4brl1Vdfvan33bVrF08//XSpj+/evTsXL17Ey8vrpq5XWpL01PIOrFCQjGzfDiYTSNOqEELY1sWLFy3Pv/vuO1555RViYmIs29zd3S3PlVIYjUYcHG78dVavXr0yxeHk5ERAQECZzhE3p9Z/9bZvD25uugNrdLStoxFCCBEQEGApXl5eGAwGy+ujR4/i4eHBqlWrCA8Px9nZmc2bN3Py5EmGDh2Kv78/7u7udOnShd9++83qfa9tpjEYDPznP/9h+PDhuLm50axZM5YvX27Zf22NxcKFC/H29mbNmjW0atUKd3d3BgwYYJU85efn8/zzz+Pt7Y2vry9Tp04lMjKSYcOG3fTP4+rVq4wZMwYfHx/c3NwYOHAgx48ft+w/c+YMQ4YMwcfHhzp16tCmTRtWrlxpOXfUqFHUq1cPV1dXmjVrxoIFC246lopS65MRBwfo2lU/l6YaIURNpxRkZNimKFV+9zFt2jTmzp1LdHQ07du3Jz09nUGDBrFu3Tr27dvHgAEDGDJkCGfPnr3u+7z22ms8/PDDHDhwgEGDBjFq1CiuXLlS4vGZmZm89dZbfPXVV/zxxx+cPXuWKVOmWPa/+eabLFq0iAULFrBlyxZSU1NZtmzZLd3r2LFj2b17N8uXL2fbtm0opRg0aBB5eXkAjB8/npycHP744w8OHjzIm2++aak9mjVrFkeOHGHVqlVER0fz8ccf4+fnd0vxVAhVDaSkpChApaSkVMj7T5+uFCj1xBMV8vZCCGETWVlZ6siRIyorK8uyLT1d/76zRUlPL/s9LFiwQHl5eVler1+/XgFq2bJlNzy3TZs26v3337e8Dg0NVe+8847lNaBmzpxZ6GeTrgC1atUqq2tdvXrVEgugTpw4YTnnww8/VP7+/pbX/v7+at68eZbX+fn5KiQkRA0dOrTEOK+9TmHHjh1TgNqyZYtlW1JSknJ1dVXff/+9Ukqpdu3aqVdffbXY9x4yZIh6/PHHS7x2eSju35lZab+/a33NCMiIGiGEqG46d+5s9To9PZ0pU6bQqlUrvL29cXd3Jzo6+oY1I+3bt7c8r1OnDp6eniQmJpZ4vJubG02aNLG8DgwMtByfkpJCQkICXc3V7ei1W8LDw8t0b4VFR0fj4OBAt27dLNt8fX1p0aIF0X/2LXj++ef5+9//To8ePZg9ezYHDhywHPvss8+yePFiOnbsyEsvvcTWKvpFJ8kIcPvt+vHoUbhO7ZwQQlR7bm6Qnm6bUp6TwNapU8fq9ZQpU1i6dCn/+Mc/2LRpE1FRUbRr147c3Nzrvo+jo6PVa4PBgMlkKtPxqjzbn27CuHHjOHXqFI899hgHDx6kc+fOvP/++wAMHDiQM2fO8MILL3DhwgX69u1r1axUVUgyAvj5QbNm+vmOHbaNRQghKpLBAHXq2KZU5LI4W7ZsYezYsQwfPpx27doREBDA6dOnK+6CxfDy8sLf359du3ZZthmNRvbu3XvT79mqVSvy8/PZUejL6fLly8TExNC6dWvLtuDgYJ555hl+/PFHXnzxRT777DPLvnr16hEZGcnXX3/N/Pnz+fTTT286nopS64f2mkVEwPHjuqlm4EBbRyOEEKIsmjVrxo8//siQIUMwGAzMmjXrujUcFeW5555jzpw5NG3alJYtW/L+++9z9erVUi1QePDgQTw8PCyvDQYDHTp0YOjQoTz11FP8+9//xsPDg2nTptGgQQOGDh0KwKRJkxg4cCDNmzfn6tWrrF+/nlatWgHwyiuvEB4eTps2bcjJyeGXX36x7KtKJBn5U/fu8OWXMqJGCCGqo7fffpsnnniC7t274+fnx9SpU0lNTa30OKZOnUp8fDxjxozB3t6ep59+mv79+2Nvb3/Dc++8806r1/b29uTn57NgwQImTpzIvffeS25uLnfeeScrV660NBkZjUbGjx/PuXPn8PT0ZMCAAbzzzjuAnitl+vTpnD59GldXV3r27MnixYvL/8ZvkUHZurGrFFJTU/Hy8iIlJQVPT8/ye2Ol4MABaNuWA4ft6dAB3N0hORlK8e9GCCGqtOzsbGJjY2nUqNFNr6Yqbo3JZKJVq1Y8/PDDvP7667YOp0Jc799Zab+/a2+fEaWgTRvo2BF27aJNG/Dw0J2sDh2ydXBCCCGqozNnzvDZZ59x7NgxDh48yLPPPktsbCyPPvqorUOr0mpvMmIw6GQEYNUq7O3BPHJKmmqEEELcDDs7OxYuXEiXLl3o0aMHBw8e5LfffquS/TSqktqbjEBBT9XVqwFZNE8IIcStCQ4OZsuWLaSkpJCamsrWrVuL9AURRdXuZKR/f/24axckJVmSkSo6J4wQQghRI9XuZKRBA2jXTvcf+fVXy+RnJ07ApUu2DU0IIYSoLWp3MgJWTTU+PmBu1tu+3XYhCSGEELWJJCMDBujH1avBZJKmGiGEEKKSVXgy8vHHH9O+fXs8PT3x9PQkIiKCVatWVfRlS69HDz25yKVLsG+fZdE86cQqhBBCVI4KT0YaNmzI3Llz2bNnD7t376ZPnz4MHTqUw4cPV/SlS8fJCfr21c9Xr7bUjOzcCdnZtgtLCCGEqC0qPBkZMmQIgwYNolmzZjRv3pw33ngDd3d3tlelThnmfiOrVtGyJTRsCFlZ8Ouvtg1LCCHEzevduzeTJk2yvA4LC2P+/PnXPcdgMLBs2bJbvnZ5vU9tUal9RoxGI4sXLyYjI4MIcxVEMXJyckhNTbUqFco8xHfbNuxSrvLgg/rlkiUVe1khhBBFDRkyhAHm/nzX2LRpEwaDgQMHDpT5fXft2sXTTz99q+FZefXVV+nYsWOR7RcvXmRgBa+6unDhQry9vSv0GpWlUpKRgwcP4u7ujrOzM8888wxLly61Wvr4WnPmzMHLy8tSgoODKzbAsDBo2RJMJvjtNx56SG9evhxycir20kIIIaw9+eSTrF27lnPnzhXZt2DBAjp37kz79u3L/L716tXDzc2tPEK8oYCAAJydnSvlWjVBpSQjLVq0ICoqih07dvDss88SGRnJkSNHSjx++vTppKSkWEpcXFzFB1loiO/tt+spSFJTpalGCCEq27333ku9evVYuHCh1fb09HSWLFnCk08+yeXLlxk5ciQNGjTAzc2Ndu3a8e233173fa9tpjl+/Dh33nknLi4utG7dmrVr1xY5Z+rUqTRv3hw3NzcaN27MrFmzyMvLA3TNxGuvvcb+/fsxGAwYDAZLzNc20xw8eJA+ffrg6uqKr68vTz/9NOnp6Zb9Y8eOZdiwYbz11lsEBgbi6+vL+PHjLde6GWfPnmXo0KG4u7vj6enJww8/TEJCgmX//v37ueuuu/Dw8MDT05Pw8HB2794N6DV2hgwZgo+PD3Xq1KFNmzasXLnypmO5EYcKe+dCnJycaNq0KQDh4eHs2rWLd999l3//+9/FHu/s7Fz5GeWAAfDOO7B6NXYGxQMPGHjvPd1UM2RI5YYihBAVRinIzLTNtd3c9LpgN+Dg4MCYMWNYuHAhM2bMwPDnOUuWLMFoNDJy5EjS09MJDw9n6tSpeHp6smLFCh577DGaNGlC165db3gNk8nE/fffj7+/Pzt27CAlJcWqf4mZh4cHCxcuJCgoiIMHD/LUU0/h4eHBSy+9xIgRIzh06BCrV6/mt99+A8DLy6vIe2RkZNC/f38iIiLYtWsXiYmJjBs3jgkTJlglXOvXrycwMJD169dz4sQJRowYQceOHXnqqadueD/F3Z85Edm4cSP5+fmMHz+eESNGsGHDBgBGjRpFp06d+Pjjj7G3tycqKgpHR0cAxo8fT25uLn/88Qd16tThyJEjuLu7lzmOUlM2cNddd6nIyMhSH5+SkqIAlZKSUnFBZWUp5eqqFCi1f7/atEk/9fRUKju74i4rhBAVJSsrSx05ckRlZWUVbExP17/cbFHS00sde3R0tALU+vXrLdt69uypRo8eXeI5gwcPVi+++KLlda9evdTEiRMtr0NDQ9U777yjlFJqzZo1ysHBQZ0/f96yf9WqVQpQS5cuLfEa8+bNU+Hh4ZbXs2fPVh06dChyXOH3+fTTT5WPj49KL3T/K1asUHZ2dio+Pl4ppVRkZKQKDQ1V+fn5lmMeeughNWLEiBJjWbBggfLy8ip236+//qrs7e3V2bNnLdsOHz6sALVz506llFIeHh5q4cKFxZ7frl079eqrr5Z47cKK/Xf2p9J+f1d4M8306dP5448/OH36NAcPHmT69Ols2LCBUaNGVfSly8bFBe66Sz9fvZru3SEoSDfVFFNzJ4QQogK1bNmS7t2788UXXwBw4sQJNm3axJNPPgnoARGvv/467dq1o27duri7u7NmzRrOnj1bqvePjo4mODiYoKAgy7biBlZ899139OjRg4CAANzd3Zk5c2apr1H4Wh06dKBOnTqWbT169MBkMhETE2PZ1qZNG+zt7S2vAwMDSUxMLNO1Cl8zODjYqs9l69at8fb2Jjo6GoDJkyczbtw4+vXrx9y5czl58qTl2Oeff56///3v9OjRg9mzZ99Uh+GyqPBkJDExkTFjxtCiRQv69u3Lrl27WLNmDXfffXdFX7rsCg3xtbODBx7QL2VUjRCixnBzg/R025Qydh598skn+d///kdaWhoLFiygSZMm9OrVC4B58+bx7rvvMnXqVNavX09UVBT9+/cnNze33H5U27ZtY9SoUQwaNIhffvmFffv2MWPGjHK9RmHmJhIzg8GAyWSqkGuBHgl0+PBhBg8ezO+//07r1q1ZunQpAOPGjePUqVM89thjHDx4kM6dO/P+++9XWCwV3mfk888/r+hLlB/zULLNmyEtjYce8uD99+Gnn/SoGukYLYSo9gwGKPQXelX28MMPM3HiRL755hu+/PJLnn32WUv/kS1btjB06FBGjx4N6D4Sx44du+5IzcJatWpFXFwcFy9eJDAwEKDI/Fdbt24lNDSUGTNmWLadOXPG6hgnJyeMRuMNr7Vw4UIyMjIstSNbtmzBzs6OFi1alCresjLfX1xcnKV25MiRIyQnJ1v9jJo3b07z5s154YUXGDlyJAsWLGD48OEABAcH88wzz/DMM88wffp0PvvsM5577rkKiVfWpimsaVNd8vNh3Tp69IDAQEhJgT/7JgkhhKgk7u7ujBgxgunTp3Px4kXGjh1r2desWTPWrl3L1q1biY6O5i9/+YvVSJEb6devH82bNycyMpL9+/ezadMmq6TDfI2zZ8+yePFiTp48yXvvvWepOTALCwsjNjaWqKgokpKSyClmPohRo0bh4uJCZGQkhw4dYv369Tz33HM89thj+Pv7l+2Hcg2j0UhUVJRViY6Opl+/frRr145Ro0axd+9edu7cyZgxY+jVqxedO3cmKyuLCRMmsGHDBs6cOcOWLVvYtWsXrf5cLXbSpEmsWbOG2NhY9u7dy/r16y37KoIkI9cqtHCeNNUIIYRtPfnkk1y9epX+/ftb9e+YOXMmt912G/3796d3794EBAQwbNiwUr+vnZ0dS5cuJSsri65duzJu3