fashion-mnist.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import torch\n",
    "import torch.optim as optim\n",
    "import torch.nn.functional as F\n",
    "\n",
    "import torchvision\n",
    "import torchvision.transforms as transforms\n",
    "from torch.utils.data import DataLoader\n",
    "\n",
    "from model import Network\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\", category=UserWarning)\n",
    "\n",
    "torch.set_printoptions(linewidth=120)\n",
    "\n",
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def imshow(img):\n",
    "    npimg = img.numpy()\n",
    "    plt.subplots(figsize=(20, 20))\n",
    "    plt.imshow(np.transpose(npimg, (1, 2, 0)))\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Loading data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_set = torchvision.datasets.FashionMNIST(\n",
    "    root='./data'\n",
    "    ,train=True\n",
    "    ,download=True\n",
    "    ,transform=transforms.Compose([\n",
    "        transforms.ToTensor()\n",
    "    ])\n",
    ")\n",
    "test_set = torchvision.datasets.FashionMNIST(\n",
    "    root='./data'\n",
    "    ,train=False\n",
    "    ,download=True\n",
    "    ,transform=transforms.Compose([\n",
    "        transforms.ToTensor()\n",
    "    ])\n",
    ")\n",
    "\n",
    "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
    "\n",
    "classes = np.array(torchvision.datasets.FashionMNIST.classes)\n",
    "train_loader = DataLoader(train_set, shuffle=False, batch_size=500, num_workers=1)\n",
    "test_loader = DataLoader(test_set, shuffle=False, batch_size=8, num_workers=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###  Dataset visualization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABH4AAAC0CAYAAADiievYAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/d3fzzAAAACXBIWXMAAAsTAAALEwEAmpwYAAA6RUlEQVR4nO3debBdVZnH/Wc1UyBzyMANmUNmIAFCIIwtgyA2hklbqEJb6aJttNFqS6TpP1TUaqkS7LJsfAUFkWoFSqEFgQikgTQzJCQkkBlCBjLPmDCv949c3zfreX7J2bm5E+d+P1VUsp6su8++Z6+99j6bs3875ZwNAAAAAAAA9edv2noFAAAAAAAA0DK48AMAAAAAAFCnuPADAAAAAABQp7jwAwAAAAAAUKe48AMAAAAAAFCnuPADAAAAAABQp/bpwk9K6dyU0oKU0uKU0rXNtVIAAAAAAADYdynn3LQfTGk/M1toZmeb2Qoze9HMLs05v7aHn2naiwEAAAAAAGB31uec+6h/2Jdv/Ewys8U559dzzu+Z2V1mNmUflgcAAAAAAIC99+bu/mFfLvwcbmbLd2mvaKwBAAAAAACgHdi/pV8gpXSlmV3Z0q8DAAAAAACA0r5c+FlpZgN3aQ9orBVyzreY2S1mZPwAAAAAAAC0pn251etFMxuRUhqaUjrQzD5vZvc3z2oBAAAAAABgXzX5Gz855w9SSl8zsz+b2X5mdlvO+dVmWzMAAAAAAADskyY/zr1JL8atXgAAAAAAAM1tRs55ovqHfbnVCwAAAAAAAO1Yiz/VC0Cpa9euRXvSpEmhz7Rp05rltY499thQe/vtt0Nt4cKFzfJ6qCalVLTVNy/PPPPMULv66quL9qxZs0Kfww47rGgvXrw49OnSpUuo9ezZs2i///77oc+wYcNC7cILLww1tJw+ffoU7SuvjA/N3LJlS6jt2LGj5rL9z6lxud9++4XagQceWLTXrl0b+jzxxBNF+7333qu5PvXkb/4m/n+2jz76KNT83KA01ze1J0+eHGqHHHJIqPntq8aAd9BBB4XaunXrQm369Ok1l4XW5fdVM7MPPvigaL/77ruhT6dOnUJt6dKlNfv069evaKtzFDXm/D716U9/OvRB06h5qMq8488jzMw2bdpUtIcPHx769O7du2h/+OGHoY8ac3PmzKm5TmhZfj9UY0dtT+8LX/hCqPlj1P77x8sWfnzNnz8/9Lnttttqvn5Tx3yVZbXm3VW18I0fAAAAAACAOsWFHwAAAAAAgDrFhR8AAAAAAIA6xYUfAAAAAACAOsXj3IEm8iGF3/jGN0KfSy+9NNR8+J0PazUz2759e6j16tVrL9fQ7J133gk1FfLqg9eefPLJ0OeXv/xl0Z46deperw928mF4KuT1qaeeCrWTTz55r19r69atoaYCXH1onhqD6ufOP//8ov2nP/1pb1cRe+Gqq64q2jfddFPos3HjxlBbtWpV0VZB3StXrizaKvR9zJgxoebnmcceeyz0eeWVV4r2nXfeGfrUsyqhzUqVczT/wAAzszPOOCPUjjnmmKL9qU99KvRZsGBBzXVQ4fCHHnpo0V6/fn3oc/DBB4eaD+1V88f9999ftJctWxb6oGm6desWakuWLAk1FdjuqeODP9apcxJ//qGOPSos3K+TeiACqvH7oQrj9XOYCls+4IADQs1vTzUPbN68uWj7QHkz/cCJW2+9tWhfc801oQ/an6OPPjrUZs+eHWrPPPNM0VbHQz8uTjnllNBHhcpXCZxuzyHNNfA4dwAAAAAAgI6GCz8AAAAAAAB1igs/AAAAAAAAdWr/2l0A3HDDDaF25ZVXFm2Vs6DydHxN5XGoe6D/8pe/FG1/77yZ2XvvvVe01b3y6uf8/fN/93d/F/pMmTKlaD/77LOhz2mnnRZqiFSmjzd+/PhQ82NF5Wh07ty5aPt7983MNmzYEGoffPBB0VaZJEcccUSojR49umiT8dOyfCbY0qVLQ58q9677zB+zODf43BYznQnic6T69+8f+syfP7/mOtUzlQ2g9rEqGQL+2DNixIjQR+33Pr/nnnvuCX3UvOOzPHwemFq2yhZTxyM/ngcNGhT63HjjjTWX82//9m+h9tZbb4UaSir7Qo1Bv839ucbuaps2bSraalz6OUW9vprn1PkVmqbKMePv//7vi/b1118f+qjslksuuaRo//jHPw59fP7YWWedFfqo7Libb765aKu5yZ/bmH2ss1vaPX9OaGbWr1+/oq0yw0444YRQ+973vle01fmHPx784z/+Y+ijPpv4LCD1OU/NaR93fOMHAAAAAACgTnHhBwAAAAAAoE5x4QcAAAAAAKBO7VPGT0ppqZltM7MPzeyD3T0zHgAAAAAAAK2vOcKdP5FzjgmjwMeYD8+85pprQp/Vq1cXbR++bFYt0POAAw4Ifd55552aNbVsX1NBd4pf9ttvvx36+PC/k046KfR54IEHQu3888+vtA4odenSJdR8mLMKuvMBvT6Y1UwHbPqAb/VzysCBAyv1Q/Pwgcvr1q0LfYYNGxZqPmRVjZ1t27YV7R49eoQ+KpDYB9uruWnOnDmh1pE0Ncj5qquuCrVevXoV7TfffDP0ef/990PN7/cqYHP69OmhdsEFFxRtf+wziyGY6nebO3duqJ177rlFe/HixaGPD4pWAdDf//73Q+2KK64INZQuvvjiUPPjy8xs+fLlRVudW6gHR/jjiOrjH2ahlt29e/dQa2hoKNrHHXdc6DNjxoxQQ9P4kGQVnv6DH/wg1B566KGi7fd5M7OhQ4fWfH01F6rQ7yoIc24atY/544PfL83MnnnmmaKtzi3Ug278QwP69u0b+vhw59mzZ4c+Bx54YKj544r6nPfkk08W7Xnz5oU+6iEr7Rm3egEAAAAAANSpfb3wk83skZTSjJTSlTV7AwAAAAAAoNXs661ep+ScV6aU+prZoyml+Tnn4nvCjReEuCgEAAAAAADQyvbpGz8555WNf641s/vMbJLoc0vOeSLBzwAAAAAAAK2ryd/4SSl1NrO/yTlva/z7J83s+mZbM6AN+aBIHwJmZvbRRx8VbRWYe9hhh9V8LR+6qpZtFoP1OnfuHPp06tSpaKuwNBWu6IObfdCvWQwnXbNmTehz2mmnhVrv3r2L9sctCK019OvXr1I/H9iqAgr99lXj0o8lszjm1LLVfqDC9tByfJDv+PHjQx81f/h9XIXR+4BeNXZUsK8Pg1VBxvPnzw+1jqRquLMPSx8wYEDo88YbbxRtFQSv+G2u5p0lS5aE2uuvv160R4wYEfps2LChaKuQ5lNPPTXUfECsP4ap2o4dO0Ifday9/PLLi/add94Z+vjt0tFCX1UAttrHfYi8mvfVccWPX7Xt/NykHm6hlu3H76RJ4f89d7hw5yrjWQXdHnvssUVbhe/688Lhw4eHPuPGjQu18847r2hv3rw59Fm1alXRHjlyZOijjBo1ao/raKZDqP1DVdT5rDqOdiQ33HBDqE2bNi3U/Dm9Ota/+uqrRXvIkCGhj5+vzeL+u3DhwtDHb/PPfOYzoc+f//znUPNBzSeeeGLoc9ZZZ9Xs8z//8z+htmjRolBrL/blVq9+ZnZf4ySzv5n9Nuc8tVnWCgAAAAAAAPusyRd+cs6vm1n8X40AAAAAAABoF3icOwAAAAAAQJ3a16d6AXWpe/fuRfvdd98NfXyWisoYuPnmm0PtlltuKdrqHnR1j72/V37btm2hz7Jly4q2ug/f53iYmTU0NBTtlStXhj7+vvtu3bqFPgcffHCoDRs2rGiT8RMdeeSRlfr5jB/1fvu8BN820zlPPhtA5buo/cBnOKFl+cyGOXPmhD4qv8dvX5XP0LNnz6Ktxom6x97zmTBmOqOjI6maF3HEEUcUbbX/7r9/eer29ttvhz4q68Lv0+rnVLbH1KnlXfynnHJK6OOzW/w67q7mszVUdp0/1qiMEjU3TZgwoWirjJ+OlunjjR49OtReeumlUPPHGrUN1Hzh5yL1c35cbNmyJfRRNb9P9e/fP/TpaKqM57Fjx4ba8ccfX7QXLFgQ+vjcktmzZ4c+KpOsa9euRfuCCy4IfV5++eWirc4r1PmOH1+HHnpo6KMyyfx5sD+3Mut456pHHXVU0VZZOd/+9rdDbenSpUVbHev9OYH/GbN4/mFmdvvttxdt/3nCLI4LP++bmT3//POhdsghhxRtlQXlPwupZf/rv/5rqP3zP/9zqLUXfOMHAAAAAACgTnHhBwAAAAAAoE5x4QcAAAAAAKBOceEHAAAAAACgThHuDAg+GNMHG5vFsFTluuuuCzUfUqhCdFWI3RNPPFG0P/GJT9R8/ddeey3UxowZE2o+PPPqq68OfX7wgx8U7XXr1oU+KtzRB4G+8MILemU7sPHjx4eaCuH249CH05nFsatCuDdu3BhqPhRSjW8VGKuChNFyfKDp8uXLQx+13/vte8kll4Q+Phhz3Lhxoc+TTz4Zaj6gXoXD+1DX7du3hz6I77k69vj9UO2r6v3187MKjlbzxapVq4r2I488Evr4QE+17CVLloSaX/d+/fqFPj78t1OnTqGPMmnSpEr9OhL/IAd1/rF27dpQ8w+KUGHl6pg1cODAoq3Gsw8ZP+CAA0IfFQzul6UCvhGpEN3FixcX7S5duoQ+/pxPzRUbNmwINR+SPHHixNDH76tz584Nffr06RNqPjh606ZNldbJj191zt3RHHfccUX73HPPDX2+/OUvh9qUKVOKtgrFnj9/ftEeNWpU6KPCpP0YGzJkSOjj56aRI0eGPmpO8/3UAy/82Jk3b17o8+CDD4Zae8Y3fgAAAAAAAOoUF34AAAAAAADqFBd+AAAAAAAA6hQXfgAAAAAAAOoU4c4djA/yUwF9PgRUUSGvPlhvxIgRoc+iRYtqLru1+dBRRb1P6j3wfvOb34SaD0JTevXqFWo+zPn6668PfbZu3Vq0L7300krLHjRoUNG+5557Qh8f7qyCnFWg54QJE0INpeOPPz7U1JjzYc4+UNXMrHv37kV75syZoY/aJj4UUQVlqjBpFS6MluPDBc8888yafczi9lQB0D54/Re/+EXoo7b3ihUrirYK2NyxY0eoITr88MOL9rZt20KfKseeNWvWhJrff1Vg7vvvvx9qPnB6zpw5oY8/rrz11luhT//+/UOtR48eRVuFO/twaRU6/sYbb4SaD+ZUx3oVSFzPDjvssKJdNZzfh3Cr/dmHw5uZvfTSS0X7yCOPDH06d+5ctNWYV+cb/vingqM7OhXS7AORzeI+poJ2/X5fNWTdh3er/dCHK6t5SAWR+88rKtRe1fxcqM5tOpozzjijaKs5ddasWaHmH1jjt7dZDOsePHhw6LN69epQ+9///d+irQKYfRj8UUcdFfqowGl/rFHHTHWM9AYMGBBqvXv3rvn6bYVv/AAAAAAAANQpLvwAAAAAAADUqZoXflJKt6WU1qaU5u5S65VSejSltKjxz54tu5oAAAAAAADYW1Uyfn5tZj8zs13DSq41s2k55x+llK5tbH+7+Vev/vn7plVNZX34HAAzs8mTJxfthx9+OPSpej93LSr/w7voootC7YYbbmiW129OKnfAU9vA35OsqO1UxWc/+9mafe68885Q8/fdq3uiZ8+eHWoNDQ1FW92j21Qq6wmlMWPGhJq6x92PQ3X/vr9X/8QTTwx9VI6Xz1BQmQrqfueNGzeGGlqOzyJQc7rP8TDTuTue374qS0aNC5+tobKnfB5ElWNIvVN5Np7KnujZs/x/bSpzp2pGhqdy2vy28q9vFnM71LmNz2Iwi2NVjQs/Dn0u0O74sXr00UeHPj6Dpt6NHDmyaKt9tcp5ojqG+PMIM7MjjjiiaL/88ss112nZsmWhjxrPfqwyp0RqX1Hzus9XUXNTnz59irYaJ2o8+eODz6I0i9tXzR8+s8ss5suo45Oq+flKndv496nex1e3bt2K9sCBA0MfNV/6c06V/bR58+aircal2sd9LqxfR7P4ucfPJ2Y618qfE6mMsCeffLJoX3zxxaGPn+PMYt7ZxyrjJ+c83cz8mf0UM7uj8e93mNkFzbtaAAAAAAAA2FdNzfjpl3P+6yW+1WZW+39bAQAAAAAAoFXt8+Pcc845pbTb53+nlK40syv39XUAAAAAAACwd5r6jZ81KaUGM7PGP9furmPO+Zac88Sc88QmvhYAAAAAAACaoKnf+LnfzL5oZj9q/POPzbZGkEHC3qmnnhpqJ5xwQtFWocU//elPm75iu+jbt2+onXPOOUV727ZtzfJaLc0H1lXlgypVMJkKd1ZBc54PFFOmTp0aasOGDSvaKgzvvPPOC7XHH3+8aKsAaB/4rH4PFeyngmZR6t69e6ip97JKuPO9997bpHXwwa8q5FXxIYloWT5QU4X/qmOIPx6oMMtZs2YVbRXgqkLt/RhQIcJqfuzo/HxtFrevCmLt3Llz0VbbqVevXqHmj1kqhFPxc72aG/w6qHMEtZ7+91MB0H7MqXML9T75OXTo0KGhT0cLdx49enTRVgG9fnyZxfdShf9WCTB97rnnQm38+PFFW81favv68fTee+/VfP2ORp1bqPfJh++qhwH4baD6qFBmvz3VvOMDelXQrvo5Pzds37499FFj1R///O9vFoOE161bF/rUE/8eqBB/9fnBvy/qHMGHh6u5eMiQITVr6kEo/gEj6rj6q1/9KtT8OdGECRNCn9NPP71on3TSSaGPGnNqvmovqjzO/Xdm9qyZjUoprUgpXWE7L/icnVJaZGZnNbYBAAAAAADQjtT8xk/O+dLd/NOZzbwuAAAAAAAAaEZNzfgBAAAAAABAO7fPT/XCvlFZCP5e6uOPPz70Ufc5rlmzpmiPGDEi9LnvvvuKtr830kzfn/nmm28W7UMPPTT08ffDrlixIvRpj1QOj6fuW/bUfZ4q38bf76yWPWrUqFD70Y/KOyqHDx9ec53mzZsXav4efzOzwYMHF+2rrroq9Jk8eXLRVmNH3Tte5f3t6FQehhpPKiPD+93vflezz7vvvhtqPhNE5UMpKmMGLcdnIag8DJ/Hpag+L7/8cs2fU8cHv05qfJHxEw0cODDUfLZFlUw4tZxly5aFmp+f1fmHqvkcGLUtBw0aVLSrZsD5saKypxoaGoq2mhvVOvnayJEjQ5+O5ogjjijaW7ZsCX1UbpvfdipD8te//nXN17/ttttC7Stf+UrRVmNQ8etUNZeuI1HHZ3We5t87Nc/37t27aK9dG5/ro85Rqpy3+G2pxoCaU/z8oeYYlRek+lX5uXo2Y8aMon3HHXeEPirjxmfqqHw5P4er/CCVWdmjR4+i3bVr19DHjxU/Ts3MBgwYEGr+M7LaV/yyVCacyodSn4/aC77xAwAAAAAAUKe48AMAAAAAAFCnuPADAAAAAABQp7jwAwAAAAAAUKcId25lPpxMBYx17ty5aF9yySWhjwrP7NSpU9FWIVg+SFiFpamw4XHjxhXt5cuXhz6bNm0q2iqksT3q06dPzT4qQNUHiqnfVwWo/vCHPyzaBxxwQOjzyU9+MtTGjx9ftI888sjQx29zFeTsQ6LNzO6+++6iPWHChNDHU+F76n1Svx9KKlROjZ0q+9Tjjz9es8+zzz4baj68u2rAZtUQaDQPH8KpQm2rBGxWCYBW4ZYq+NWH/6rjGsGrkQrI9XPo1q1bQ5+DDjqoaPsHK6jlmMX5Q22TKvO6f331c9u2bQt9VKCnH2MqVNa/Byq8U4UU+3MZfwztiPxY8cHsZnr+8GNHHdf/8z//s+brv/jii6Hmx5c6L1Vj1YcUM8dE/nOBmX6f/H7Yr1+/0Mfvv2qfUw9+8ceMKscHFUCt5jQ/VtR8efrpp4eaf5CBGvNVHurycXXUUUeF2uc///mirR4UovZNPzeoceHPN6qGyvtalc8T6pxUvZ4fh2p8+XE4derU0Ec9xOcTn/hE0b7zzjv1yrYBvvEDAAAAAABQp7jwAwAAAAAAUKe48AMAAAAAAFCnuPADAAAAAABQpz4e6bstyId3qYAvFWbl+6mfUyGJVcLnvvKVrxTtNWvWhD4qdHPw4MFFW4W6+WVVDej14Z0qeM2HBqoASB9crZbd2hoaGmr2qRIqp4J3VaDYddddV/P11M/5bTd27Niay1m9enWoqTBrNZ48P8arjh2vqfsFYrCdCklUwe/e0qVLQ+2UU04p2lWDDdVYRctZv3590a56zPIhiVX2eRXQq8aFX9Zbb70V+lSZGzqaLl26hJo/tvqHJpiZDRo0qGjff//9lZbtx4oKBlfHbV9TAZt+LlLHQ7VsPy7UuFywYEHRPv/882suxyweV1RwdEfjt7k6/1Jzin8AgTq3eP3115u0Tj6MVc0xGzduDDUfJMz2jVRg7vbt20PNv+cqMN5v8x49eoQ+auz4fVPNO/6YpZajfq7KAy/UA3IWLlxYtNUxS32Gqhfqs5gPKf6Hf/iH0Oe8884Lte9973tF27+3ZvHzizqGHH744aHmH0SiPiusW7euaKu5YtGiRaG2du3aoq0ePnDfffcV7TFjxoQ+6qEBM2bMKNqEOwMAAAAAAKDFceEHAAAAAACgTtW88JNSui2ltDalNHeX2ndTSitTSrMa/4vf/QIAAAAAAECbqpLx82sz+5mZ/cbVf5Jz/nGzr1EzqpLfo2peS+aWXHrppaHm77N8+eWXQx91X6u/39bfN20W733s3bt36NO1a9dQU7+f5+/R9feEm5mNGDEi1GbNmlVz2S1JZd5U4bMYpk2bFvqcdtppobZixYqircaJui/bb3OVv+Gp+2hVZpTPXlDL9lkuEyZMCH3UmPOGDBkSakuWLKn5cx2Nmpv89mzq++bHoFm1e+zR9latWlW01VyhMjL8fKzmBk8dZ1QmyNatW4u2yhhCpDJvduzYUbRVjpffvq+99lroc+qpp4ba22+/XXOd1PmOP7dQuUP+59R6V/ldFJ8Zoc4t1HJ83ln37t1rvla98+eAVeYBs5gZNXXq1GZbJ58do8agz/EwM+vVq1fRZt6J1Byj9l+//4waNSr08flbKk9Q5SxV+fzg+1TJ1TSrNqddeOGFoXbjjTcWbZVbqnLS6oU6Zvj80UceeST0UfvhxRdfXLRV7qM/51Tb97LLLgs1nxs2fPjw0MfntKpjnxrzAwcOLNrqs68/D37ooYdCn8cffzzU1PvbXtScJXPO080sJiUBAAAAAACgXduXy+NfSym90ngrWIzCBgAAAAAAQJtq6oWfn5vZcDObYGarzOzG3XVMKV2ZUnoppfRSE18LAAAAAAAATdCkCz855zU55w9zzh+Z2a1mNmkPfW/JOU/MOU9s6koCAAAAAABg71UJdw5SSg05578mTF5oZnP31L+tVAkn9WFhKjxMhe/6ZVcJcjYz+/KXv1y0R44cGfosX768aB966KGhjwoy9KFqK1euDH18eJUK2Nq+fXuoderUqebrV3m/zznnnFBr63BnH1ypqJA3H1Z2xx13hD7nnRcfeKfeX0+NQ/+eq+BVT20T9XM+AFCFcN5+++1FW4U7V6ECxQl3jtQ26Ny5c9GeO7dpU68KqLvmmmuKNkGZ7ZOfP9R8ogIv/fb0waiKCnJWYaE+5LNKyHtHo+ZdFazrQ07Vsfb9998v2m+99VboUyU0WYUkq3nH91Pb1x9r1LHHr7daTxUEu3jx4j2uj5mer/x77udPs3hsrxIW+3HmH9zQs2dMalBj1YeqfvOb36z0en67qHPON954o2gffvjhoc/69etrrueAAQMqrVNHovZDH8ZvFuf1oUOHhj5+7Khjgf+sYBb3e7VOflxU/Uzlj39q3vPh4WZxjL3yyiuhTz2fA6kH7fjPo2ob9O3bN9T8nK3mcP/5VC3bhy2bmY0bN65ojx49OvTx41CNAXWsHTRoUNFW50Q+pFk9HEe9l0cffXTRVuOrrdT85JhS+p2Z/a2Z9U4prTCz75jZ36aUJphZNrOlZvZPLbeKAAAAAAAAaIqaF35yzvF542a/aoF1AQAAAAAAQDOq3++xAQAAAAAAdHBc+AEAAAAAAKhTTQp3bmtVA7d8gJgKfPKBYip4ror+/fuH2kUXXRRqPuBq0aJFoY8PG1QBairw+b333ivaKkBNhSJ6KnTLh3eqPj4IVL2XJ598cs3Xb20q0Mu/d+p9W7duXdHetGlTpdfzQXcqSLFKUHYVajkqeM33O/DAA0Of559/vkmvt2PHjqJdz4F5zalKwLcPxaxq9uzZoea3uQrDU1QAMFqOn3vV+18l6NbPX4o6PvljmFkcKyrgs6NTofZVHpKg5mt/rFfHEFXzwc1qnlfhxn4OV3OD3+YqyFmdN/j3QPXx4dVVg1/feeedoq2OT4cddljR9kHS9caPHbWvqodZ+DnFh57ujh+/6rzw1VdfLdoqWFgFEvfp06doVz0Hq2d+e6pjgT+fNzPr1q1bzWX782B17FHh8H4uqHJuUfW82I9nFQyuPp9VCQKv53NVFUjs50u1nT73uc+F2rXXXlu0/f5sZrZ58+aird5b9aCK3/72t0X7mGOOCX38eqv54+GHHw61Z599tmirz4I/+clPar6++nzox7x6iJB/T1pL/Y5qAAAAAACADo4LPwAAAAAAAHWKCz8AAAAAAAB1qt1l/Kj72f393E3N4amSm+LvGTYzGzJkSKiNGjWqaDc0NIQ+/t5Ts3ifsrrvz99rq+6zVLk//n0ZPHhw6OOXpe4xVPfm+2Wr+zN9DoDaltu2bQu1cePGFW11f2hLUtvA3wOt7oP3WQhjxoyp9Hr+HuiqWSpNyf2pkiGhauo9qfL66vX8WFFZFx3dihUrQk3dN+z3Q599UZW6D99T+69Cxk/bUlkI6l51369KHobK8VDZCP6Y5Y8F0HOq2nY+r0D93PLly4u2Oq527tw51NasWVPz9dWx3Z+DqeOhz35SOTxq3vHHv65du4Y+PnNm7dq1oY86L/RzmPrd+vbtW7TrPeNnzpw5RXvSpEmhjzq/9Hlfq1evrvR6Vc7XH3zwwaJ99dVXhz5qPPfr169ob9iwodI61bMqx22VlzRy5MiaP+fndZUVpF7f779qbvD7pho3VbLrVq5cGfqosaoybrwqeWtV88bam+OOOy7UNm7cWLRVlqz/7GsW5/Uzzjgj9FmwYEHRVjlip59+eqi9/PLLRVuNU3+MVOs9ffr0UJs8eXLRVp/Zly1bVrRVxo8ac/5zjrq2QMYPAAAAAAAAmhUXfgAAAAAAAOoUF34AAAAAAADqFBd+AAAAAAAA6lS7C3euEpTlA93MdJCxD4NT4XA+kHDo0KGhjwpZ9QHIPujXTAeRde/efY+vbxaDstTrb9++PdR80NqBBx4Y+qxatWqP67O71/NBoCqYq2fPnkVbhb4edthhoaaCuFqTCmyrEmTsw8qGDx9e6fX8stU4Ua+vgpP39rXM9O/rx44aFypQs8qy/XqrkLOOzoeumunx5N/fKoGMigqx86qGFqr5Aq1HzZ8+iNXM7FOf+lTR/sUvflFz2TNnzgw1FQbrww3VnNbRqblYnTf4uVjt4/7Yo5bT1AB39bABv+4q1NUHv6oA6CrHWn8eYRbPJXxAsZkOhfbnLSowVp3L1LO77767aH/pS18KfdTc7wPcVYDrI488EmpVzlv8ePbh5WbVwn79OkLPO+rzw7HHHlvz5/w+rj6/qLHjPy9VObdQ85caA1XGl/osokKKPT++VBj+xzXc+dlnnw215557rmgfeeSRoc9TTz0Van6eVT/nP4+qcwS1Lf0xQ41d/5miSgi4WTzWqfNif2xVc0yV41GVz0+thbMzAAAAAACAOsWFHwAAAAAAgDpV88JPSmlgSunxlNJrKaVXU0pfb6z3Sik9mlJa1Phn/H4uAAAAAAAA2kyVjJ8PzOybOeeZKaWuZjYjpfSomf2DmU3LOf8opXStmV1rZt9uiZU866yzinb//v3jSor7Qf19f+r+cn9/plrOtm3bQs3fF66ya9T9igcddFDR9vdGmsX7E9U96Op38fexqvv+t2zZUrT79u0b+lSh1tvff6vu//W/v1m1LIKW1NT7dhcuXFi0TzvttCa/nqfGjq9VySFSy1H3v1bZBitWrNhj26xaXlNHy1So4sUXXwy1MWPGhJrP1hg/fnyLrZPaVxWV94HWc/rpp4eayofyGT+XX355zWXPnTs31Hr16hVqX/3qV4v2K6+8EvrMmDGj5uvVMzU3quOMP26qvDX//qrcNDXP+mOGOhap/d7nKqjjhV+2Ov6rY4/P/1A5HoMGDSrar7/+eugzefLkUPPrMH/+/NCno+XC+DHn338zPXb8Nlfzh8r4qXJusX79+qJdNcfTr+c777xT87Xqnc8tUeeA6tzRz+vq5/xnIZWbqrJF/ecVlaVS5by4ynypxvOGDRua9Hp+LmpKzmZ7NWHChFBbvHhxzT4+z8/MrKGhoWgPGDAg9Fm9enXRVplsfp5Xy1I5vL6POvaoOcW/nhon/nOeysBTn4X8fKWy6/zn8dZS8xs/OedVOeeZjX/fZmbzzOxwM5tiZnc0drvDzC5ooXUEAAAAAABAE+xVxk9KaYiZHWNmz5tZv5zzXx8RtdrM4uU0AAAAAAAAtJnKj3NPKXUxsz+Y2Tdyzlt3/cpbzjmnlOR9JymlK83syn1dUQAAAAAAAOydSt/4SSkdYDsv+vx3zvnexvKalFJD4783mJl8SH3O+Zac88Sc88TmWGEAAAAAAABUU/MbP2nnV3t+ZWbzcs437fJP95vZF83sR41//rE5VuiTn/xkqF1xxRVFWwX0rVq1KtR8EJkKFvQhYyo0uUrImQo0UyGFPtBKLdsHU6nlqIApHzCtwqzGjh1btNV6q/fJ80HSZmaHHHJI0VZBeypweu1aec2w1aj1rBLu7LfL6NGjQx8VNFfl/W0qv2wV4qfGU5Xf94gjjijaPqzNTIec+33MjxOYTZ8+PdS+9KUvhZofT8cee2yzrYMfA2ouVNR4Qsvxxwy1nUaMGBFqPrixShCqCmZVYcMnnHBC0VbHp45O7audOnWqWVPHcf9whYkT4/9X27FjR6j5fVUdH9R48nO4Ckb1xx61HBUE72tqzPkQ+61bt4Y+6vf151IqjPa4444r2r///e9Dn3qmgpxVOKqfLyZNmtRi66T2C7X/+PNXtd4djX9P1H6o9jH/3qm5wYfRqrlJ/ZwfYypAvsrcpI4rft5Rc4MK0R04cGCoef59aslz99b26U9/OtT8ucXXv/710OfPf/5zqPkHN6hzwpkzZxZt9f6/8MILofbqq68WbbUN/PZVx6fZs2eHmg80Vw8t8g9OuOmmm0If9dnPP4TqP/7jP0KfpUuXhlprqHKr18lmdrmZzUkpzWqsXWc7L/jck1K6wszeNLPPtcgaAgAAAAAAoElqXvjJOT9lZrt7ht2Zzbs6AAAAAAAAaC718701AAAAAAAAFLjwAwAAAAAAUKcqP869tahwpxNPPLFoH3XUUaHPySefXHPZKtDMhzRv3Lgx9FE1HyalQpJVcPOhhx5atEeNGhX6+PDbbt26hT4q+MwHIL7yyiuhjw+TOuuss0IfFbymXs/z7+/KlStDHxW85gOvW5saF1WCbX3QnN+2Zmbbt29v0rKrqLJNFBW8VmWdpkyZUrRVMNkxxxwTan49e/bsWfO1Oppnnnkm1FT4rh+rzRmM7udCNX8p9RR4+HHg9yd17FEhpypYtxa1bBWc6AOfVZ+OTj0QQYXYHn744UVbHR99UKU/9puZbd68OdSqBOur/d6fE6jjhT+uqAc5+JBoM7O+ffvucTlmZoMHDy7aDzzwQOjzy1/+MtTuueeeoq22gXpIQUfy9NNPh9pll10Wahs2bCjaavs2lzfffDPUfBCrWZyfOBbF/bfqeWJDQ0PR9g8DUMtS+6o6n/a1KufcVc8/qjxcYt68eaGmPnt59Rzu/K1vfSvUnn322aKtgt+XLFkSaj169CjaKoTbn8+q45Oai/3nSLUNfJBy1c/My5cvL9rqeOznGHWceeqpp0LNr6fq01bqZxQDAAAAAACgwIUfAAAAAACAOsWFHwAAAAAAgDrV7m7EV/f9XX/99TV/Tt2LeMIJJxRtdU/nSSedVLSHDBkS+hx99NGh1rlz56Kt7kdV9xT6+1FVftCcOXOK9mOPPRb6PPTQQ6GmMkFquf/++0Nt0KBBobZ+/fqi7fNAVE3dx6tyJhYuXFhzPVvShx9+GGrqXk9v9OjRRVvlYajf1+dfqHuUq9zfrPo09f7uKhk/ft9QGVKXXHJJqPl1UPf/dnQq00DlYfmsDTVOhw0bVrRff/31Suvw/vvvF+2qOS3NlVmFplG5Keoed5VvUosfE2Z6vvT7dEfPTVFuv/32Sv38uYzfn81izsJFF10U+mzatKnmstUxRJ2D9e7du2irY52vqflD1fzxb926daGPz3m85ZZbQh+fFWQWc2iaco5U7372s5+FmjqO++3kcz3M9FitevzZlTq/VFlX/tijxnxH4/fpKhk4ZvG8f8WKFTWXrfJAVQaLz5xTffx5YpU+atmKyqPyc5E6j/HHunrKrlP7qp8f1XuyYMGCUDvzzDOLtjoeHXvssUXb5/KYmX3xi18MNZ8JOnDgwNBnzJgxRVudF6vXmzBhQtFWOWKPPvpo0e7Tp0/o069fv1Dzx0M1X6pjXWvgGz8AAAAAAAB1igs/AAAAAAAAdYoLPwAAAAAAAHWKCz8AAAAAAAB1qm6SqlR417Rp0/bYNjO7+eabW2ydPg4+85nPtPUqtDkVjlolXNmHjqmQObXsKmF7VfqooDtfU32qhEJv2bIl9Jk8eXLRrhrK7dehShgfdHCiD9tTfZoa7rxq1aqirYLuVRg94c5ta8eOHaGmwg2bEmxbdW70QZwqFBrV+HMZFaLvg25VKKXaV30I95o1a0KfQw45JNT88tUY8GNFHcPUfKXGWK118qGcZvqBF6ht5cqVoaYCvv0DTVTA96RJk0KtKeHOakyoYF2/Dmp8dXRqOyn+vGzRokWhjw87Vg8vUeec/udUnyrrqR4sUIV6sIH/fdW85x9QU/W9/Djw+7NZDC5WQcYvvfRSqM2cObNoq7Hz1FNPFe3x48eHPupc5q677ira48aNq/n66pz0t7/9baj530UdR6dOnVrz9dXDpfz7254+9/CNHwAAAAAAgDrFhR8AAAAAAIA6VfPCT0ppYErp8ZTSaymlV1NKX2+sfzeltDKlNKvxv/NafnUBAAAAAABQVZWMnw/M7Js555kppa5mNiOl9Gjjv/0k5/zjlls9AAAAAAAANFXNCz8551Vmtqrx79tSSvPM7PCWXjGgtaggUh8ypsK7brrppqJ95plnhj4q0KupAXVNDW72VPCZX6du3bqFPk888UTR/tOf/hT6fOc736m57HoKyGsqv53UtrzvvvtC7bLLLtvjcszMTjnllKL92GOPVVonFYDoqdfbtGlTpeWjZRx22GGhpvZxH8BchXpoggrt9a+nQhoRVQnKVscLv49XDdPevn37Hl/LLIbDm5ktXbq05rL79etXc9kqdNyvkxo7PoD41FNPDX1UuHOVebajqfKePPLII6F2ySWXFG0VwDxlypRQ8+GsVah5p6GhIdT8GGvKHFdv/D5W9XzTP8zhmWeeCX2GDh1atNU2UfuvP0dQQd3+GKL6+HD63dWqrFP37t33+PpmMdy5nvgHBJiZDRgwoGgfccQRoY+fr83MzjnnnKKt3ku/PdV5y7x580LNz0/q9efMmVO01TFMBdavXbu2aPtjmFkc49u2bQt9Bg0aFGr+M6M69rWVvZolU0pDzOwYM3u+sfS1lNIrKaXbUko9d/+TAAAAAAAAaG2VL/yklLqY2R/M7Bs5561m9nMzG25mE2znN4Ju3M3PXZlSeimlFJ8BBwAAAAAAgBZT6cJPSukA23nR579zzveameWc1+ScP8w5f2Rmt5rZJPWzOedbcs4Tc84Tm2ulAQAAAAAAUFvNjJ+084bgX5nZvJzzTbvUGxrzf8zMLjSzuS2zikDLOuSQQ0LN3xetMhT8vcXr168PfUaMGBFqS5YsKdpNvS+9Sp6P6qMyOvy9zL169Qp9/P2w6vdV/Hs5ePDgSj9Xz6rkLPzxj38MtS984QtFW43Liy++uGh/97vfrbRO/h5stU6q9u6771ZaPlrGmjVrQq1v376h1pS8ApXfpDIjDjrooKLt5wpoan+qkskxatSoor1169bQR2Wp+WWr49Obb74Zaj7/q3///qGPHwPq2KNyDnw/lR3jayofQvHvr1qnjpb7UyVDSuUlffazny3aKjfFZ4Q01ZYtW0LNjy8zs40bNxZtdd7S0fjj+DvvvBP6qAwWv2++9FK8ScPvP+rYr85n/XZReYL+5zp37hz6qFwafz6r5piZM2eG2urVq4v2wIEDQ58FCxYU7Sp5Qh8XPhfHzOy5554r2v44Y6bPOf12UX18ptKJJ54Y+mzYsCHUzj777KKtxsUbb7xRtE844YTQ59FHHw01v819zpWZ2cKFC4v29OnTQ5+xY8eGmj8m+899banKU71ONrPLzWxOSmlWY+06M7s0pTTBzLKZLTWzf2qB9QMAAAAAAEATVXmq11Nmpr5aEP+XAAAAAAAAANoNnn0IAAAAAABQp7jwAwAAAAAAUKeqZPwAde2ZZ54JtcmTJxdtFZDnQ79GjhzZvCvWzg0bNizUtm3bFmo+bO/FF19ssXX6uPBBhipw++GHHw41H7arAi/VsqqYO7fM5z/qqKNCHxXo2dDQ0KTXQ/NQ42TixPgQzaaMC7U/qyBhv4+rgGBU44NXVfjuoEGDirYKcl68eHGo+THgj2FmMTDXLIZXqrHkg0/Ver/99tuh5oN81e/i5zn1QAY1F/rwWcKdq80DTz/9dKitXLmyaPuwVjMduj1+/PiiPXv27Jqvr+aYgw8+ONR8iKwKo+9o/HhW41uFs/v97ve//33zrlgzUOG/Vaigah8SfMYZZ4Q+PgBZBQt/XC1btizUzjzzzKLtjzNmev7w+/hbb70V+vg5e+jQoaFPlf1XBWz7uUHNFSoY3K+TCviuEmjer1+/UPPzZXuam/jGDwAAAAAAQJ3iwg8AAAAAAECd4sIPAAAAAABAnSLjBx3eCy+8EGr+3s/33nsv9Glqlkq9UPfaqpwFf++4ynnoaFT+RRX+vuwTTzwx9PH3oZ900kmhj8q18tkiPrfFTG/z3r1765VFq1D5Y2rbNXXMeer+eT/mVqxY0Syv1RFVyZz593//96L9rW99K/Q599xzQ61Hjx5F+4033gh9Pvjgg1Dz42ndunWhT8+ePYu2ylTo1atXqPl8BJ/5Y2a2fv36ov1f//VfoY/KXvA6+jHbrOmZRv7Yc/7554c+auycffbZRbtKxo8aO2re8VTWRkczePDgoq2ymPw8YGb2/e9/v6VWqV366U9/WrTVXOgzq3w2o1n7ym7ZGz6/yMzsX/7lX4r28ccfX2lZv/nNb4q2Oi/15x9dunQJfVSG0/Dhw4u2z/Uyi5/XVAacmvv9ZxO1LefPn1+0jz766NBH5WEuXbq0aLenLDm+8QMAAAAAAFCnuPADAAAAAABQp7jwAwAAAAAAUKe48AMAAAAAAFCnCHdGh7dy5cpQmzlzZtFWAap/+ctfai57//3jLuZDzlJKNZfT2tQ6+fVevHhx6PPggw+Gmg8XfO655/Zx7T7+mhr0duuttxZtHzxnZnbXXXcVbRXkrNx5551FW4VCqmDu//u//6u0fLQMH6xoZnbKKaeE2sMPP9wsr3f//ffX7KOCI1FNlQDiHTt2FO3rr7++0rIHDRpUtMeOHRv6qIDcbt26FW0VcuqpEE5V86HBTz/9dOjDAwHa3g9/+MOivXr16tBHbd8nn3xyr1/r7rvvDrU1a9aE2ubNm4v2tGnT9vq16o3fV9QDGbZu3RpqTzzxxF6/ljpPbE8htnvyhz/8oWirB7j4B17UExXEfu+99xbtVatWVVrW3Llz99hWbrvttlCbMWNGqPnxqz6v+SBltd6vvfZazZ974IEH1KrWXEd1zF6+fHnRbk/7Bd/4AQAAAAAAqFNc+AEAAAAAAKhTXPgBAAAAAACoUzUv/KSUOqWUXkgpzU4pvZpS+l5jfWhK6fmU0uKU0t0ppQNbfnUBAAAAAABQVaoVOJR2pnd1zjm/nVI6wMyeMrOvm9m/mtm9Oee7Ukr/j5nNzjn/vMay2k+6EQAAAAAAQH2YkXOeqP6h5jd+8k5/jYk/oPG/bGZnmNnvG+t3mNkF+76eAAAAAAAAaC6VMn5SSvullGaZ2Voze9TMlpjZ5pzzX58Ht8LMDt/Nz16ZUnoppfRSM6wvAAAAAAAAKqp04Sfn/GHOeYKZDTCzSWY2uuoL5JxvyTlP3N1XjgAAAAAAANAy9uqpXjnnzWb2uJlNNrMeKaX9G/9pgJmtbN5VAwAAAAAAwL6o8lSvPimlHo1/P9jMzjazebbzAtAljd2+aGZ/bKF1BAAAAAAAQBPsX7uLNZjZHSml/WznhaJ7cs5/Sim9ZmZ3pZR+YGYvm9mvWnA9AQAAAAAAsJdqPs69WV+Mx7kDAAAAAAA0t6Y/zh0AAAAAAAAfT1Vu9WpO683sTTPr3fh3AO0L+ybQ/rBfAu0T+ybQPrFvoqMavLt/aNVbvf6/F03pJR7vDrQ/7JtA+8N+CbRP7JtA+8S+CUTc6gUAAAAAAFCnuPADAAAAAABQp9rqws8tbfS6APaMfRNof9gvgfaJfRNon9g3AadNMn4AAAAAAADQ8rjVCwAAAAAAoE61+oWflNK5KaUFKaXFKaVrW/v1AeyUUlqaUpqTUpqVUnqpsdYrpfRoSmlR458923o9gXqXUrotpbQ2pTR3l5rcF9NOP208hr6SUjq27dYcqG+72Te/m1Ja2XjsnJVSOm+Xf/u3xn1zQUrpnLZZa6C+pZQGppQeTym9llJ6NaX09cY6x01gD1r1wk9KaT8z+y8z+5SZjTWzS1NKY1tzHQAUPpFznrDLIy+vNbNpOecRZjatsQ2gZf3azM51td3ti58ysxGN/11pZj9vpXUEOqJfW9w3zcx+0njsnJBzfsjMrPF89vNmNq7xZ25uPO8F