DjeeOMNq2Puu+8+XnjhBSZMmEDHjh3ZunUrs2bNsjrmgQceYMCAAdx1113Uq1ev2OHFbm5urFmzhitXrtClSxcefPBB+vbtywcffFC2H0Yx0tPT6dSpk1UZMmQIBoOBn376CR8fH+6880769etH48aN+e677wCwt7fn8uXLjBkzhubNm/Pwww8zcOBAXnvtNUAnOePHj6dVq1YMGDCA5s2b89FHH91yvCUxKKVUhb17OUlNTcXLy4uUlBQ8PT0r9mIrVsC990JICJw+zR+bDPTqBd7ekJCgl7IRQojqIDs7m9jYWBo1aoSLi4utwxE11PX+nZX2+1tqRq511116ZM3Zs3DgAD16QEAAJCdLU40QQghRESQZuZabW8FaNT/+iL29NNUIIYQQFUmSkeLcf79+/LOjknmtmmXLoIJGdAkhhBC1liQjxbn3XnBwgIMH4fhx7rgD/P11U826dbYOTgghhKhZJBkpTt260Lu3fr50qTTVCCGEEBVIkpGSXNNU8+CD+uXy5XoaEiGEqC6qwaBJUY2Vx78vSUZKMmyYnqlw+3Y4f56ePcHHBy5fhq1bbR2cEELcmHl68UxbrdIragXzv69rp7MviwqfDr7aCgyE22+Hbdtg2TIcxo/n3nvhq6/09PB33mnrAIUQ4vrs7e3x9va2LLbm5uZmmU5diFullCIzM5PExES8vb2tFvkrK0lGruf++3UysnQpjB/PsGE6GVm2DN56S1ecCCFEVRYQEABw06u/CnEj3t7eln9nN0tmYL2ekyf1WjX29pCQQLqzL35+etG8gwehbdvKC0UIIW6F0WgkLy/P1mGIGsbR0fG6NSKl/f6WmpHradIE2reHAwfg559xHzuWu++GX37RtSOSjAghqgt7e/tbqkYXoiJJB9YbuWZUzdCh+uVPP9koHiGEEKKGkWTkRszJyJo1kJ7OkCG6r8ju3XDunG1DE0IIIWoCSUZupG1b3W8kJwdWrcLfHyIi9K7ly20bmhBCCFETSDJyIwYDDB+un0tTjRBCCFHuJBkpDXNTzS+/QE4Ow4bpl+vXQ0qKzaISQgghagRJRkqja1cICoK0NFi3jubNoWVLyMuDVatsHZwQQghRvUkyUhp2dliqQ378ESh4uWyZLQISQgghag5JRkrL3FFkzRpQyvJy5Urdt1UIIYQQN0eSkdLq2ROcnfV43mPH6NoVAgJ0y82GDbYOTgghhKi+JBkpLVdX6NFDP//tN+zs4L779EsZVSOEEELcPElGyqJfP/24bh1Q0G9k+XIwmWwTkhBCCFHdSTJSFuZk5PffwWikTx9wd4fz52HPHtuGJoQQQlRXkoyUxW23gbe3nlxkzx6cnWHgQL3rz0E2QgghhCgjSUbKwt4e+vTRz3/7DYAHHtAvf/gBlLJRXEIIIUQ1JslIWZmbav5MRgYNAhcXOHECDhywYVxCCCFENSXJSFmZk5EtWyAzEw8PGDBAb/rhB9uFJYQQQlRXkoyUVdOmEBICubmweTMADz2kdy1ZIk01QgghRFlJMlJWBkORppp77wUnJ4iJgcOHbRibEEIIUQ1VeDIyZ84cunTpgoeHB/Xr12fYsGHExMRU9GUr1jXJiKcn9O+vN0lTjRBCCFE2FZ6MbNy4kfHjx7N9+3bWrl1LXl4e99xzDxkZGRV96YpjHlGzbx8kJQHw4IN6kyQjQgghRNkYlKrcXg6XLl2ifv36bNy4kTvvvLNU56SmpuLl5UVKSgqenp4VHGEptW8PBw/Cd9/Bww+TnAz160NeHhw5Aq1a2TpAIYQQwrZK+/1d6X1GUlJSAKhbt26Jx+Tk5JCammpVqpxrmmq8veHuu/UmqR0RQgghSq9SkxGTycSkSZPo0aMHbdu2LfG4OXPm4OXlZSnBwcGVGGUpXZOMgDTVCCGEEDejUptpnn32WVatWsXmzZtp2LBhicfl5OSQk5NjeZ2amkpwcHDVaqZJTwcfH8jPh5MnoXFjrlwBf3+9KSYGmje3dZBCCCGE7VS5ZpoJEybwyy+/sH79+usmIgDOzs54enpalSrH3R0iIvTzP2tH6taFvn31pv/9z0ZxCSGEENVMhScjSikmTJjA0qVL+f3332nUqFFFX7LyXKepZskSG8QjhBBCVEMVnoyMHz+er7/+mm+++QYPDw/i4+OJj48nKyuroi9d8czJyLp1YDIBMGyYXk9v3z7deiOEEEKI66vwZOTjjz8mJSWF3r17ExgYaCnfffddRV+64nXpAh4ecOUKREUB4OcHvXvr3dJUI4QQQtxYpTTTFFfGjh1b0ZeueI6OcNdd+nmhdpnCa9UIIYQQ4vpkbZpbZU6qPv0U/pxVdtgwsLOD3bvh2DGbRSaEEEJUC5KM3Kr77oPGjXVTzZdfAnp478CBevcHH9gwNiGEEKIakGTkVtnbw6RJ+vk771g6sk6cqDctWAB/TjorhBBCiGJIMlIeHn8cvLzg+HFYuRLQA21at9Zzo33xhY3jE0IIIaowSUbKg7s7/OUv+vnbbwNgMBTUjrz/PhiNNopNCCGEqOIkGSkvEyboJpv16/UkI8Do0XpW1thY+OUXG8cnhBBCVFGSjJSX4GB4+GH9/J13AHBzg6ee0pvefddGcQkhhBBVnCQj5WnyZP24eDFcuADA+PEFFSYHDtgwNiGEEKKKkmSkPHXuDD17Ql4efPghoCtMHnhA75baESGEEKIoSUbKm7l25JNPLJOgmTuyLloEly7ZKC4hhBCiipJkpLwNGVJkErSICF1pkpOjJ2oVQgghRAFJRspb4UnQ5s8Hk8lqmO9HH0Furq2CE0IIIaoeSUYqgnkStGPHYONGQA+0CQjQ/Vp/+MHG8QkhhBBViCQjFcHdvWCY79dfA+DkBH/9q970z39Cfr6NYhNCCCGqGElGKsro0frxhx8gKwuAZ58Fb2/Yv18W0BNCCCHMJBmpKHfcASEhkJpqmX7Vz0/XigDMnAlxcTaMTwghhKgiJBmpKHZ2MGqUfv5nUw3Ak09Cjx561O+ECaCUjeITQgghqghJRiqSualm5UpISgJ0jvLpp+DoCMuXw9KlNoxPCCGEqAIkGalIrVtDp066t+qSJVabp07Vz597DlJSbBSfEEIIUQVIMlLRzLUjhZpqAF5+GZo21UN9Z860QVxCCCFEFSHJSEV75BHdNrN1K5w6Zdns6qpnjAe9jM2OHTaKTwghhLAxSUYqWlAQ9O2rn3/zjdWuvn1hzBjdifXpp/X6ekIIIURtI8lIZSjcVHPN8Jm33gJfXzhwAObNs0FsQgghhI1JMlIZhg/X7TIxMbBnj9WuevXgnXf089mzYdcuG8QnhBBC2JAkI5XBwwOGDdPPr+nICrri5KGH9KCbUaMgPb1ywxNCCCFsSZKRymJuqvn22yIL0xgM8O9/Q8OGcPx4waK/QgghRG0gyUhluftu3SaTmAi//VZkt48PfPWVTkw+/xz+9z8bxCiEEELYgCQjlcXRUQ/zBfjii2IP6d27YDK0p56Cc+cqJzQhhBDCliQZqUyRkfpxyRJYuLDYQ157DTp3hqtX9bBfk6nywhNCCCFsQZKRyhQeDq+8op8//TRs3lzkECcnWLQI3Nxg/Xo99FcIIYSoySQZqWyzZ8ODD+oZzoYPh9jYIoc0bw7vvaefz5gBGzdWcoxCCCFEJZJkpLLZ2cF//6trSZKS4L77IC2tyGFPPAEPP6wH3tx3H+zfb4NYhRBCiEpQKcnIH3/8wZAhQwgKCsJgMLBs2bLKuGzV5eYGP/0EgYFw6BA8+igYjVaHGAy6W0nPnpCaCgMGFFuJIoQQQlR7lZKMZGRk0KFDBz788MPKuFz10KCBTkhcXOCXX2DatCKHuLrC8uXQrh3Ex8M99+iRwUIIIURNUinJyMCBA/n73//O8OHDK+Ny1UeXLgWjat56Cz74oMgh3t6wejWEhsKJEzB4cLGtOkIIIUS1VSX7jOTk5JCammpVaqwRI+DVV/Xz556DuXOLHBIUBGvW6AX1du+GBx6A3NzKDVMIIYSoKFUyGZkzZw5eXl6WEhwcbOuQKtYrr8CsWfr59Om6XLO6b4sWsHKl7m6ydq2eXV4SEiGEEDVBlUxGpk+fTkpKiqXExcXZOqSKZTDA3/4G8+bp13PnwoQJRWY869pVTxPv4KDnTevfX0+OJoQQQlRnVTIZcXZ2xtPT06rUClOm6BXzDAb46CMYO7bIonoDBsCKFXoh4A0bICICTp2ySbRCCCFEuaiSyUit9vTTegpWe3u9ct5DDxVpj7nnHj15a8OGEBMDt98OO3bYKF4hhBDiFlVKMpKenk5UVBRRUVEAxMbGEhUVxdmzZyvj8tXPyJGwdCk4O8OyZfDCC0UOad9eJyCdOsGlS3qRPVnpVwghRHVUKcnI7t276dSpE506dQJg8uTJdOrUiVfM67SIooYM0R1DzE02//lPkUOCguCPP/Rw3+xsXYkyfTpkZNggXiGEEOImGZS6ZthGFZSamoqXlxcpKSm1p/+I2RtvwMyZ4OioO4l0717kkPx8mDQJzHPKhYTA22/D/ffrXEYIIYSwhdJ+f0ufkaru5ZcLFtZ74AE4f77IIQ4O8P778OOPOhE5e1af0r+/7lMihBBCVGWSjFR1BgMsWFAwJ/zw4bpNppjDhg+H6Gg9ZYmzs56PpF07mDpVZm0VQghRdUkyUh24u+t1bOrWhV274JlnikyKZubmpqcsOXxY9yXJy4N//hOaNdPdTq5Zj08IIYSwOUlGqotGjeD77/WQ3//+F/7xjyJzkBTWpIlef2/5cmjaFBIS4Kmn9Oib336rxLiFEEKIG5BkpDrp21cvqAe6U2uzZnpxvczMEk8ZMkTXkrzzjl507+BBuPtuvf3o0coJWwghhLgeSUaqm4kT9VCZevXg9Gm9uF5ICLz2Gly+XOwpTk56tM2JE/D887rD6y+/QOvWMHQorF9fYquPEEIIUeFkaG91lZWlO7a+9RbExuptbm56vvjevXVp0wbsiuabMTG6U+tPPxVs69hRJyyPPKI7vwohhBC3qrTf35KMVHf5+Xrq1TffhH37rPf5+UGvXnqM79ixeq6SQo4ehffeg4ULdW4D4O+vVwQeOlRPaWJvXyl3IYQQogaSZKS2UQp27oTff9ftLlu2WPclefRRvdZNMTUlV67Ap5/q7ieFpzHx9YV774X77tPr4bi7V8J9CCGEqDEkGantcnNh92749Vc9i2t+vu4wMn9+idOy5uXp0TfLlumVga9eLdjn7AxjxsD//Z/uNyuEEELciCQjosA338CoUfr5G2/oWV1vIC9PV64sX677lpw6pbcbDHqa+alToUuXCoxZCCFEtSfTwYsCjz4K776rn8+YAZ99dsNTHB11H9i339ajcDZt0k02SukuKl27Qp8+OlFJTKzY8IUQQtRsUjNSm8ycqWtG7Oz0isD331/mtzh0CObN05UthedcCwzUI3I6ddKPnTtDWJgs1CeEELWZNNOIopSCv/xF14w4OelqjXvuKbZT642cPau