0Lw+MLNv5pzHmtmJZvbVxv2P4yawB639jZ9JZrY45/x6zvk9M7vLzKa08joA2L0pZnZH49/vMLML2m5VgI4h5zzdzDa68u72xSlm9pu803Nm1iOl1NAqKwp0MLvZN3dnipndlXN+N+f8hpkttp3nvQCaUc55Vc55ZuPft5nZPDM73DhuAnvU2hd+Djez5bu0VzTWALS+bGaPpJRmpJSubKz1yzmvavz7ajPr1zarBnR4u9sXOY4Cbe9rjbeM3LbLLdHsm0ArSykNMbNjzOx547gJ7BHhzkDHdUrO+Vjb+RXYr6aUTtv1H/POR/7x2D+gjbEvAu3Kz81suJlNMLNVZnZjm64N0EGllLqY2R/M7Bs55627/hvHTSBq7Qs/K81s4C7tAY01AK0s57yy8c+1Znaf7fxK+pq/fv218c+1bbeGQIe2u32R4yjQhnLOa3LOH+acPzKzW+3/v52LfRNoJSmlA2znRZ//zjnf21jmuAnsQWtf+HnRzEaklIamlA60nSF497fyOgAdXkqpc0qp61//bmafNLO5tnN//GJjty+a2R/bZg2BDm93++L9ZvaFxqeUnGhmW3b5ajuAFuayQS60ncdOs5375udTSgellIbaziDZF1p7/YB6l1JKZvYrM5uXc75pl3/iuAnswf6t+WI55w9SSl8zsz+b2X5mdlvO+dXWXAcAZrbzvuf7dh47bX8z+23OeWpK6UUzuyeldIWZvWlmn2vDdQQ6hJTS78zsb82sd0pphZl9x8x+ZHpffMjMzrOdwbHbzexLrb7CQAexm33zb1NKE2znbSRLzeyfzMxyzq+mlO4xs9ds51OHvppz/rANVhuodyeb2eVmNielNKuxdp1x3AT2KO28BRIAAAAAAAD1hnBnAAAAAACAOsWFHwAAAAAAgDrFhR8AAAAAAIA6xYUfAAAAAACAOsWFHwAAAAAAgDrFhR8AAAAAAIA6xYUfAAAAAACAOsWFHwAAAAAAgDr1/wIRAzcQtpp6GQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 1440x1440 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ground truth vector ['Ankle boot' 'Pullover' 'Trouser' 'Trouser' 'Shirt' 'Trouser' 'Coat'\n",
      " 'Shirt']\n"
     ]
    }
   ],
   "source": [
    "images, targets = next(iter(test_loader))\n",
    "imshow(torchvision.utils.make_grid(images, nrow=10))\n",
    "print(f'Ground truth vector {classes[targets]}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train(model, loader, optimizer, device):\n",
    "    for batch, labels in loader:\n",
    "        batch, labels = batch.to(device), labels.to(device)\n",
    "        output = model(batch) #pass batch\n",
    "        loss = F.cross_entropy(output, labels)\n",
    "        \n",
    "        optimizer.zero_grad() #set the gradients to zero\n",
    "        loss.backward() #calculate gradients\n",
    "        optimizer.step() #update weights\n",
    "        \n",
    "    return loss.item()\n",
    "\n",
    "@torch.no_grad()\n",
    "def evaluate(model, loader, device):\n",
    "    correct, total = 0, 0\n",
    "    model.eval()\n",
    "    \n",
    "    for batch, labels in loader:\n",
    "        batch, labels = batch.to(device), labels.to(device)\n",
    "        output = model(batch)\n",
    "        correct += output.argmax(dim=1).eq(labels).sum().item() #sum up the the correct predictions\n",
    "        total += labels.size(0)\n",
    "            \n",
    "    return 100 * correct / total \n",
    "\n",
    "def train_stats(epoch, train_loss, train_accuracy, test_accuracy):\n",
    "    print(f\"\"\"\n",
    "            Epoch {epoch}:\n",
    "            Train loss = {train_loss}, Train accuracy = {train_accuracy} || Test accuracy = {test_accuracy}\n",
    "            \"\"\"\n",
    "         )\n",
    "    \n",
    "@torch.no_grad()\n",
    "def predict(model, loader, device):\n",
    "    for batch, labels in loader:\n",
    "        batch, labels = batch.to(device), labels.to(device)\n",
    "        output = model(batch)\n",
    "        yield output.argmax(dim=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Train/Test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "            Epoch 0:\n",
      "            Train loss = 0.3415835201740265, Train accuracy = 85.52166666666666 || Test accuracy = 84.77\n",
      "            \n",
      "\n",
      "            Epoch 1:\n",
      "            Train loss = 0.3294743597507477, Train accuracy = 88.54166666666667 || Test accuracy = 87.33\n",
      "            \n",
      "\n",
      "            Epoch 2:\n",
      "            Train loss = 0.31111836433410645, Train accuracy = 89.575 || Test accuracy = 87.96\n",
      "            \n",
      "\n",
      "            Epoch 3:\n",
      "            Train loss = 0.2877389192581177, Train accuracy = 89.92666666666666 || Test accuracy = 88.28\n",
      "            \n",
      "\n",
      "            Epoch 4:\n",
      "            Train loss = 0.2920975685119629, Train accuracy = 90.38666666666667 || Test accuracy = 88.66\n",
      "            \n",
      "\n",
      "            Epoch 5:\n",
      "            Train loss = 0.2861175239086151, Train accuracy = 90.78333333333333 || Test accuracy = 88.75\n",
      "            \n",
      "\n",
      "            Epoch 6:\n",
      "            Train loss = 0.27454856038093567, Train accuracy = 90.96166666666667 || Test accuracy = 88.61\n",
      "            \n",
      "\n",
      "            Epoch 7:\n",
      "            Train loss = 0.24221056699752808, Train accuracy = 91.50333333333333 || Test accuracy = 88.82\n",
      "            \n",
      "\n",
      "            Epoch 8:\n",
      "            Train loss = 0.21225318312644958, Train accuracy = 91.855 || Test accuracy = 88.89\n",
      "            \n",
      "\n",
      "            Epoch 9:\n",
      "            Train loss = 0.2345620095729828, Train accuracy = 91.99166666666666 || Test accuracy = 88.85\n",
      "            \n"
     ]
    }
   ],
   "source": [
    "network = Network()\n",
    "network.to(device)\n",
    "optimizer = optim.Adam(network.parameters(), lr=0.01)\n",
    "\n",
    "for epoch in range(10):\n",
    "    train_loss = train(network, train_loader, optimizer, device)\n",
    "    train_accuracy = evaluate(network, train_loader, device)\n",
    "    test_accuracy = evaluate(network, test_loader, device)\n",
    "    train_stats(epoch, train_loss, train_accuracy, test_accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Inference"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Predictions ['Ankle boot' 'Pullover' 'Trouser' 'Trouser' 'Shirt' 'Trouser' 'Coat'\n",
      " 'Shirt']\n"
     ]
    }
   ],
   "source": [
    "preds = next(predict(network, test_loader, device))\n",
    "imshow(torchvision.utils.make_grid(images, nrow=10))\n",
    "print(f'Predictions {classes[preds.cpu()]}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Confusion matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix, without normalization\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAArkAAALICAYAAAB7IOjQAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/d3fzzAAAACXBIWXMAAAsTAAALEwEAmpwYAACsrklEQVR4nOzdd3gUVRfH8e9JQkea9ISOdCEk9CYKgtJBpSogIGDvvWHvDbH7WlHBThMEaQKCdGxIUVAIIAQJUk3hvn/sEgNCGklmd/19nmef7N6dnTmzk3Jy9swdc84hIiIiIhJKwrwOQEREREQkpynJFREREZGQoyRXREREREKOklwRERERCTlKckVEREQk5ER4HYCIiIiIZE14sSrOJR/yOgzcoV1fOufO8zqOE1GSKyIiIhJkXPIhCtTu63UYHF79QmmvYzgZtSuIiIiISMhRkisiIiIiIUftCiIiIiJBx8BUq0yP3h0RERERCTmq5IqIiIgEGwPMvI4ioKmSKyIiIiIhR0muiIiIiIQctSuIiIiIBCOdeJYuvTsiIiIiEnKU5IqIiIhIyFG7goiIiEgw0uwK6VIlV0RERERCjiq5IiIiIkFHVzzLiN4dEREREQk5SnJFREREJOSoXUFEREQkGOnEs3SpkisiIiIiIUdJroiIiIiEHLUriIiIiAQbQ7MrZEDvjoiIiIiEHCW5IiIiIhJy1K4gIiIiEnRMsytkQJVcEREREQk5quSKiIiIBCOdeJYuvTsiIiIiEnKU5IqIiIhIyFG7goiIiEgw0oln6VIlV0RERERCjpJcEREREQk5alcQERERCTqm2RUyoHdHREREREKOKrkiIiIiwcbQiWcZUCVXREREREKOklwRERERCTlqVxAREREJRjrxLF16d0REREQk5CjJFREREZGQo3YFERERkaCjeXIzondHREREREKOKrkiIiIiwShM8+SmR5VcEREREQk5SnJFREREJOSoXUFEREQk2Bg68SwDendEREREJOQoyRURERGRkKN2BREREZFgZJpdIT2q5IqIiIhIyFGSKyIiIiIhR+0KIiIiIkFHl/XNiN4dEREREQk5quSKiIiIBCOdeJYuVXJFREREJOQoyRURERGRkKN2BREREZFgpBPP0qV3R0RERERCjpJcEREREQk5alcQERERCTZmml0hA6rkioiIiEjIUSVXREREJBjpxLN06d0RERERkZCjJFdEREREQo7aFURERESCkU48S5cquSIiIiIScpTkioiIiEjIUbuCiIiISNAxza6QAb07IiIiIhJylOSKiIiISMhRu4KIiIhIMNLsCulSJVdEREREQo4quSIiIiLBxtCJZxnQuyMiIiIiIUdJroiIiIiEHLUriIiIiAQdzZObEb07IiIiIhJylOSKSEgxs0JmNsXM9prZR6ewnkFmNjMnY/OKmbU1s3VexyEikpeU5IqIJ8xsoJktN7P9ZrbdzKabWZscWPWFQDngdOfcRdldiXPuPedcpxyIJ1eZmTOzmukt45xb4JyrnVcxiUgeMfP+FsCU5IpInjOzG4BngYfxJaSVgReBnjmw+irAeudccg6sK+iZmc69EJH/JCW5IpKnzKw4cD9wpXPuU+fcAedcknNuinPuZv8yBczsWTPb5r89a2YF/M+1N7OtZnajme30V4Ev9T93H3AP0M9fIR5uZmPMbHya7Vf1Vz8j/I+HmtmvZrbPzDaZ2aA04wvTvK6VmS3zt0EsM7NWaZ6bZ2YPmNki/3pmmlnpk+z/0fhvSRN/LzPrYmbrzexPM7sjzfLNzGyxmSX4lx1nZvn9z33tX2yNf3/7pVn/rWa2A3jz6Jj/NTX824jxP65oZrvMrP2pHFcR8YCFeX8LYIEdnYiEopZAQeCzdJa5E2gBRAONgGbAXWmeLw8UByKB4cALZlbSOXcvvurwROdcUefc/9ILxMyKAGOB851zpwGtgNUnWK4UMM2/7OnA08A0Mzs9zWIDgUuBskB+4KZ0Nl0e33sQiS8pfw24GIgF2gJ3m1k1/7IpwPVAaXzvXQfgCgDnXDv/Mo38+zsxzfpL4atqj0y7YefcL8CtwHgzKwy8CbztnJuXTrwiIkFHSa6I5LXTgfgM2gkGAfc753Y653YB9wGXpHk+yf98knPuC2A/kN2e0yNAAzMr5Jzb7pz78QTLdAU2OOfedc4lO+c+AH4GuqdZ5k3n3Hrn3CHgQ3wJ+skkAQ8555KACfgS2Oecc/v82/8JX3KPc26Fc26Jf7ubgVeAszKxT/c65/72x3MM59xrwEbgW6ACvn8qRERCipJcEclru4HSGfSKVgR+S/P4N/9Y6jqOS5IPAkWzGohz7gDQDxgNbDezaWZWJxPxHI0pMs3jHVmIZ7dzLsV//2gS+kea5w8dfb2Z1TKzqWa2w8z+wlepPmErRBq7nHOHM1jmNaAB8Lxz7u8MlhWRQOT1SWc68UxE5BiLgb+BXukssw3fR+1HVfaPZccBoHCax+XTPumc+9I5dy6+iubP+JK/jOI5GlNcNmPKipfwxXWGc64YcAeQ0V8Wl96TZlYU34l//wPG+NsxRERCipJcEclTzrm9+PpQX/CfcFXYzPKZ2flm9rh/sQ+Au8ysjP8ErnuA8SdbZwZWA+3MrLL/pLfbjz5hZuXMrKe/N/dvfG0PR06wji+AWv5pzyLMrB9QD5iazZiy4jTgL2C/v8p8+XHP/wFUz+I6nwOWO+dG4Os1fvmUoxQRCTBKckUkzznnngJuwHcy2S5gC3AV8Ll/kQeB5cB3wPfASv9YdrY1C5joX9cKjk1Mw/xxbAP+xNfrenwSiXNuN9ANuBFfu8UtQDfnXHx2Ysqim/Cd1LYPX5V54nHPjwHe9s++0DejlZlZT+A8/tnPG4CYo7NKiEiQMPN+ZoUAn13BnEv3Uy0RERERCTBhJau6Au3vynjBXHb488tWOOeaeB3HiWiScBEREZFgFOAnfnktsOvMIiIiIiLZoCRXREREREKO2hVEREREgpCpXSFdSnJzgOUv6qxQaEwz2bhmWa9DyDGhdEplKJ0fGqbfyQEpJZS+yYBw/fGXXLZy5Yp451wZr+OQk1OSmwOsUCkKtE7vMvXBY9Hka7wOIcckp5xoutPglHIkdBKQAvnCvQ5BTuDA4fSushx8ihTUnzfJXYXy2fFXQZQAo98CIiIiIkHGULtCRnTimYiIiIiEHCW5IiIiIhJy1K4gIiIiEmzMf5OTUiVXREREREKOKrkiIiIiQcd04lkGVMkVERERkZCjJFdEREREQo7aFURERESCkNoV0qdKroiIiIiEHCW5IiIiIhJy1K4gIiIiEoTUrpA+VXJFREREJNeY2WYz+97MVpvZcv9YKTObZWYb/F9L+sfNzMaa2UYz+87MYtKsZ4h/+Q1mNiSj7SrJFREREQlCZub5LQvOds5FO+ea+B/fBsx2zp0BzPY/BjgfOMN/Gwm85N/XUsC9QHOgGXDv0cT4ZJTkioiIiEhe6wm87b//NtArzfg7zmcJUMLMKgCdgVnOuT+dc3uAWcB56W1ASa6IiIiI5CYHzDSzFWY20j9Wzjm33X9/B1DOfz8S2JLmtVv9YycbPymdeCYiIiISbMx/817po322fq865149bpk2zrk4MysLzDKzn9M+6ZxzZuZyOjBVcj3285tDWfbiQJY8P4CFz/UD4J5LWrD0Bd/YlAd7UaFUEQD6t6/N0hcGsuzFgcx98iLOrFY6dT3Fi+Tn/Tu6sPqVi1n18sU0r1Pek/3JjISEBAb0u5BGDeoQfWZdlixe7HVI6bp85HCqVSpPs5iGqWN//vknPbp0Irp+bXp06cSePXtSn1swfx6tmsXQtPGZnNfxbC9CPqkrR42gZpUKtGzSKHXs++/WcG771rRqGk2/C3ry119/ATB39izOatWMVk2jOatVM+bPm+NV2Fk29tlniGlUn9joBgy+eACHDx/2OqRs2bJlC507nk3jhvWIaVSfcWOf8zqkDB0+fJhz27fkrJYxtG7aiEcfuu+Y52+/+TqqlC+R+vjO226kfatY2reKpVl0PapHlSYYzPxyBg3r16Z+nZo88fijXoeTJaNGDKNyxbLERjdIHbvv3rtp2rghzWOj6XZ+J7Zt2+ZhhNkXzMcliMU755qkuR2f4OKci/N/3Ql8hq+n9g9/GwL+rzv9i8cBldK8PMo/drLxkzLncjxx/s8JK17ZFWh9U7Ze+/ObQ2l97QR2//XPH+HTCuVn36FEAK7o0Yg6lUtxzbi5tKhbnp+37CFh/990alKFuwY1p931HwLw2g3nsujHbbz15Y/kiwijcIEI9h5IzHI8eyZfk639yIoRlw6hdZu2XDp8BImJiRw8eJASJUrk+HaSU47kyHoWLviaokWLMnL4UJau/A6Au+64lZIlS3Hjzbfy1BOPkZCwhwceepSEhAQ6tm/DZ5O/oFLlyuzauZMyZcuecgwpR3Lm53TRwq8pUqQol192KYuXrwHg7DYteOCRx2jT9izefftNftu8ibvuvZ81q1dRtmw5KlSsyE8//sAFPbqw9pffTzmGAvnCT3kd6YmLi6ND+zas+u4nChUqxKABfTnvvC5cMmRorm43N2zfvp0d27fTOCaGffv20ap5LB9+/Dl169XL8W0dOJycI+txznHgwAGKFi1KUlISXTudxcOPPU2TZi1YtXI5r770PF9MmcRvOxL+9drXXh7H92tWM/al1085jiIFc++DypSUFM6sV4tp02cRGRVFmxZNeXv8B7lyXHLDwgW+3wMjhg1mxeofAPjrr78oVqwYAC88P5af1/7E8y++7GWYWZbXx6VQPluR5iSqPBd+ejVXtPP9Xm0+1V8fDE73fTCzIkCYc26f//4s4H6gA7DbOfeomd0GlHLO3WJmXYGrgC74TjIb65xr5j/xbAVwdLaFlUCsc+7Pk21bldwAdDTBBShcMB9H/xFZsnYHCfv/BmDpzzuIPL0oAMUK56dNg4q89eWPACQlH8lWgpsX9u7dy8KFXzN02HAA8ufPnysJbk5q07YdJUuWOmZs2pTJDLp4MACDLh7M1MmTAPho4gf06NmbSpUrA+RIgpuTWrdpR8lSx+7LLxvX07pNOwDO7tCRKZM+A6BRdGMqVKwIQN169Tl0+BB///133gacTcnJyRw6dMj39eDB1P0INhUqVKBxjO/3+WmnnUadOnXZti3dwoXnzIyiRX2/m5KSkkhKSsLMSElJYcxdt3HvAyevrn360UT6XNQ/r0LNtmVLl1KjRk2qVa9O/vz5uahff6ZOmeR1WJnWpm07Sh33e+Boggtw8OCBoJx/NdiPS1YZ3s+skMnvk3LAQjNbAywFpjnnZgCPAuea2Qago/8xwBfAr8BG4DXgCgB/MvsAsMx/uz+9BBfUk+s55xxTHuyFc/C/6d/zxgxfojpmcEsGdajD3gOJnHfbp/963dBO9fhyxW8AVC1fjPi9h3j1+o6cWb0Mqzbu5KaX53Pw75ypzOSkzZs2Ubp0GUYOv5Tvv1tD45hYnnzmOYoUKeJ1aFmya+cflK9QAYBy5cuza+cfAGzcsJ6kpCTOP/cc9u/fx+VXXs1AfzIcqOrUrce0KZPp1qMnn3/6MXFbt/xrmcmff0qj6MYUKFDAgwizJjIykuuuv4la1StTqFAhOnTsRMdzO3kd1in7bfNmVq9eRdNmzb0OJUMpKSl0aNuMTb/+wrDLLie2aXNeeXEs53XpRvnyFU74mi2//8Zvv22m7VmB1eJzItu2xREV9c+nppGRUSxd+q2HEeWMe+++k/fGv0Px4sWZMWuu1+FkWagel2DnnPsVaHSC8d34qrnHjzvgypOs6w3gjcxuO9cruWZ2un/y39VmtsPM4tI8zp/O66qa2Q8nee5+M+t4kueGmlnF48b6m9mdZtbezFqd2h7lrA43f0yraybQ655JjOrWkNYNfKGPeWcxZwx5kwnz1jG6e8NjXtOuYRRDOtXnrjcWARARHkZ0zbK89sX3tLz6Aw4eTuKmvp59gpKu5ORkVq9ayWWjLmfJ8lUULlKEJ4O8byrtf7PJycmsWrWSjz+fwmdTpvP4Iw+xYcN6jyNM37iXX+d/r73EWa2asX/fPvLlP/bHcu1PP3LvXbfz7PMveRRh1uzZs4epUyaxdsMmfv19GwcOHuCD98Z7HdYp2b9/PwP6XsATTz17TMUtUIWHhzPvmxV89/NmVq5YxjcLFzD5s0+4bPRVJ33NZx9/SI9efQgPz912Fjm5+x54iI2bttB/wCBefnGc1+FIJnhdxQ30in+uJ7nOud3+yX+jgZeBZ44+ds5l6zN159w9zrmvjh83s3BgKHD8Z5PnAzOA9kBAJbnbdh8AYNfeQ0xe/CtNa5U75vmJc3+mV+uaqY8bVD2dl67twEUPTOXPfb4+3rj4/cTF72fZOl818bOFG4muUSaP9iBrIqOiiIyKollzXzWq9wUXsnrVSo+jyroyZcuxY7tv5pMd27dTuoyvLaFiZBQdO3aiSJEilC5dmlZt2vLDd2u8DDVDtWrX4bMpM5j/zVIu7NufatWqpz4Xt3UrF/e/kJdff5Nq1Wt4GGXmzZn9FVWrVqNMmTLky5ePXr36sGTxN16HlW1JSUkM6HsB/QYMolfvPl6HkyXFS5SgTbv2LFwwj02//kLTRnVoXL8mBw8epGmjOscs+9knE+lzYT9vAs2iihUj2ZrmE4+4uK1ERqY7k1FQ6TdgEJ9/9onXYWRZqB8XybqA6Mk1s/pmttRf3f3OzM7wPxVuZq+Z2Y9mNtPMCvmXf8vMLvTf32xmj5nZSmAA0AR4z7+uQub7NyMa+BMYDVzvf66tv1o8x7/N2WZWOc36Xzaz5Wa23sy65cZ+Fy4QQdFC+VLvd2xcmR9/+5MaFYunLtOtRXXWb/WduV+pTFEm3NWV4U9+yca4hNRl/thzkK279nFGZAkA2kdX4uff021T8Uz58uWJiqrE+nXrAJg3ZzZ16gbHyRppdenWnffGvwPAe+PfoWv3HgB07d6Dxd8sIjk5mYMHD7J82VJq16nrZagZ2rXTd0LrkSNHeOKxh7l0xCjANwtG3wt6cO/9D9OiZWsvQ8ySSpUqs3TpEg4ePIhzjrlzZgf8MTgZ5xyjLxtO7Tp1ufb6G7wOJ1Pid+1ib0ICAIcOHWL+nK9oFB3DT79sZdWPG1n140YKFy7MsjX/zCC0Yd3PJCQk0LR5S4+izpomTZuyceMGNm/aRGJiIh9NnEDXbj28DuuUbNywIfX+1MmTqFW7TjpLB6ZQPC5yagKlJ3c08Jxz7j1/C0M4vkblM4ABzrnLzOxD4ALgRJ877nbOxQCY2QjgJufc0WsjxwBrnHObzOxlYL9z7kn/c1OAt51zb5vZMGAs/1xxoyq+KS5qAHPNrKZzLnUKBPNNZuyb0LhguleVO6myJQsz8a6ugK/lYOK8dcxa8Rsf3NmFMyJLcsQ5ft+5j2vG+aZuun1gc0qdVpBnr/D1rCUfOUKbaycCcMPL83nzls7kjwhn8469jHzmX4XugPH0s89z6eBBJCYmUrV6dV59/U2vQ0rXpZcMZMGC+eyOj6d2jcrccde93HDTrQwZ1J9333qDSpWr8PZ7EwCoU6cuHTt1pkWTaMLCwhhy6XDq1W+QwRbyzvAhg1j49Xx2746nXs0q3HbXvRw4sJ/XX/G1InTv2YuLBw8F4LWXX2DTLxt5/JEHefyRBwH4bMr0gDuZ7njNmjend58LadkshoiICBo1aszwy0Zm/MIA9M2iRbz/3rs0aHAmzWOjAbjvwYc57/wu3gaWjj/+2M5Vo4aRkpLCkSOOnn0upPP5XdN9zaeffEjvC/oG/EefR0VERPDMc+Po3rUzKSkpDBk6jHr163sdVqYNvngAC+bPIz4+nhpVo7j7nvuYMeMLNqxfR5iFUblKFca+EFwzK0DwH5fsCJafGa/k6RRiZjaGNElmmvGBwJ3AO8CnzrkNZlYV3+XbzvAvcyuQzzn3oJm9BUx1zn1sZpuBs5xzv/mXm8exSe4dwCbn3AfHb9/M4oEKzrkkM8sHbHfOlfav/2t/gzNm9jVwjXNu9Yn261SmEAs0eTGFWF7JqSnEAkFOTSEWCHJ7CjHJnpyaQixQ5OYUYiLg/RRiEadXd8W6POjV5lPtGT/I0/chPZ60K5hZ7zQnnzVxzr0P9AAOAV+Y2Tn+RdPOV5TCySvPB9LZXCdgZjbCPD6rCJ0sQ0RERCTEefKvrnPuM3xXvADAzKoDvzrnxvr7YhvimyMtO/YBp/nXWxyI8E9TcfS5tKcmfwP0B94FBgEL0jx3kZm9DVQDqgPrshmPiIiISI5Tu0L6AuLEM6Av8IOZrQYa4GtbyK63gJf96+oBpG1OnQIcrSK3Ba4GLjWz74BLgGvTLPs7vkmLpwOj0/bjioiIiEhgy9NKrnNuzEnGH+WfK10c9Se+hPfoMk+muT80zf2qx63rE+ATADN7HXg9zXPr8VWJ0zqHE/vKOTf6JM+JiIiISAAL6c5859wIr2MQERERyXHmv8lJhXSSm11pK8UiIiIiEnyU5IqIiIgEIZ14lr5AOfFMRERERCTHKMkVERERkZCjdgURERGRIGOY2hUyoEquiIiIiIQcJbkiIiIiEnLUriAiIiIShNSukD5VckVEREQk5KiSKyIiIhKMVMhNlyq5IiIiIhJylOSKiIiISMhRu4KIiIhIsDGdeJYRVXJFREREJOQoyRURERGRkKN2BREREZEgpHaF9KmSKyIiIiIhR0muiIiIiIQctSvkgMY1y7Jo8jVeh5EjSja9yusQcsyeZeO8DiHHhJnzOgQJcUUK6s+BSLBRu0L6VMkVERERkZCjf91FREREgoxhquRmQJVcEREREQk5SnJFREREJOSoXUFEREQkGKlbIV2q5IqIiIhIyFGSKyIiIiIhR+0KIiIiIsHGNE9uRlTJFREREZGQo0quiIiISBBSJTd9quSKiIiISMhRkisiIiIiIUftCiIiIiJBSO0K6VMlV0RERERCjpJcEREREQk5alcQERERCUbqVkiXKrkiIiIiEnJUyRUREREJQjrxLH2q5AaBUSOGUbliWWKjG3gdSrp+nnYfyz68gyUTbmPhe7cAcGatSOa9fSPLPryDj58dxWlFCgJwTvM6LHrvFpZ9eAeL3ruFs5rWSl1Pvohwxt01gO8+v4fVn95Frw7RXuxOpqSkpNCiSWP69OzmdSjZkpCQwKD+F9H4zLrENKzHt0sWc/+Yu2kW24gWTRvTvUtntm/b5nWYWbJlyxY6dzybxg3rEdOoPuPGPud1SKdk5pczaFi/NvXr1OSJxx/1OpxsO3z4MG1aNqNZTCNiGtXngfvu9TqkUxIqxyVY/r5kVqgcF8kZ5pzzOoagFxvbxC36dnmurX/hgq8pUqQoI4YNZsXqH3JtOwAlm16V7df+PO0+Wg96nN0JB1LHFo6/mdue+YyFKzYyuGcLqkaezv0vTqNR7Sh2/rmP7bv2Uq9GBaa8eCU1Ot8FwF2juxAeFsZ9L07FzChVvPAx68