7n6xcCceOFT9PSUAA9OihR+V0764TFRk2LIQQtYckI6J4RiM8/LBe4hf03CStWunSurUu7duXqVojPV3P7BoVpcvevbB/v+53UpizM/TsqadCGTBAX0pqToQQouaSZESULDtbzzuydKkedVMcb2+47TZdnWF+bNZMT99aCllZsGeP7gS7dasuSUnWxzRsqJOSfv2gbVu9ho7UnAghRM0hyYi4sfx8PUzmyJGCcviwLtdWa4Du1dq0KbRsqUurVjpJadv2hpdSSk+y9uuvsHo1bNigc6LC7Oz0eoDmt2/ZUk8i27o1eHmVzy0LIYSoPJKMiJuXm6sTk717ddm3T7e7ZGQUf3ynTnpJ4EcfLXXWkJUFf/yhE5PNm/UU9WlpJR/fsKFOStq0gRYtdE7UtKneLrPECiFE1STJiChfJhOcO6erN8wlOlq3v5ibetzcdH+Up5+G228vU4cQpSA+Xiclhd/+8GHrWWGv5eQEjRvrFqTwcOjWTc9/4ut7i/crhBDilkkyIipHUpKeZv6zz3T2YFa/fkHHWHNp3lwnLpcv6znozcVkgjvu0DUsxVRzJCcXtCAdOaLnPTlxQrcwldTlpUkTnZi0bg0eHnoq+zp1dHF31xU4vr5Qt67eJh1phRCi/EkyIiqXUrBtm05KvvuuYOW9svDx0TOp9eunS5Mm180SjEZdWXPihE5Sdu2CHTv0UOOycHLSSYmvrx6OHBCg500xPzZooPOogABJWoQQoiwkGRG2k5mps4PoaOty8qRuyqlbt6D4+urjN26E1FTr9/Hy0ssI16una1rMj40a6Y4jLVsW2x5z9WpBYnL6tO7qkpEBuSlZBCYdpGHyQRKzPNmUcRtH8xoDpcswPDz0Zc2lSRPdZyU4WCcsTk63/qMTQoiaRJIRUfUoVXLVQn6+Xtjvt9902bq1+BE91/L1LejR6uVV0BZjLikpBROgHD2qm4QKh+TlRU6rTqQ26URCw3CONezDqaxA4uPh4kXdj+XMGZ3UXHOqFYNB500NG+qkxcXFutSpAyEhevqW0FD9WK+e1LQIIWo2SUZE9ZaRoad5TUyES5cKHuPjdbtMTAzExZX9fevVgw4ddEeUAweK73QSHg6DB+vSuTPY2ZGToy977Ji+dEyMTlDi4nRTUU5O2UNxddXhODrq6VsKFycnPeeKuZiTGl9f8PMrKPXq6QomL6+CXOwmJtQVQogKIcmIqPkyMuD4cZ0ZxMbqqWDNbTIZGfq1i4tOPjp21CUwsKA6Ii9PNx+ZhzBv26ZrZwrz94c779Tf9G5uOoNwddXPGzaE229HhYaRdNlAXJwe+ZORoedQycnRj9nZuoLm7FmdwJw+DRcuFD+F/q0yGMDTUxdvb118fKwfza1dhYuXl44zM7OgZGTo/sT16uni4lL+8QohajZJRoS4GQkJsGoV/PKLnqHtepOfmNWrp4fu3H67HldsMumanMIlI0O3zzRrBk2bkhvajLOmhlxJtsNo1K1U5pKXpytszMlMTo4umZl68FFSki6XLunHK1d0slN44cLiGDARQDxJ+JFH2Tu4uLvrW/Xz0zU0Pj7WpU4dHYu5IstcmWUy6dtu3rygv03z5joxUkrvNz8aDBWU9CQm6uFXskCSEJVKkhEhblVuLmzapCd8y8zUI4TMjxkZukYmKqp0fVuK4+ysqyUMBt22YjAUlDp1dHWFt3fB47XVGubnHh6o7Byyr2aRnpRNRlIWGUlZ5J86i33MEZxjo/E8F41P4lGc8rPINzhwzqUZMfatOZDfmt3ZbYgjGF8uE0A8DR0TaGCfQIBdAiaTgZicME6qRpwmjNOEcYZQsnHBHiMO5ONAPvYYcSWLIC7QkHMEE0dDztGQcxhQ7KYz27mdfXQil+snA3XqQFCQrsQKDNTP/fx085Wdna6tsbMr+JEpVVDLZH70zYunybmNNDi+Ab8jG3E7rYedK09PuO8+DA89pBeJlOoeISqUJCNCVIbsbD1D7Y4duuzbp7/gCreB+Pvrpp3YWN2sdPy4/iv9RlUZNVCuwYnDDh35I+920nGnHpesihcpHKM527mdHXRjO7dzgQaW8324QiuiLcWfBOqQYVW8SSaMM0WufQUf6nLV8jrN4MEm7yEc8b0Tg70dDnYm7O0U9nYmHAwmrrgEcdazLRdcm2DE3lJ7U7jGyvxoZ6dHVYWFFZTQUF0DVL9+xf9chaiqJBkRoirLz9edSC5fLvjT3lxMJl3zkpys219SUvTz5GTrNpDExILzQX8jmvu0uLjoiVFat7aeeK5RI90JuPBaREeO6KFDfn46cQoI0I/+/jpO83Ci2NiCvjklCQjQfWnMY54bNtQ1TDt3wvbtOv4ySnZvQGKdRvinHMcrO6FU55gwcNK9Aztde7PJvhe/5/bkTKoPnfO38RBLeJAfaMh1pvYtJBNXjtCag7QjhhY4koc3yfhwFW+S8SYZO0ycojEnaMpxmnGCppygKel4UK+eXr6pXTtd2rTRlV0ODrqWx/xob//nR6kUKlu3y9llZ6IU5NfxwuTmjkkZLP9MXF1136Aq3WnZZNIdxTdu1Os/xMToprK+ffWcQiEhto5QVDBJRoSoDfLzdbORi4sellPRlNJJUX6+9bepuVzvW1Epncxs364ngsnPL+gday516sDBg7qWaft2/fzaMdXBwQXJVXBw0eHc7u46CfPxKXL57Ow/87orJvK3bMdj9Q84nT2Owg5T4aLA4/Jp6l44jEPeTUzg96dMXMnFiTwcycXJ8tyAwh6jVXEgH1eycCMTO4r+WjZiRyqepOBFMt6cpwEnaUIsjbng2oREjyZc9Qojz8EVg53upG1u9bPDRGj+SZplH6BF9n6aZx2gSdZBjAYHzri04LRzS864tCDWuSXnXZpQ3yWVRk7nCbY7T5A6T/3883jmXsagTBiUyZw1YVAK7O0xuZiTYFcMbq7YOTvgcWIfPgc34ZiRXOLPJ6VeE84170NKcDt83HLwdsrE0yETNzIxZGfpcfKhoRAaSl5QKOm+oaQ7+mAwgH1eNg6ZqThkpemSmYpzTip2aSl6ziJzqVOnYGHPpk3LPCGQ0ahbYkts0TMadV+zCxd0D/YrV/T1OnbUP5NaTpIRIUT1l54Oe/boX/LNm+tf8u7ulXd9o1E3qR08qMuJE3oklbkPj7kopSf1O35cH3P8uO5dfItycMIOE46UrUkv58/Ex1w8SMOdEha6rGBpuLOZO9hIL6JpRVd20off6cIuHDCW+f0yccWJ3Js614g9F9yacN6jJTg64WKXgyvZOJODEznYoUhyCuK8oSEncoOJTgvmYHIw2SZHGjnE0dwljsYOZwk2xNHAGIdf7nl8cuKxV0VjMdo5kBTQjsSwLlxu0pWMkNbYuzji4GyPvZM9Ds665CSlkXEqgdy4BEwXE7BPSsApLYk8Rzdy69Ql36suJu+6GHzr4urphH/2GfxST+F9NRb3S7E4XziFwc5AXtPW5DZvQ17ztuQ2a6OLiyd52UbyckyWR2N2Hh6mFNzzrlInLxnX7Ku4ZF/FMT0ZwzN/Kfd2RUlGhBDCllJS9HTAubkFQ6TMjwZDQdtM4VKnjk52zMPIHRx0opOVVdBk9+f7qjNnyY85ifHYSTh1EsezJ7HPKHn0l9HJhfTQtqQ26kBKaHuSQ9pjp4y4X4jB49xR3M/HUOd8DG6JZ8hzcSfdqwHJdRpw2aUBCQ4NSDLUw2RwQKGrW8yPGI3Y5WRhn1tQHPKzOevSnCivXpzy7IjB0cGqOcrBAdxNqbRP/oMOSevwTo8jOceNK9luXM5yIwM3snDFh6uEcsZS/Ekscl+peJD2Z0nBi1Q8LSUND7xJphXRtOQoHlynifEW5GNPPAFcIIgUvGjHQQIoXZNiVRL1yXY6/qVbub5nab+/Hcr1qtfx4YcfMm/ePOLj4+nQoQPvv/8+Xbt2razLCyFE5TLPRHerDIaCBCUwsGAz4PhnAXTSkpqq26Jyc62Lqyv2TZrgZW+PFxBsdYG+1tfLz8fRwQEfwAdodOt3UAJP4N4/i9XlSUjQ3ZhAV4S5u4OzO+TZZeGYdFGPRPP0RLnVoQ52uBjBOw/8/hzoVniunJwcSMyFczkK+4QLuJ6Oxu38cXKzTWTkO5Oe70xGnjNpuc7k5ynCnC4QYojDP+8cPulxuF2Jwy4/l7yAELLqBZPuHcxVzxAuuwZzxa0hV12DuOrkT3aevaVD88/ZCrcr5wg6v4vQhJ00StqFf8ZJDCYjBmXCThktJdu+Dmlu/mR7+WP09ccQ4I9joB8qMwtT0hUMV69gn3IFx/QrGHKyiXcM4ax9I07SmON5jTic2Yi8HBNtOEwbdYjW6jCt1SGaqBPYU3Ta6HzsSbXzJgVvrigfrihvruJDMt509/OusE/7RiqlZuS7775jzJgxfPLJJ3Tr1o358+ezZMkSYmJiqF+KKiGpGRFCCCHKIC9PZ3aFx8IX06fLaNSJW3q6nj+ovNfYqlLNNN26daNLly588MEHAJhMJoKDg3nuueeYNm3aDc+XZEQIIYSofkr7/V3hA8Jyc3PZs2cP/fr1K7ionR39+vVj27ZtFX15IYQQQlRxFd5nJCkpCaPRiL+/v9V2f39/jh49Wuw5OTk55BRaeSz12qXlhRBCCFFjVMmpcubMmYOXl5elBAcH3/gkIYQQQlRLFZ6M+Pn5YW9vT0KC9TCnhIQEAgICij1n+vTppKSkWErczSwVL4QQQohqocKbaZycnAgPD2fdunUMGzYM0B1Y161bx4QJE4o9x9nZGedCK2ua+9hKc40QQghRfZi/t280VqZS5hmZPHkykZGRdO7cma5duzJ//nwyMjJ4/PHHS3V+2p/LuEtzjRBCCFH9pKWl4XWdeXcqJRkZMWIEly5d4pVXXiE+Pp6OHTuyevXqIp1aSxIUFERcXBweHh4YDIabiiE1NZXg4GDi4uJq7PBguceaQe6xZpB7rBnkHm+NUoq0tDSCgoKue1ylzcA6YcKEEptlbsTOzo6GDRuWSxyenp419h+UmdxjzSD3WDPIPdYMco8373o1ImZVcjSNEEIIIWoPSUaEEEIIYVO1JhlxdnZm9uzZVqN0ahq5x5pB7rFmkHusGeQeK0elrE0jhBBCCFGSWlMzIoQQQoiqSZIRIYQQQtiUJCNCCCGEsClJRoQQQghhU7UiGfnwww8JCwvDxcWFbt26sXPnTluHdNP++OMPhgwZQlBQEAaDgWXLllntV0rxyiuvEBgYiKurK/369eP48eO2CfYmzZkzhy5duuDh4UH9+vUZNmwYMTExVsdkZ2czfvx4fH19cXd354EHHiiyGGNV9vHHH9O+fXvLJEMRERGsWrXKsr+6319x5s6di8FgYNKkSZZt1f0+X331VQwGg1Vp2bKlZX91vz+z8+fPM3r0aHx9fXF1daVdu3bs3r3bsr+6/94JCwsr8jkaDAbGjx8P1IzP0Wg0MmvWLBo1aoSrqytNmjTh9ddft1ozxqafo6rhFi9erJycnNQXX3yhDh8+rJ566inl7e2tEhISbB3aTVm5cqWaMWOG+vHHHxWgli5darV/7ty5ysvLSy1btkzt379f3XfffapRo0YqKyvLNgHfhP79+6sFCxaoQ4cOqaioKDVo0CAVEhKi0tPTLcc888wzKjg4WK1bt07t3r1b3X777ap79+42jLpsli9frlasWKGOHTumYmJi1Msvv6wcHR3VoUOHlFLV//6utXPnThUWFqbat2+vJk6caNle3e9z9uzZqk2bNurixYuWcunSJcv+6n5/Sil15coVFRoaqsaOHat27NihTp06pdasWaNOnDhhOaa6/95JTEy0+gzXrl2rALV+/XqlVM34HN944w3l6+urfvnlFxUbG6uWLFmi3N3d1bvvvms5xpafY41PRrp27arGjx9veW00GlVQUJCaM2eODaMqH9cmIyaTSQUEBKh58+ZZtiUnJytnZ2f17bff2iDC8pGYmKgAtXHjRqWUvidHR0e1ZMkSyzHR0dEKUNu2bbNVmLfMx8dH/ec//6lx95eWlqaaNWum1q5dq3r16mVJRmrCfc6ePVt16NCh2H014f6UUmrq1KnqjjvuKHF/Tfy9M3HiRNWkSRNlMplqzOc4ePBg9cQTT1htu//++9WoUaOUUrb/HGt0M01ubi579uyhX79+lm12dnb069ePbdu22TCyihEbG0t8fLzV/Xp5edGtW7dqfb8pKSkA1K1bF4A9e/aQl5dndZ8tW7YkJCSkWt6n0Whk8eLFZGRkEBERUePub/z48QwePNjqfqDmfI7Hjx8nKCiIxo0bM2rUKM6ePQvUnPtbvnw5nTt35qGHHqJ+/fp06tSJzz77zLK/pv3eyc3N5euvv+aJJ57AYDDUmM+xe/furFu3jmPHjgGwf/9+Nm/ezMCBAwHbf46VtlCeLSQlJWE0GousDuzv78/Ro0dtFFXFiY+PByj2fs37qhuTycSkSZPo0aMHbdu2BfR9Ojk54e3tbXVsdbvPgwcPEhERQXZ2Nu7u7ixdupTWrVsTFRVVI+4PYPHixezdu5ddu3YV2VcTPsdu3bqxcOFCWrRowcWLF3nttdfo2bMnhw4dqhH3B3Dq1Ck+/vhjJk+ezMsvv8yuXbt4/vnncXJyIjIyssb93lm2bBnJycmMHTsWqBn/TgGmTZtGamoqLVu2xN7eHqPRyBtvvMGoUaMA239/1OhkRFR/48eP59ChQ2zevNnWoZS7Fi1aEBUVRUpKCj/88AORkZFs3LjR1mGVm7i4OCZOnMjatWtxcXGxdTgVwvxXJUD79u3p1q0boaGhfP/997i6utowsvJjMpno3Lkz//jHPwDo1KkThw4d4pNPPiEyMtLG0ZW/zz//nIEDB95wyfvq5vvvv2fRokV88803tGnThqioKCZNmkRQUFCV+BxrdDONn58f9vb2RXo9JyQkEBAQYKOoKo75nmrK/U6YMIFffvmF9evX07BhQ8v2gIAAcnNzSU5Otjq+ut2nk5MTTZs2JTw8nDlz5tChQwfefffdGnN/e/bsITExkdtuuw0HBwccHBzYuHEj7733Hg4ODvj7+9eI+yzM29ub5s2bc+LEiRrzOQYGBtK6dWurba1atbI0R9Wk3ztnzpzht99+Y9y4cZZtNeVz/L//+z+mTZvGI488Qrt27Xjsscd44YUXmDNnDmD7z7FGJyNOTk6Eh4ezbt06yzaTycS6deuIiIiwYWQVo1GjRgQEBFjdb2pqKjt27KhW96uUYsKECSxdupTff/+dRo0aWe0PDw/H0dHR6j5jYmI4e/ZstbrPa5lMJnJycmrM/fXt25eDBw8SFRVlKZ07d2bUqFGW5zXhPgtLT0/n5MmTBAYG1pjPsUePHkWG1h87dozQ0FCg5vzeAViwYAH169dn8ODBlm015XPMzMzEzs76K9/e3h6TyQRUgc+xwrvI2tjixYuVs7OzWrhwoTpy5Ih6+umnlbe3t4qPj7d1aDclLS1N7du3T+3bt08B6u2331b79u1TZ86cUUrpoVne3t7qp59+UgcOHFBDhw6tVkPslFLq2WefVV5eXmrDhg1Ww+0yMzMtxzzzzDMqJCRE/f7772r37t0qIiJCRURE2DDqspk2bZrauHGjio2NVQcOHFDTpk1TBoNB/frrr0qp6n9/JSk8mkap6n+fL774otqwYYOKjY1VW7ZsUf369VN+fn4qMTFRKVX9708pPSzbwcFBvfHGG+r48eNq0aJFys3NTX399deWY2rC7x2j0ahCQkLU1KlTi+yrCZ9jZGSkatCggWVo748//qj8/PzUSy+9ZDnGlp9jjU9GlFLq/fffVyEhIcrJyUl17dpVbd++3dYh3bT169croEiJjIxUSunhWbNmzVL+/v7K2dlZ9e3bV8XExNg26DIq7v4AtWDBAssxWVlZ6q9//avy8fFRbm5uavjw4erixYu2C7qMnnjiCRUaGqqcnJxUvXr1VN++fS2JiFLV//5Kcm0yUt3vc8SIESowMFA5OTmpBg0aqBEjRljNv1Hd78/s559/Vm3btlXOzs6qZcuW6tNPP7XaXxN+76xZs0YBxcZdEz7H1NRUNXHiRBUSEqJcXFxU48aN1YwZM1ROTo7lGFt+jgalCk2/JoQQQghRyWp0nxEhhBBCVH2SjAghhBDCpiQZEUIIIYRNSTIihBBCCJuSZEQIIYQQNiXJiBBCCCFsSpIRIYQQQtiUJCNCCCGEsClJRoQQQghhU5KMCCGEEMKmHMp6wh9//MG8efPYs2cPFy9eZOnSpQwbNuy652zYsIHJkydz+PBhgoODmTlzJmPHji31NU0mExcuXMDDwwODwVDWkIUQQghhA0op0tLSCAoKKrJqcGFlTkYyMjLo0KEDTzzxBPfff/8Nj4+NjWXw4ME888wzLFq0iHXr1jFu3DgCAwPp379/qa554cIFgoODyxqqEEIIIaqAuLg4GjZsWOL+W1ooz2Aw3LBmZOrUqaxYsYJDhw5Ztj3yyCMkJyezevXqUl0nJSUFb29v4uLi8PT0vNlwhRBCCFGJUlNTCQ4OJjk5GS8vrxKPK3PNSFlt27aNfv36WW3r378/kyZNKvGcnJwccnJyLK/T0tIA8PT0lGRECCGEqGZu1MWiwjuwxsfH4+/vb7XN39+f1NRUsrKyij1nzpw5eHl5WYo00QghhBA1V5UcTTN9+nRSUlIsJS4uztYhCSGEEKKCVHgzTUBAAAkJCVbbEhIS8PT0xNXVtdhznJ2dcXZ2rujQhBBCCFEFVHgyEhERwcqVK622rV27loiIiIq+tBBCCFE15edDcjK4uIC7e/m//6VLkJ0NTk7WxcEBquAUGWVORtLT0zlx4oTldWxsLFFRUdStW5eQkBCmT5/O+fPn+fLLLwF45pln+OCDD3jppZd44okn+P333/n+++9ZsWJF+d2FEEKIms1kgosX4dw58PKCwEDw9CzbF2tGBmzbBtu3g6Mj1K8P9erpx/r1wc0NTp+GEyfg5MmCx6wsCAkpWkJD9XnXzJ+hFOTnmsg7fAzTth3Y7dyO3YljGJKvYJd8BbuUq9hnpBUc7+ODITS04H0bNICcHLhyBS5f1o9XrqCyc8gJbkpqUEsSfVtxzqMVJx1akJ2eT4P4PTQ4v5PAuJ34n92JR/K5En8M+S51yHdxJ89Zl1wnXep89Ba+vdqW9ZMpF2Ue2rthwwbuuuuuItsjIyNZuHAhY8eO5fTp02zYsMHqnBdeeIEjR47QsGFDZs2aVaZJz1JTU/Hy8iIlJUVG0wghBEBqKsTF6S9n8+OFC5CWpr9009MLipMTdOwInTvr0qEDmJvJU1Jg717Yswd274YjR6BuXf1FGxZWUAID9besyQRGoy7m5yaTdVEKWrXSX9QlycmBn3+GhQvh2DFM3j5kufqS4uhLksmX+Ny6uGZfJSDzJH7JJ/FMOoV9XrbVWxidXcmpG0iOTwCZXkEkuwVxxTmQBPsgLhBEnDEQj5TztEnaSJukDTS9shMHlV+uH0OOwZl4pxDO2YVwWoWSkOdLa+MBurEDH5LL9VrFMaGTMTtUke25OOFCTnGnFevQFztp+3iXco2vtN/ftzTPSGWRZEQIYXNKwdWr+q9ye/uSj7t0SX+x790LZ89CUpLeZn68ckVXzXt5WRcPD/2+dna6mJ/n5urq/JQU/WguGRk3fSsmO3sS/drgmJ+J75UTNz7hJh2078BG53vY7HI3u13uIM/BlTb5+3koYwHD0r7Gx3S5TO+Xjz0XCMKLFLxIvamYzhLMH9xJPg7UJ9GquJDDORpwkiacoKnlMQtXgokjlDOEcJYQzhLKGQK5iD2mEq+ViSu76cxOunHYoQNX7OuRYudDqr0PaQ4+pNl7k5+aSUDeWcv7hnCWBpwnC1euUJfL+FoejdjTgmN0dI6mjX00TfOi8crTP8OkOqGc9O3K8bpdOe7dlWPut5GOO7k5ivwcI+TmonJyscvNxt0uEy/7dDwMBcWddAZ8cC/Nb697Uz/XkkgyIoSonpTSf+HHxBSU5GRo0QLatIG2bfVf6teZWrpcYoiL00mFuezerRMKBwdo2LCgmj40VG/bt08fd67k6vHydhVv4gjmHA2JI5gLBJGMNxnUQX+96OJJKrexl87spgu78CfR6n1iCWMP4ewhnAO0x5NUwjhtVfxJwIQdRuytinmbCTtLcSCfZlgnOVm4cJYQWnDMsu08QfyXSNZyN56kEuh4mWZ1LxPqcZkg5yvkOHty1rEJJ2nC0bwmxGSFkJTiiJ0deNhnEmS4SKAhngB1kUB1gYb2+tEv/yJ1sy/glX6BPFdPzjfvzYWmvTjfrDepdcPAYKBOHetc0NND4UA+p+IciYmBY8cKSkYG+PsXlPr19aOvZx7++efwyzhL3bQzeF49g2v6JVSrVhhuvx37Dm1xdnfE3r7k1iSldEvM+fO6XLgA8fG6xahuXevi66uLQ+EOFpcu6cd69crhX1T5k2REiJrMZIKjR3Xb9/nz+jdUvXrg51fwWExbdrGMRv0l6+d3/b/4zbKz9TmOjvq3YmmTgvx83V7/yy+werWuIXBwKHgfBwf9mzk29sZ/9bu5QevW0LRpwTdEQEDBY8uW+piSZGXBTz+hvl6EOnwYlZun4/uzGHJzsMsufh6k0rjs15wTXuHsS2/OocR6JKp6JOFHEn5coS7O5Pz5131B8SCt0Nd5QcnHgWS8ScabFLwsj/EEkIE77u6664Snp+4Haf5x2tvrYn7u7KyLk6PCP/88jZP3kmvnwjGPcJKUL9nZ+seSk6O/+AICrIuPj/7Y8/Ksi8GgW3xcXXWFj/nR4UoiLpt/w23rWty3/orTpQsAmBwcudJzGPEDHyfptnswGexxdYXGjfU/2SrYt1LcAklGhLCFhASoU+fGveMzMuD4cf1XTZ71FyH5+fo3cuEvaUdH/S2xe7dOQHbs0H0GrsfTE26/HSIioHt36NZN/wmYnQ07d8LmzbBpE2zdqt/L3h6CgiA4WP/lHxyst128qP9cu3BBP09Otr6OnZ2Oz8VF1xY0aqS/WcyPaWk6AVm1SjdzlIa9PapxY3JDm3OlXgtS7H3wOH8UzzOHqBN3FLu867eDmxwcudykK2dD7+RIvV7sc+3O+eQ6NDz1Bz3PfEXfKz/goa7/88vHnkO0ZQ/h7KYzewjnMG3w4SohnCWMMzR1PEMj+zM4G7PYldeBPYQTRUfSsP49FRYGPXroEhGhc0fQH7O5KKU/evOXvPm5UvrHe20x/zMrTf5oc0rpvigxMXDnnTrxFbWCJCNCVIb8fP1l/ssvujPe0aN6u5+f/jI2l7p19V/8x47pX8jnz9/6td3coEsXaNJEf8kX7ptw+bL+AijMYNDHnj2r+yHYQK5HXY43HcTvbvcSa98UZ/t8nOzycTLk4Wyfjx0mjqSHsDOpMcfPOJGdXfQ97MmnCSdpw2FCOYM/CfiTQADx+JNAQ85Rn0tW5+RjzxXqWm0/TShfM5o19CcLV4wGRwyOBTU1CXaBZJhcLX01jUb9cZtK7iJAQIDOvxo31j/qNm10AhIUVF4/QSGqF0lGhCgvSum/7hMTC0pCgq5VWLVKNzfcDD8//S1lrgEpXL9e+M9kc20JQLt2urbj9tv1c6vG40Ly8+HwYZ0omcupU5bdeX4BXG3TkwuN7+Bk4B2c925DoGMSDUxx1M+Jo27mOTyS41D5Rq7WacAl+0DOqyBOZQdxLC2QC5ccuZKQx5WEPJIv5ZGfnYcbmYRyhpaOp2jjFktT+1OEGGNRClab+vNt+r1s53ZMlP5PeYNBj3L09y/4kRSuPTAP3DAP8lAKlEnR1u0UfRz+4PacjbS9+gf10mIByHX1JKHnQ6QOfQyHu3pS188Od3c92KS0NQx5eZCZqZs0zI8Gg+46UqdOqW9NiFpBkhEhbpZScPAgrFihy549FPsnulndujBoEAwZAvfco7edPq1rQszl8mVdQ9KiBTRvrkvdW++1npam3zo727pkZelOcIVDSDuRQP2EA5yiMadoDJRv47yjo/6ivp5mzaBTJz2y1M/POrEwJxeBgQWjSoODdaJwy+LidDLWtWvBkFYhRIWTZETUXsnJuj/GiRP6z93+/XVfievJyoLff9fJxy+/6C+va7m7W0+Q1Lo13Huv7gRQhob7vDzdWnPggK5ksbPTf1kXfszP1y0peXn6MTdXh1i4+8b583oKibKys9N5kJ+f7rvg56dbfC5f1q085mJeONvLS3chMZcGDXTCULi/qL+/fo/0dF1pFB+vHxMSdG7XoQO0b69Hrwohao/Sfn9X+HTwQpS7zEzd76FwMc+aePy47jNRmJOTrrF46CG47z7w9tbbL17UycfPP8Patfrb3szFBfr21clG3776G/h6ozP+DCsxUfcFLTzfVHq6/nI/eFAnIIcPl2+XjcIjGMzF2VnnTYW7rTRqpGsb/PxuPABGKR23wVC2mard3XVp0uSWbkkIUctIMiKqnuxsWLMGDh3Sf1qb+2iYS2n6aAQE6DaBS5d0p9JfftHF0VEnF5cvw65d1uc0bKiTj3vvhbvuKpJ8mEx6ZmjzfFaxsQW1APHxusmktNzddU1BcLB+be7zYO73YG9vvZyEo6NOMAICdDeTBg30Y2BgxdQ2GAxSiyGEqDySjIiqwWiEjRvhm2/ghx/0bJPX4+6uOxaYJ54KCdFzTphL4W/Sw4f1ey5Zop+vXl2wr0sX3ddjyBDo0AGFgatX4dwJPXfVuXO6ScWcgNxoNK2zs654cXcvGHrp7q63tWmjE5AOHXTIFTlnlxBCVCfSZ0SUP6V0U8nx4wXl2DE4c0a3J1w7reClS7B4se4IYdawIfTrp//0N093aH5s2FB3ZLjB7EiZmTqZKDxng8OJaDw3rSRZebEvaDDH0wMtMx+eO6cfMzNLfk9nZ51MhIfrebXME0KZ+06Udd0uIYSoyaTPiKh8OTnwxRfwz3/qPhxl5e2t+3WMGgU9e5a56iAzU0/wuX49bNig5/UqOrqj1Z/l+urVK+iwGRoKt92mS+vWuslECCFE+ZFkRNy6nBz4/HOYM8d6XY6QEN1vw1waNdI9N/9cDttSDAYYOhQGDNBVD6WQl6e7gkRF6bJrl56U9NqOoR4e+i2vnXW8bt2CkSENGhQ8Nz+6uJTbT0cIIcQNSDIibl52tq4JKZyENGgA06fD2LHlMgOU0ag7ih49qkt0NOzfr/u25hQzI3iDBrrv6V13Qe/eOv+RZhMhhKjaJBkRpZefr3txrlun5+TYvLlgMjBzEvLkkzdVraCUzmf279flwAG9lMWxYyUPg/XwgI4ddenUSbfsNGkiyYcQQlQ3koyIAomJOtlISNDDRlJSCh7Pn9fTn187nCQkBF566aaSkPPn4auv9OCWAwdKXkPNxUVPXNqypX5s314nII0ayYgUIYSoCSQZqa2ysvR6Jbt2FZTiZh29lre3bv/o2xf69IFWrcpUFfHnyu0sXKjnGSu86JiDg044OnTQpU0b/fYhIdVkZVIhhBA3RZKR2ujKFT2F+bFj1tsNBp0NhIToobOengWPdevqxdk6dSpzZnD+vM571q6F77+3nkKkZ089eKZbN514lLL/qhBCiBpEkpHaRil46imdiPj66hqOLl2gc2c9dvUW53HJzdXzim3bBlu26HLmjPUxISEQGQljxuj5yYQQQtRukozUNp9+Cj/+qMe5rlmjZ++6Sbm5sHs37Nuny969epTLtXN72NnpZpcePeD++6FXL+nrIYQQooAkI7XJ4cMwaZJ+PnfuTSciRiP8978we7b1tCJm3t66sqVHD126dZN1ToQQQpRMkpHaIisLHnlED8UdMKAgKSkDpfQCt9On62G3oFt6unXTLTydOukSFibDa4UQQpTeTVWWf/jhh4SFheHi4kK3bt3YuXPndY+fP38+LVq0wNXVleDgYF544QWyzfNTiMoxZYpuQ/H310NZythOsmWL7mw6dKhOROrWhX/9S9eMrFgBr7+um2BkkjEhhBBlVeaake+++47JkyfzySef0K1bN+bPn0///v2JiYmhfv36RY7/5ptvmDZtGl988QXdu3fn2LFjjB07FoPBwNtvv10uNyFuYOlS+Ogj/fzLL3VCUgrHjunFbn/4QfcJAb3O3aRJemoRb+8KiVYIIUQtU+ZVe7t160aXLl344IMPADCZTAQHB/Pcc88xbdq0IsdPmDCB6Oho1q1bZ9n24osvsmPHDjZv3lyqa8qqvbfg5EndgePqVfi//9OL2F1HTIwefvvDD3oiMjN7ez2v2ezZEBRUwTELIYSoEUr7/V2muvrc3Fz27NlDv379Ct7Azo5+/fqxbdu2Ys/p3r07e/bssTTlnDp1ipUrVzJo0KASr5OTk0NqaqpVEWVw7hy8954ettKsmU5EunSBv/+9xFMOH4YHH9TTjLzyik5EHBygf3/47DOIj4d//1sSESGEEOWvTM00SUlJGI1G/K+p5vf39+fo0aPFnvPoo4+SlJTEHXfcgVKK/Px8nnnmGV5++eUSrzNnzhxee+21soQmLl/WfUF++AG2b7feFxEBixaBk1OR044fh9deg2++0R1UDQadgDz8sO4fUrdu5YQvhBCi9qrw2R42bNjAP/7xDz766CP27t3Ljz/+yIoVK3j99ddLPGf69OmkpKRYSlxppimvrWJj4bnn9ExiU6boRMRg0GNq334bTp/W0582amR12tmzMG6cnvV00SKdiNx/v64RWbUKHn9cEhEhhBCVo0w1I35+ftjb25OQkGC1PSEhgYCAgGLPmTVrFo899hjjxo0DoF27dmRkZPD0008zY8YM7IoZ1eHs7IyzzAt+fbt3w7x5uibEvMBLp046wxg+HAIDiz1NKd3cMmUKZGTobYMGwd/+dkvznwkhhBA3rUw1I05OToSHh1t1RjWZTKxbt46IiIhiz8nMzCyScNj/ubZJGfvOCtDzhNx3n+4D8v33OhHp3x9++w327IG//rXEROT8eRg4EJ59Vicid9yhh+yuWCGJiBBCCNsp89DeyZMnExkZSefOnenatSvz588nIyODxx9/HIAxY8bQoEED5syZA8CQIUN4++236dSpE926dePEiRPMmjWLIUOGWJISUQYvvaRnHnNwgJEjdRVH+/bXPUUp3SdkwgRITgYXF5gzB55/XqZlF0IIYXtlTkZGjBjBpUuXeOWVV4iPj6djx46sXr3a0qn17NmzVjUhM2fOxGAwMHPmTM6fP0+9evUYMmQIb7zxRvndRW2xYgW8/75+vny5rua4gaQkXRPyww/6defOeqqRVq0qME4hhBCiDMo8z4gtyDwjwMWLugYkKUnPOvbOO9c9XCn47jvdtzUpSVekzJqlp3J3dKyckIUQQtRupf3+lrVpqgOTCcaM0VlFhw56kbvrOHdOdx35+Wf9um1bPepX+oUIIYSoiqTHQHXwr3/pDqqurvDtt1DCSCOTCT79FNq00YmIo6OeQ2TPHklEhBBCVF1SM1LV7d4N5gni3n23xM4e8fG6P+uGDfp1t27w+ec6MRFCCCGqMklGqrL0dHj0UcjPhwce0HOIFCMhAe66C44eBTc3eOMN3VdEBisJIYSoDiQZqaqUgmee0fO1N2yo218MhiKHJSZCnz46EWnYENatg+bNbRCvEEIIcZMkGamq3nhDz9Nub68fi5mb/dIl6NsXjhyBBg1g/Xpo2tQGsQohhBC3QDqwVkXffKPH4QJ88AHceWeRQ5KSdCJy6JCecFUSESGEENWVJCNVzaZNepU60LOrPvNMkUMuX4Z+/eDgQQgI0IlIs2aVHKcQQghRTiQZqUqOHYNhwyA3Vy+h++abRQ5JT4d77oH9+8HfXyciLVpUfqhCCCFEeZFkpKpISoLBg+HKFejaFb76qsjCMUrB2LGwdy/Uqwe//w4tW9omXCGEEKK8SDJSFWRn6xqREycgLEyvO+PmVuSwN96A//0PnJzgp5+gdetKj1QIIYQod5KMVAUTJsCWLeDlpRfD+3PRwcKWLy/o0/rRRxARUckxCiGEEBVEkhFb++orPVWqwQBLlhRb3REdDaNH6+fjx8OTT1ZyjEIIIUQFkmTElo4cKRgt8+qrcPfdRQ65ehWGDoW0NOjV64aL9QohhBDVjiQjtpKRAQ89BJmZepzujBlFDjEa9Wzwx49DaKiuOHF0tEGsQgghRAWSZMQWlIK//lXXjAQGFsy0eo1XXoHVq/VivcuW6RE0QgghRE0jyYgtLFgAX36ph+4uXgz16xc5ZO9emDu34PCOHSs3RCGEEKKySDJS2Q4e1L1QAf7+92Knes/Ph6efBpMJRo6EESMqOUYhhBCiEkkyUpmysnQ/kexsGDgQpk4t9rAPP4Q9e8DbG95+u3JDFEIIISqbJCOVaelSiInR/UTMzTTXiIuDmTP18zff1GvPCCGEEDWZJCOVadky/Th2LPj5FXvI88/r9Wd69IBx4yotMiGEEMJmJBmpLNnZsGqVfj58eLGHLFumi4MD/PvfxVacCCGEEDXOTX3dffjhh4SFheHi4kK3bt3YuXPndY9PTk5m/PjxBAYG4uzsTPPmzVm5cuVNBVxtrVunqzwaNIDw8CK709Lguef085degjZtKjk+IYQQwkYcynrCd999x+TJk/nkk0/o1q0b8+fPp3///sTExFC/mCGqubm53H333dSvX58ffviBBg0acObMGby9vcsj/upj6VL9OGxYsVUes2bBuXPQuHFBnxEhhBCiNjAopVRZTujWrRtdunThgw8+AMBkMhEcHMxzzz3HtGnTihz/ySefMG/ePI4ePYrjTU4fmpqaipeXFykpKXh6et7Ue9iU0ag7rV66BGvX6hlXC9mzB7p21UN516yBe+6xUZxCCCFEOSrt93eZmmlyc3PZs2cP/Qp9mdrZ2dGvXz+2bdtW7DnLly8nIiKC8ePH4+/vT9u2bfnHP/6B0Wgs8To5OTmkpqZalWpt61adiHh76wVmClFKL9prnlNEEhEhhBC1TZmSkaSkJIxGI/7XLHHv7+9PfHx8seecOnWKH374AaPRyMqVK5k1axb/+te/+Pvf/17idebMmYOXl5elBAcHlyXMqsc8iubee4ssLvPNN7B9O9SpA2+9VfmhCSGEELZW4eM1TCYT9evX59NPPyU8PJwRI0YwY8YMPvnkkxLPmT59OikpKZYSFxdX0WFWHKUK+otcM4omI6Ng3rMZMyAoqJJjE0IIIaqAMnVg9fPzw97enoSEBKvtCQkJBJQwO1dgYCCOjo7YF1oIrlWrVsTHx5Obm4uTk1ORc5ydnXF2di5LaFXXwYMQGwsuLtC/v9WuN9+E8+ehUSN44QUbxSeEEELYWJlqRpycnAgPD2fdunWWbSaTiXXr1hEREVHsOT169ODEiROYTCbLtmPHjhEYGFhsIlLjmGtF7rlHt8X86cwZmDdPP583T+cqQgghRG1U5maayZMn89lnn/Hf//6X6Ohonn32WTIyMnj88ccBGDNmDNOnT7cc/+yzz3LlyhUmTpzIsWPHWLFiBf/4xz8Yb14srqYz9xcZNsxq80sv6XnQeveG+++v7KCEEEKIqqPM84yMGDGCS5cu8corrxAfH0/Hjh1ZvXq1pVPr2bNnsSs0j0ZwcDBr1qzhhRdeoH379jRo0ICJEycytYRF4mqU2FiIitLzigwZYtn8xx/w/fd68/z5YDDYLEIhhBDC5so8z4gtVNt5RubP151BevWCDRsAPeVIly6wbx/85S9wnX68QgghRLVWIfOMiDIqPOvqnxYs0ImIlxe8/rptwhJCCCGqEklGKsqlS7B5s37+ZzKSmqqH8ALMng316tkmNCGEEKIqkWSkovz8s55WtWNHCAsD4L//hcREaNoUakv/XSGEEOJGJBmpKOZRNH9OdKYUfPSR3jRpEtSGUc1CCCFEaUgyUhEuXdIr3oGliWb9ejh6FNzd4bHHbBeaEEIIUdVIMlIRPv8ccnOhc2do3x4oqBV57DGoTgOChBBCiIomyUh5MxoLxuv+2THk/PmCVptnn7VNWEIIIURVJclIeVu5Us/1XrcujBgBwGef6RylZ09o187G8QkhhBBVjCQj5e3DD/XjE0+Aqyt5efDpp3qTjKARQgghipJkpDwdP647rhoMlvaYZcvg4kXw97cMrBFCCCFEIZKMlKePP9aPAwdC48ZAQcfVp56S4bxCCCFEcSQZKS+ZmXqud7C0xxw5opeksbfX69AIIYQQoihJRsrLt99CcjI0agT9+wMFtSL33QcNG9ouNCGEEKIqk2SkPChV0HH12WfB3p60NPjyS73pr3+1XWhCCCFEVSfJSHnYvl0vxeviokfRAIsWQVoaNG8OffrYOD4hhBCiCpNkpDyY22MeeQR8fYGCvqx//SvYyU9ZCCGEKJF8Td6qxET4/nv9/M+Oq0ePwoED4OgIY8bYMDYhhBCiGpBk5FZ98YVeh6ZLF70WDbB0qd7Vty/4+NgwNiGEEKIakGTkVv30k3586inLJnMyIpOcCSGEEDcmycitSE+HXbv087vvBiAuTm8yGGDoUBvGJoQQQlQTkozcis2b9Qp4YWG6ULA6b48eegp4IYQQQlyfJCO3YsMG/di7t2WTNNEIIYQQZXNTyciHH35IWFgYLi4udOvWjZ07d5bqvMWLF2MwGBg2bNjNXLbqMScjd90FQFISbNyoN0kyIoQQQpROmZOR7777jsmTJzN79mz27t1Lhw4d6N+/P4mJidc97/Tp00yZMoWePXvedLBVSloa7N6tn/fqBcDPP4PJBB076lnhhRBCCHFjZU5G3n77bZ566ikef/xxWrduzSeffIKbmxtffPFFiecYjUZGjRrFa6+9RuM/V7Ot9sz9RRo1gtBQoKCJ5v77bRiXEEIIUc2UKRnJzc1lz5499OvXr+AN7Ozo168f27ZtK/G8v/3tb9SvX58nn3yyVNfJyckhNTXVqlQ51zTRpKXBr7/qTdJEI4QQQpRemZKRpKQkjEYj/tcME/H39yc+Pr7YczZv3sznn3/OZ599VurrzJkzBy8vL0sJDg4uS5iVY/16/fhn59XVqyEnB5o2hTZtbBeWEEIIUd1U6GiatLQ0HnvsMT777DP8/PxKfd706dNJSUmxlLi4uAqM8iakpsKePfr5n8nIjz/ql/ffr+cYEUIIIUTpOJTlYD8/P+zt7UlISLDanpCQQEBAQJHjT548yenTpxkyZIhlm8lk0hd2cCAmJoYmTZoUOc/Z2RlnZ+eyhFa5Nm/WPVWbNIHgYHJyYMUKvUuaaIQQQoiyKVPNiJOTE+Hh4axbt86yzWQysW7dOiIiIooc37JlSw4ePEhUVJSl3Hfffdx1111ERUVVzeaX0rimieb333WfkaAg6NrVdmEJIYQQ1VGZakYAJk+eTGRkJJ07d6Zr167Mnz+fjIwMHn/8cQDGjBlDgwYNmDNnDi4uLrRt29bqfG9vb4Ai26uVazqvmptohg0DO5lGTgghhCiTMicjI0aM4NKlS7zyyivEx8fTsWNHVq9ebenUevbsWexq8jdySgrs3auf9+qF0ViwVp400QghhBBlZ1BKKVsHcSOpqal4eXmRkpKCp6enbYNZsQLuvVcPmzl+nE2b4M47wccHEhLA0dG24QkhhBBVRWm/v2twFUYFMfcXuaaJZsgQSUSEEEKImyHJSFkVWhxPqYJVeqWJRgghhLg5koyURXIy7Nunn/fuzcGDcPo0uLrCPffYMjAhhBCi+pJkpCw2bdLzizRvDkFBllqRe+4BNzebRiaEEEJUW5KMlEWhJhooaKIZNswGsQghhBA1RJmH9tZqheYXOXNGt9jY2enBNUIIURKj0UheXp6twxCi3Dk6OmJvb3/L7yPJSGldvVrQX6RXL35aop/ecQeUYdkdIUQtopQiPj6e5ORkW4ciRIXx9vYmICAAwy0szCbJSGlt3AhKQcuWEBhomehMmmiEECUxJyL169fHzc3tln5ZC1HVKKXIzMwkMTERgMDAwJt+L0lGSmvpUv3Yvz9XrujcBGDoUNuFJISouoxGoyUR8fX1tXU4QlQIV1dXABITE6lfv/5NN9lIB9bSyM0tmPP9wQdZsQKMRmjfHho3tm1oQoiqydxHxE2G2okazvxv/Fb6RUkyUhrr1uk1aQIDoXt3GUUjhCg1aZoRNV15/BuXZKQ0fvhBP95/P1k5dqxerV9KMiKEEELcOklGbiQvr2BCkQcfZN06yMyEkBDo2NGWgQkhRPUQFhbG/PnzS338hg0bMBgMMgqpFpFk5EbWr4crV6B+fejZ05KXDB0KUvsqhKhJDAbDdcurr756U++7a9cunn766VIf3717dy5evIiXl9dNXe9mtGzZEmdnZ+Lj4yvtmqKAJCM3UqiJxog9y5frl9JEI4SoaS5evGgp8+fPx9PT02rblClTLMcqpcjPzy/V+9arV69MHXmdnJxued6Ksti8eTNZWVk8+OCD/Pe//62Ua15PbZwgT5KR68nPLxjS++CDbNsGly6Bjw/07Gnb0IQQorwFBARYipeXFwaDwfL66NGjeHh4sGrVKsLDw3F2dmbz5s2cPHmSoUOH4u/vj7u7O126dOG3336zet9rm2kMBgP/+c9/GD58OG5ubjRr1ozl5r/0KNpMs3DhQry9vVmzZg2tWrXC3d2dAQMGcPHiRcs5+fn5PP/883h7e+Pr68vUqVOJjIxkWCn+cvz888959NFHeeyxx/jiiy+K7D937hwjR46kbt261KlTh86dO7Njxw7L/p9//pkuXbrg4uKCn58fwwst424wGFhmrlL/k7e3NwsXLgTg9OnTGAwGvvvuO3r16oWLiwuLFi3i8uXLjBw5kgYNGuDm5ka7du349ttvrd7HZDLxz3/+k6ZNm+Ls7ExISAhvvPEGAH369GHChAlWx1+6dAknJyfWrVt3w59JZZNk5Hr++AOSksDXF3r1sjTR3HsvODraNDIhRDWkFGRkVH5RqvzuYdq0acydO5fo6Gjat29Peno6gwYNYt26dezbt48BAwYwZMgQzp49e933ee2113j44Yc5cOAAgwYNYtSoUVy5cqXE4zMzM3nrrbf46quv+OOPPzh79qxVTc2bb77JokWLWLBgAVu2bCE1NbVIElCctLQ0lixZwujRo7n77rtJSUlh06ZNlv3p6en06tWL8+fPs3z5cvbv389LL72EyWQCYMWKFQwfPpxBgwaxb98+1q1bR9euXW943WtNmzaNiRMnEh0dTf/+/cnOziY8PJwVK1Zw6NAhnn76aR577DF27txpOWf69OnMnTuXWbNmceTIEb755hv8/f0BGDduHN988w05OTmW47/++msaNGhAnz59yhxfhVPVQEpKigJUSkpK5V742WeVAqXGjVMmk1JNmuiXP/xQuWEIIaqfrKwsdeTIEZWVlWXZlp6uf4dUdklPL3v8CxYsUF5eXpbX69evV4BatmzZDc9t06aNev/99y2vQ0ND1TvvvGN5DaiZM2cW+rmkK0CtWrXK6lpXr161xAKoEydOWM758MMPlb+/v+W1v7+/mjdvnuV1fn6+CgkJUUOHDr1urJ9++qnq2LGj5fXEiRNVZGSk5fW///1v5eHhoS5fvlzs+REREWrUqFElvj+gli5darXNy8tLLViwQCmlVGxsrALU/PnzrxunUkoNHjxYvfjii0oppVJTU5Wzs7P67LPPij02KytL+fj4qO+++86yrX379urVV1+94XXKqrh/62al/f6WmpGSGI3w44/6+YMPcuQInDwJzs7Qv79tQxNCCFvp3Lmz1ev09HSmTJlCq1at8Pb2xt3dnejo6BvWjLRv397yvE6dOnh6elqmFS+Om5sbTZo0sbwODAy0HJ+SkkJCQoJVjYS9vT3h4eE3vJ8vvviC0aNHW16PHj2aJUuWkJaWBkBUVBSdOnWibt26xZ4fFRVF3759b3idG7n252o0Gnn99ddp164ddevWxd3dnTVr1lh+rtHR0eTk5JR4bRcXF6tmp71793Lo0CHGjh17y7FWBJkOviRbtkBCgu4g0qcP3/9db777bnB3t21oQojqyc0N0tNtc93yUqdOHavXU6ZMYe3atbz11ls0bdoUV1dXHnzwQXJzc6/7Po7XtHUbDAZL00dpj1e32P505MgRtm/fzs6dO5k6daplu9FoZPHixTz11FOW6c5LcqP9xcVZXAfVa3+u8+bN491332X+/Pm0a9eOOnXqMGnSJMvP9UbXBd1U07FjR86dO8eCBQvo06cPoaGhNzzPFqRmpCTmUTRDh6IcHDH3G3rkEduFJISo3gwGqFOn8ktFDkrZsmULY8eOZfjw4bRr146AgABOnz5dcRcshpeXF/7+/uzatcuyzWg0snfv3uue9/nnn3PnnXeyf/9+oqKiLGXy5Ml8/vnngK7BiYqKKrE/S/v27a/bIbRevXpWHW2PHz9OZmbmDe9py5YtDB06lNGjR9OhQwcaN27MsWPHLPubNWuGq6vrda/drl07OnfuzGeffcY333zDE088ccPr2ookI8UxmeB//9PPH3yQvXvh+HFwdZWF8YQQorBmzZrx448/EhUVxf79+3n00UevW8NRUZ577jnmzJnDTz/9RExMDBMnTuTq1aslDg/Oy8vjq6++YuTIkbRt29aqjBs3jh07dnD48GFGjhxJQEAAw4YNY8uWLZw6dYr//e9/bNu2DYDZs2fz7bffMnv2bKKjozl48CBvvvmm5Tp9+vThgw8+YN++fezevZtnnnmmSC1PcZo1a8batWvZunUr0dHR/OUvfyEhIcGy38XFhalTp/LSSy/x5ZdfcvLkSbZv325JoszGjRvH3LlzUUpZjfKpam4qGfnwww8JCwvDxcWFbt26WfXuvdZnn31Gz5498fHxwcfHh379+l33+Cph2za4cAE8PaFfP0utyJAh0kQjhBCFvf322/j4+NC9e3eGDBlC//79ue222yo9jqlTpzJy5EjGjBlDREQE7u7u9O/fHxcXl2KPX758OZcvXy72C7pVq1a0atWKzz//HCcnJ3799Vfq16/PoEGDaNeuHXPnzrWsTtu7d2+WLFnC8uXL6dixI3369LH6jvvXv/5FcHAwPXv25NFHH2XKlCmlmnNl5syZ3HbbbfTv35/evXtbEqLCZs2axYsvvsgrr7xCq1atGDFiRJF+NyNHjsTBwYGRI0eW+LOoEsraa3bx4sXKyclJffHFF+rw4cPqqaeeUt7e3iohIaHY4x999FH14Ycfqn379qno6Gg1duxY5eXlpc6dO1fqa1b6aJpJk3QX9NGjldGoVMOG+uU1HaKFEKJE1xthICqe0WhUzZs3txq1UxvFxsYqOzs7tWfPngq7RnmMpjEoVbYeQN26daNLly588MEHgJ50JTg4mOeee45p06bd8Hyj0YiPjw8ffPABY8aMKdU1U1NT8fLyIiUlBU9Pz7KEW3ZKQWgoxMXBsmX84TOUXr3Ay0v3Z3V2rtjLCyFqhuzsbGJjY2nUqFHV/ou0hjhz5gy//vorvXr1Iicnhw8++IAFCxawf/9+WrVqZevwKl1eXh6XL19mypQpxMbGsmXLlgq71vX+rZf2+7tMzTS5ubns2bOHfv36FbyBnR39+vWztJ/dSGZmJnl5eSUOk7K5HTt0IlKnDtxzj6WJ5v77JRERQoiqys7OjoULF9KlSxd69OjBwYMH+e2332plIgK6A2xgYCC7du3ik08+sXU4N1Smob1JSUkYjUbLDG9m/v7+HD16tFTvMXXqVIKCgqwSmmvl5ORYzRqXmppaljBvjXkUzZAh5Dm4smSJfjlyZOWFIIQQomyCg4Mr9K//6qZ37963PPS5MlXqaJq5c+eyePFili5det1qyzlz5uDl5WUpwcHBlROgUgXJyEMPsXYtXL6sF+y9667KCUEIIYSobcqUjPj5+WFvb281vAggISGBgICA65771ltvMXfuXH799VermfeKM336dFJSUiwlLi6uLGHevF274MwZ3UQzcKCliebhh8FBpocTQgghKkSZkhEnJyfCw8OtJlkxmUysW7eOiIiIEs/75z//yeuvv87q1auLTHlbHGdnZzw9Pa1KpTC3ydx7L5nK1bIw3qOPVs7lhRBCiNqozH/vT548mcjISDp37kzXrl2ZP38+GRkZPP744wCMGTOGBg0aMGfOHECvpPjKK6/wzTffEBYWRnx8PADu7u64V6VJOwo30Tz4ICtW6Gmbw8Lg9tttGpkQQghRo5U5GRkxYgSXLl3ilVdeIT4+no4dO7J69WpLp9azZ89iZ1dQ4fLxxx+Tm5vLgw8+aPU+s2fP5tVXX7216MvTnj1w+rRexGHQIL79c92kRx6p2KmUhRBCiNrupnpCTJgwgQkTJhS7b8OGDVavK3uNgptmbqIZPJiUPDdWrtQvZRSNEEIIUbFkbRrQTTTmZOShh1i6FHJyoHVraNfOtqEJIUR107t3byZNmmR5HRYWxvz58697jsFgYJm5o94tKK/3EZVLkhGAffsgNlavhDdokGUUzciR0kQjhKg9hgwZwoABA4rdt2nTJgwGAwcOHCjz++7atYunn376VsOz8uqrr9KxY8ci2y9evMjAgQPL9VolycrKom7duvj5+VnNjSXKTpIRKKgVGTSIhPQ6/PabfvnII7YLSQghKtuTTz7J2rVrOXfuXJF9CxYsoHPnzjecmqE49erVK9XicOUhICAA50qaLvt///sfbdq0oWXLljavjVFKkZ+fb9MYboUkI9c00Xz3HZhM0LUrNG1q29CEEKIy3XvvvdSrV4+FCxdabU9PT2fJkiU8+eSTXL58mZEjR9KgQQPc3Nxo164d35qrk0twbTPN8ePHufPOO3FxcaF169asXbu2yDlTp06lefPmuLm50bhxY2bNmkVeXh4ACxcu5LXXXmP//v0YDAYMBoMl5mubaQ4ePEifPn1wdXXF19eXp59+mvT0dMv+sWPHMmzYMN566y0CAwPx9fVl/Pjxlmtdz+eff87o0aMZPXo0n3/+eZH9hw8f5t5778XT0xMPDw969uzJyZMnLfu/+OIL2rRpg7OzM4GBgZa+mKdPn8ZgMBAVFWU5Njk5GYPBYOmXuWHDBgwGA6tWrSI8PBxnZ2c2b97MyZMnGTp0KP7+/ri7u9OlSxd+M/+F/aecnBymTp1KcHAwzs7ONG3alM8//xylFE2bNuWtt96yOj4qKgqDwcCJEydu+DO5WTKVV1QUnDwJLi4weDBf99GbR4+2aVRCiJpIKcjMrPzrurmVqs3ZwcGBMWPGsHDhQmbMmIHhz3OWLFmC0Whk5MiRpKenEx4eztSpU/H09GTFihU89thjNGnShK5du97wGiaTifvvvx9/f3927NhBSkqKVf8SMw8PDxYuXEhQUBAHDx7kqaeewsPDg5deeokRI0Zw6NAhVq9ebfmi9fLyKvIeGRkZ9O/fn4iICHbt2kViYiLjxo1jwoQJVgnX+vXrCQwMZP369Zw4cYIRI0bQsWNHnnrqqRLv4+TJk2zbto0ff/wRpRQvvPACZ86cITQ0FIDz589z55130rt3b37//Xc8PT3ZsmWLpfbi448/ZvLkycydO5eBAweSkpJyU9PZT5s2jbfeeovGjRvj4+NDXFwcgwYN4o033sDZ2Zkvv/ySIUOGEBMTQ0hICKCn4Ni2bRvvvfceHTp0IDY2lqSkJAwGA0888QQLFixgypQplmssWLCAO++8k6YV+Rd6+S8mXP5KuwTxTZk+XSlQ6v77VUyMfmpvr1RCQvlfSghRexS7rHp6uv4lU9klPb3UcUdHRytArV+/3rKtZ8+eavTo0SWeM3jwYPXiiy9aXvfq1UtNnDjR8jo0NFS98847Siml1qxZoxwcHNT58+ct+1etWqUAtXTp0hKvMW/ePBUeHm55PXv2bNWhQ4cixxV+n08//VT5+Pio9EL3v2LFCmVnZ6fi4+OVUkpFRkaq0NBQlZ+fbznmoYceUiNGjCgxFqWUevnll9WwYcMsr4cOHapmz55teT19+nTVqFEjlZubW+z5QUFBasaMGcXui42NVYDat2+fZdvVq1etPpf169crQC1btuy6cSqlVJs2bdT777+vlFIqJiZGAWrt2rXFHnv+/Hllb2+vduzYoZRSKjc3V/n5+amFCxeW+P7F/lv/U2m/v2t3M801TTSLFumn99yj16MRQojapmXLlnTv3p0vvvgCgBMnTrBp0yaefPJJAIxGI6+//jrt2rWjbt26uLu7s2bNGs6ePVuq94+OjiY4OJigoCDLtuJm8P7uu+/o0aMHAQEBuLu7M3PmzFJfo/C1OnToQJ06dSzbevTogclkIiYmxrKtTZs22NvbW14HBgaSmJhY4vsajUb++9//MrpQFfro0aNZuHAhJpMJ0E0bPXv2xNHRscj5iYmJXLhwgb59+5bpfopz7azm6enpTJkyhVatWuHt7Y27uzvR0dGWn11UVBT29vb06tWr2PcLCgpi8ODBls//559/Jicnh4ceeuiWY72e2t1Mc+AAnDgBLi6oQYNZNFNvHjXKtmEJIWooNzc9tbMtrlsGTz75JM899xwffvghCxYsoEmTJpYvr3nz5vHuu+8yf/582rVrR506dZg0aRK5ubnlFu62bdsYNWoUr732Gv3798fLy4vFixfzr3/9q9yuUdi1CYPBYLAkFcVZs2YN58+fZ8SIEVbbjUYj69at4+6778bV1bXE86+3D7BMHKoKrbpbUh+WwokWwJQpU1i7di1vvfUWTZs2xdXVlQcffNDy+dzo2gDjxo3jscce45133mHBggWMGDGiwjsg1+6aEXOtyIAB7DjiwcmTeo28YcNsGpUQoqYyGPQvmcouZZyj4OGHH8bOzo5vvvmGL7/8kieeeMLSf2TLli0MHTqU0aNH06FDBxo3bsyxY8dK/d6tWrUiLi6OixcvWrZt377d6pitW7cSGhrKjBkz6Ny5M82aNePMmTNWxzg5OWE0Gm94rf3795ORkWHZtmXLFuzs7GjRokWpY77W559/ziOPPEJUVJRVeeSRRywdWdu3b8+mTZuKTSI8PDwICwuzWuetsHr16gFY/YwKd2a9ni1btjB27FiGDx9Ou3btCAgIsJp8tF27dphMJjZu3FjiewwaNIg6derw8ccfs3r1ap544olSXftW1N5kpIQmmmHD9P9dIYSordzd3RkxYgTTp0/n4sWLjB071rKvWbNmrF27lq1btxIdHc1f/vKXIiu5X0+/fv1o3rw5kZGR7N+/n02bNjFjxgyrY5o1a8bZs2dZvHgxJ0+e5L333mPp0qVWx4SFhREbG0tUVBRJSUnFzvMxatQoXFxciIyM5NChQ6xfv57nnnuOxx57zLKESVldunSJn3/+mcjISNq2bWtVxowZw7Jly7hy5QoTJkwgNTWVRx55hN27d3P8+HG++uorS/PQq6++yr/+9S/ee+89jh8/zt69e3n//fcBXXtx++23M3fuXKKjo9m4cSMzZ84sVXzNmjXjxx9/JCoqiv379/Poo49a1fKEhYURGRnJE088wbJly4iNjWXDhg18//33lmPs7e0ZO3Ys06dPp1mzZtddCLe81N5kBOBf/4KxY8nrfy/ffac3ySgaIYTQTTVXr16lf//+Vv07Zs6cyW233Ub//v3p3bs3AQEBDCtDdbKdnR1Lly4lKyuLrl27Mm7cON544w2rY+677z5eeOEFJkyYQMeOHdm6dSuzZs2yOuaBBx5gwIAB3HXXXdSrV6/Y4cVubm6sWbOGK1eu0KVLFx588EH69u3LBx98ULYfRiFffvklderUKba/R9++fXF1deXrr7/G19eX33//nfT0dHr16kV4eDifffaZpUkoMjKS+fPn89FHH9GmTRvuvfdejh8/bnmvL774gvz8fMLDw5k0aRJ///vfSxXf22+/jY+PD927d2fIkCH079+f2267zeqYjz/+mAcffJC//vWvtGzZkqeeesqq9gj055+bm2tZBLeiGVThRqkqKjU1FS8vL1JSUvD09Cz391+5EgYP1p1Wz58Hh9rdk0YIUQ6ys7OJjY2lUaNGuLi42DocIcpk06ZN9O3bl7i4uBvWIl3v33ppv7/laxcsTTSPPCKJiBBCiNorJyeHS5cu8eqrr/LQQw/ddHNWWdXuZhp0x3bzZH0yikYIIURt9u233xIaGkpycjL//Oc/K+26tT4ZWbZMT4jYrBl06WLraIQQQgjbGTt2LEajkT179tCgQYNKu26tT0a+/lo/jholK/QKIYQQtlCrk5GEBDCvzyRNNEIIIYRt1OpkZPFivUJvt26yQq8QomJcbyZPIWqC8vg3XqvHjphH0cjcIkKI8ubk5ISdnR0XLlygXr16ODk5WWYxFaImUEqRm5vLpUuXsLOzw8nJ6abfq9YmIyYTDB8Oubnw8MO2jkYIUdPY2dnRqFEjLl68yIULF2wdjhAVxs3NjZCQEMuaOjdDJj0TQogKpJQiPz//huuoCFEd2dvb4+DgUGKtn0x6JoQQVYDBYMDR0bHYpeSFEFqt7sAqhBBCCNuTZEQIIYQQNiXJiBBCCCFsqlr0GTH3sU1NTbVxJEIIIYQoLfP39o3GylSLZCQtLQ2A4OBgG0cihBBCiLJKS0vDy8urxP3VYmivyWTiwoULeHh43PSkQampqQQHBxMXF1djhwfLPdYMco81g9xjzSD3eGuUUqSlpREUFHTdeUiqRc2InZ0dDRs2LJf38vT0rLH/oMzkHmsGuceaQe6xZpB7vHnXqxExkw6sQgghhLApSUaEEEIIYVO1JhlxdnZm9uzZODs72zqUCiP3WDPIPdYMco81g9xj5agWHViFEEIIUXPVmpoRIYQQQlRNkowIIYQQwqYkGRFCCCGETUkyIoQQQgibqhXJyIcffkhYWBguLi5069aNnTt32jqkm/bHH38wZMgQgoKCMBgMLFu2zGq/UopXXnmFwMBAXF1d6devH8ePH7dNsDdpzpw5dOnSBQ8PD+rXr8+wYcOIiYmxOiY7O5vx48fj6+uLu7s7DzzwAAkJCTaKuOw+/vhj2rdvb5lkKCIiglWrVln2V/f7K87cuXMxGAxMmjTJsq263+err76KwWCwKi1btrTsr+73Z3b+/HlGjx6Nr68vrq6utGvXjt27d1v2V/ffO2FhYUU+R4PBwPjx44Ga8TkajUZmzZpFo0aNcHV1pUmTJrz++utWa8bY9HNUNdzixYuVk5OT+uKLL9Thw4fVU089pby9vVVCQoKtQ7spK1euVDNmzFA//vijAtTSpUut9s+dO1d5eXmpZcuWqf3796v77rtPNWrUSGVlZdkm4JvQv39/tWDBAnXo0CEVFRWlBg0apEJCQlR6errlmGeeeUYFBwerdevWqd27d6vbb79dde/e3YZRl83y5cvVihUr1LFjx1RMTIx6+eWXlaOjozp06JBSqvrf37V27typwsLCVPv27dXEiRMt26v7fc6ePVu1adNGXbx40VIuXbpk2V/d708ppa5cuaJCQ0PV2LFj1Y4dO9SpU6fUmjVr1IkTJyzHVPffO4mJiVaf4dq1axWg1q9fr5SqGZ/jG2+8oXx9fdUvv/yiYmNj1ZIlS5S7u7t69913LcfY8nOs8clI165d1fjx4y2vjUajCgoKUnPmzLFhVOXj2mTEZDKpgIAANW/ePMu25ORk5ezsrL799lsbRFg+EhMTFaA2btyolNL35OjoqJYsWWI5Jjo6WgFq27Zttgrzlvn4+Kj//Oc/Ne7+0tLSVLNmzdTatWtVr169LMlITbjP2bNnqw4dOhS7rybcn1JKTZ06Vd1xxx0l7q+Jv3cmTpyomjRpokwmU435HAcPHqyeeOIJq23333+/GjVqlFLK9p9jjW6myc3NZc+ePfTr18+yzc7Ojn79+rFt2zYbRlYxYmNjiY+Pt7pfLy8vunXrVq3vNyUlBYC6desCsGfPHvLy8qzus2XLloSEhFTL+zQajSxevJiMjAwiIiJq3P2NHz+ewYMHW90P1JzP8fjx4wQFBdG4cWNGjRrF2bNngZpzf8uXL6dz58489NBD1K9fn06dOvHZZ59Z9te03zu5ubl8/fXXPPHEExgMhhrzOXbv3p1169Zx7NgxAPbv38/mzZsZOHAgYPvPsVoslHezkpKSMBqN+Pv7W2339/fn6NGjNoqq4sTHxwMUe7/mfdWNyWRi0qRJ9OjRg7Zt2wL6Pp2cnPD29rY6trrd58GDB4mIiCA7Oxt3d3eWLl1K69atiYqKqhH3B7B48WL27t3Lrl27iuyrCZ9jt27dWLhwIS1atODixYu89tpr9OzZk0OHDtWI+wM4deoUH3/8MZMnT+bll19m165dPP/88zg5OREZGVnjfu8sW7aM5ORkxo4dC9SMf6cA06ZNIzU1lZYtW2Jvb4/RaOSNN95g1KhRgO2/P2p0MiKqv/Hjx3Po0CE2b95s61DKXYsWLYiKiiIlJYUffviByMhINm7caOuwyk1cXBwTJ05k7dq1uLi42DqcCmH+qxKgffv2dOvWjdDQUL7//ntcXV1tGFn5MZlMdO7cmX/84x8AdOrUiUOHDvHJJ58QGRlp4+jK3+eff87AgQMJCgqydSjl6vvvv2fRokV88803tGnThqioKCZNmkRQUFCV+BxrdDONn58f9vb2RXo9JyQkEBAQYKOoKo75nmrK/U6YMIFffvmF9evX07BhQ8v2gIAAcnNzSU5Otjq+ut2nk5MTTZs2JTw8nDlz5tChQwfefffdGnN/e/bsITExkdtuuw0HBwccHBzYuHEj7733Hg4ODvj7+9eI+yzM29ub5s2bc+LEiRrzOQYGBtK6dWurba1atbI0R9Wk3ztnzpzht99+Y9y4cZZtNeVz/L//+z+mTZvGI488Qrt27Xjsscd44YUXmDNnDmD7z7FGJyNOTk6Eh4ezbt06yzaTycS6deuIiIiwYWQVo1GjRgQEBFjdb2pqKjt27KhW96uUYsKECSxdupTff/+dRo0aWe0PDw/H0dHR6j5jYmI4e/ZstbrPa5lMJnJycmrM/fXt25eDBw8SFRVlKZ07d2bUqFGW5zXhPgtLT0/n5MmTBAYG1pjPsUePHkWG1h87dozQ0FCg5vzeAViwYAH169dn8ODBlm015XPMzMzEzs76K9/e3h6TyQRUgc+xwrvI2tjixYuVs7OzWrhwoTpy5Ih6+umnlbe3t4qPj7d1aDclLS1N7du3T+3bt08B6u2331b79u1TZ86cUUrpoVne3t7qp59+UgcOHFBDhw6tVkPslFLq2WefVV5eXmrDhg1Ww+0yMzMtxzzzzDMqJCRE/f7772r37t0qIiJCRURE2DDqspk2bZrauHGjio2NVQcOHFDTpk1TBoNB/frrr0qp6n9/JSk8mkap6n+fL774otqwYYOKjY1VW7ZsUf369VN+fn4qMTFRKVX9708pPSzbwcFBvfHGG+r48eNq0aJFys3NTX399deWY2rC7x2j0ahCQkLU1KlTi+yrCZ9jZGSkatCggWVo748//qj8/PzUSy+9ZDnGlp9jjU9GlFLq/fffVyEhIcrJyUl17dpVbd++3dYh3bT169croEiJjIxUSunhWbNmzVL+/v7K2dlZ9e3bV8XExNg26DIq7v4AtWDBAssxWVlZ6q9//avy8fFRbm5uavjw4erixYu2C7qMnnjiCRUaGqqcnJxUvXr1VN++fS2JiFLV//5Kcm0yUt3vc8SIESowMFA5OTmpBg0aqBEjRljNv1Hd78/s559/Vm3btlXOzs6qZcuW6tNPP7XaXxN+76xZs0YBxcZdEz7H1NRUNXHiRBUSEqJcXFxU48aN1YwZM1ROTo7lGFt+jgalCk2/JoQQQghRyWp0nxEhhBBCVH2SjAghhBDCpiQZEUIIIYRNSTIihBBCCJuSZEQIIYQQNiXJiBD/324dCwAAAAAM8rcexp6iCICVjAAAKxkBAFYyAgCsZAQAWMkIALCSEQBgFZ0kXT241FAEAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(2)\n",
    "history = pd.DataFrame(history)\n",
    "\n",
    "history = pd.concat([pd.read_csv(\"history-kan-70.csv\"), history]).reset_index(drop=True)\n",
    "\n",
    "fig.suptitle('Training History')\n",
    "print(ax[0])\n",
    "\n",
    "ax[0].plot(range(1, len(history)+1), history[\"train_loss\"], color = 'b', label = \"Training Loss\")\n",
    "ax[0].plot(range(1, len(history)+1), history[\"val_loss\"], color = 'r', label = \"Validation Loss\")\n",
    "\n",
    "ax[1].plot(range(1, len(history)+1), history[\"train_acc\"], color = 'b', label = \"Training Accuracy\")\n",
    "ax[1].plot(range(1, len(history)+1), history[\"val_acc\"], color = 'r', label = \"Validation Accuracy\")\n",
    "\n",
    "ax[0].legend()\n",
    "ax[1].legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "d8c65b1b",
   "metadata": {},
   "outputs": [],
   "source": [
    "history.to_csv(f\"history-kan-{prev_epoch+epoch}.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4507a4c4",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}