ysPcvGZXtfMuu5Z55m5crl7PvrLz6dNDXXtnPkSO78nF42fCitW7dh6LARJCYmcvDgQcLCwihWrBgAL44by89rf2LsCy/n2DbDwnK38rB9+3Z2bN9O45gY9u3bR6vmsXz48efUrVcvV7ebG1JSUjizXi2mTZ9FZFQUbVo05e3xHwTlvjjnOHDgAEWLFiUpKYlzzmrDk08/R/MWLbwOLctC6bjk5d+X3JbXx6VQPlvhnGuSKyvPhPxla7ry/Z72avOptozr6en7kB5VcoNAm7btKFWqlNdhZEvNymVZuGIjAHOW/JxalV2zbivbd+0F4KdftlOwQD7y5/N1zwzp2ZIn3pgJ+P4wZifBzQtbt25lxvRpXDpshNehZMvevXtZtOBrhlw6HID8+fNTokSJ1AQX4MDBA0H3cViFChVoHBMDwGmnnUadOnXZti3O46iyZ9nSpdSoUZNq1auTP39+LurXn6lTJnkdVraYGUWLFgUgKSmJ5KSkoPveOiqUjksw/305Xigdl8wws4C4BTIluZJjnHNMefEqFr13C8P6tAZg7a/b6d6+IQB9zo0hqlzJf72ud8doVv+8hcSkZIoXLQTAvVd245v3b+W9x4dRttRpebcTWXDzjdfx0COPExYWnD9GmzdvonSZMoy6bBgtm8VwxegRHDjg+4dizD13UqtGZSZ+8D533Xu/x5Fm32+bN7N69SqaNmvudSjZsm1bHFFRlVIfR0ZGERcXnAk7+CptzWOjqVyxLOd0PJdmzXVcJOfouMjxAvavs5mdbmar/bcdZhaX5nF+r+OTf+tw6TO0GvgYva56kVH92tI6pgajxrzHyL5tWfTeLRQtXIDEpJRjXlO3enkevKYnVz04AYCIiDCiypdkyZpfaTXwMb79bjOPXN/bi91J1xfTplK2TFliYmO9DiXbUpKTWb1qJZeNHM3ipSspXLgITz3h62Ebc/9DrP/ld/oNGMgrL+V+20du2L9/PwP6XsATTz17THVavBMeHs63K1azcfNWli9byo8/BPfH4yIS2AI2yXXO7XbORTvnooGXgWeOPnbOJZpZns4MYWbhebm9YLTN336wa89+Js/5jqb1q7J+8x90v+IFWg96nA9nrGDT1l2py0eWLcHEp0cy4u532bQ1HoDdCQc4cOhvPp+9BoBPZ60kum6lf2/MY4u/WcTUqZOpXbMqgwf1Z97cOVw6+GKvw8qSipFRREZFpVY5e/e5kNWrVh2zTP/+g/j8s0+9CO+UJCUlMaDvBfQbMIhevft4HU62VawYydatW1Ifx8VtJTIy0sOIckaJEiU4q/3ZzJw5w+tQsiVUj0uw+y8eF69bFdSukIPM7C0ze9nMvgUeN7NoM1tiZt+Z2WdmVtK/3Dwza+K/X9rMNvvv1zezpf5q8HdmdoZ//OI0468cTWjNbL+ZPWVma4CWnux0kChcMD9FCxdIvd+xZR1+/GUbZUr6evDMjNsu68xrHy8EoHjRQnz6/GjuHjuJxWt+PWZdX3z9A+2anAFA+2a1+fnX7Xm4J5nzwEOP8MvmrazbuJl33ptA+7PP4c13xnsdVpaUL1+eqKhKrF+3DoB5c2dTp25dNm7YkLrM1CmTqF27jlchZotzjtGXDad2nbpce/0NXodzSpo0bcrGjRvYvGkTiYmJfDRxAl279fA6rGzZtWsXCQkJABw6dIjZX80Kuu+to0LpuIQSHRc5XjDOkxsFtHLOpZjZd8DVzrn5ZnY/cC9wXTqvHQ0855x7z9/yEG5mdYF+QGvnXJKZvQgMAt4BigDfOuduPH5FZjYSGAlQqXLlHNy9fxt88QAWzJ9HfHw8NapGcfc99zF02PBc3WZWlT39NCY+fRkAEeHhTJy+nFnfrOXKAe0Z1a8dAJPmrOadSUsAGN2/HTUqleH2kedz+8jzAeh++Th27dnPXc99zv8eHMITN11A/J79jBoTXMljMHnymbEMG3oxiYmJVKtWnZdfe4MrR1/G+vXrCAsLo3LlKowd95LXYWbJN4sW8f5779KgwZk0j40G4L4HH+a887t4G1g2RERE8Mxz4+jetTMpKSkMGTqMevXrex1WtuzYvp3Lhg0hJSWFI+4IF1zYly5dg3PqvVA6LsHw9yWzQum4SM4IiinEzGwMsB9oAMx1zr1tZsWB751zlf3L1AA+cs7FmNk84Cbn3HIzKw0sd85VNbOBwJ34EthPnXMbzOwq4A5gp39zhYAPnHNjzCwZKOCcO7aR9Di5PYVYXjqVKcQCTV5MIZZXcmsKMS/k9hRiIiJ5wespxAqUO8NVHPCsV5tPtfm5bgE7hVgwVnIzM59UMv+0YhQ8Ouice9/f6tAV+MLMRuG78vPbzrnbT7CewxkluCIiIiISeIKqJzct59xeYI+ZtfUPXQLM99/fDBw97f3Co68xs+rAr865scAkoCEwG7jQzMr6lyllZlVyfw9EREREToEFwC2ABWMlN60hwMtmVhj4FbjUP/4k8KG/b3ZamuX7ApeYWRKwA3jYOfenmd0FzDSzMCAJuBL4La92QkRERERyVlAkuc65MScZXw3865qQzrmf8VVpj7rLP/4o8K+LWTvnJgITTzBeNFsBi4iIiIingiLJFREREZFjBfo8tV4L2p5cEREREZGTUZIrIiIiIiFH7QoiIiIiwcbUrpARVXJFREREJOSokisiIiISZAxQITd9quSKiIiISMhRkisiIiIiIUftCiIiIiJBx3TiWQZUyRURERGRkKMkV0RERERCjtoVRERERIKQuhXSp0quiIiIiIQcJbkiIiIiEnLUriAiIiIShDS7QvpUyRURERGRkKNKroiIiEiwMZ14lhFVckVEREQk5CjJFREREZGQo3aFHOCAI0ec12HkiD3LxnkdQo4p2eVJr0PIMX9Ou9HrEHLM4cQUr0PIMQXzh3sdQo45FELHBaBQCB0bkRMxICxM/QrpUSVXREREREKOklwRERERCTlqVxAREREJQppdIX2q5IqIiIhIyFElV0RERCQI6Ypn6VMlV0RERERCjpJcEREREQk5alcQERERCTa6rG+GVMkVERERkZCjJFdEREREQo7aFURERESCjKHZFTKiSq6IiIiIhBxVckVERESCjqmSmwFVckVEREQk5CjJFREREZGQo3YFERERkSCkboX0qZIrIiIiIiFHSa6IiIiIhBwluQEsJSWFls1iuKBXdwDmzZ1Dq+axNGl8JpcNH0pycrLHEWbdli1b6NzxbBo3rEdMo/qMG/uc1yGd0M/vXMayV4aw5KXBLBx38THPXXtBEw7NvInTixUCoFalUsx7diAJU6/jugubHLPslb1iWP7qUFa8OpSresfkWfyZcfjwYdq2ak7z2GhiGzXggfvuBWDzpk20a92CBnXP4JKB/UlMTPQ40n/bunUL3c/vQIvYM2nZpCEvvzAWgIfuv4fWzRrTtkUsfbqfx/bt21Jfs/DrebRtEUvLJg3p2vlsr0LPklEjhlG5Yllioxt4HUqmbN26hR4nOC6PPnQf9WtWpl2LWNq1iGXWjC8A+GjC+6lj7VrEcnrRfHy/ZrWHe5B5M7+cQcP6talfpyZPPP6o1+GckoSEBAb0u5BGDeoQfWZdlixe7HVI2RZKxyUzzMzzWyAz55zXMQS9mNgmbuHiZTm+3rHPPs3KlSvY99dffPTpJOqcUZVp07/ijFq1eOC+e6hcuQpDLh2eo9sMC8vdb9jt27ezY/t2GsfEsG/fPlo1j+XDjz+nbr16Ob6tkl2ezPZrf37nMlpfNZ7dfx06ZjyqzGm8eH0nalc6nVZXvsvuvw5RpkRhKpctRvdWNUnYf5hnP14OQL2qpXnnjm60vXo8iUkpTH74Qq4eO4tftyVkOZ4/p92Y7X05GeccBw4coGjRoiQlJdGhfVuefPpZxj77DD179eaifv25+srRnNmwESNHXZ5j2/076cgpr2PH9u38sWM7jRr7vo/ObtOM8RM+oWJkFMWKFQPglRef5+ef1/LM2BfZm5BA5w5t+ejzaVSqVJldO3dSpmzZU46jYP7wU15HehYu+JoiRYoyYthgVqz+IVe3dSgx5ZTXcfxxOadNM96d8Amff/oRRYoU5errTv59/NMP33Nx/wtY+cP6U44DoFAuHpuUlBTOrFeLadNnERkVRZsWTXl7/Ae58nssL4y4dAit27Tl0uEjSExM5ODBg5QoUcLrsLIsr49LoXy2wjnXJOMlc0fhirVd7VEvebX5VKvHdPD0fUiPKrkBKm7rVmZM/4Kh/iR29+7d5M+XnzNq1QLgnA7n8vlnn3oZYrZUqFCBxjG+iuZpp51GnTp12bYtzuOoMu/x0Wdz5+tfk/afw10JB1mxfgdJKccmb3UqlWLZz9s59HcyKUccC77fQq/WZ+R1yCdlZhQtWhSApKQkkpKSwIz58+bQ+4ILAbj4kiFMnTzJyzBPqHyFCjRq/M/3Ua3addi+LS41wQU4cOBAapXhow8/oFuPXlSqVBkgRxLcvNCmbTtKlSrldRiZdrLjkhmffDSBPhf2zc3wcsyypUupUaMm1apXJ3/+/FzUrz9TpwTez0lm7N27l4ULv2boMN/fmvz58wdlgguhdVwkZyjJDVC33HQ9Dz3yGGFhvkNUunRpklOSWbnCVyX87NOP2bp1i5chnrLfNm9m9epVNG3W3OtQ/sUBUx65kEUvXMywLg0B6NayBtvi9/H9r7sytY4fN8fTukEkpU4rSKECEZzXtDpRZU7LxaizLiUlheZNGlMlshwdOnSkevUaFC9RgogI38QrkZFRbIsL7H9Cfv9tM9+tWU1sU9/30QNj7qJ+rap8NPED7rhrDAC/bNhAQkIC3c47h/atmzHhvXc9jPi/4fjj8vorL9KmWWOuGj2ChD17/rX8Z598RJ+L+ud1mNmybVscUVGVUh9HRkYRF+A/JyezedMmSpcuw8jhl9KiSWMuHzmCAwcOeB1WtoTScckU882u4PUtkAVUkmtmKWa22sx+MLOPzKxwBsvPM7Mm/vubzax03kSau6ZPm0qZMmVoHBObOmZmvP3uB9x68w20a92c0047jfDw3P2oNDft37+fAX0v4Imnnj2m+hYoOlz/Aa2ufJded37KqO7RtD4zilsGtOD+txdleh3rtvzJUx8uZcqjFzL54QtY88tOUo4EVntQeHg43y5fxYZNW1i+fBnr1/3sdUhZsn//fgYP7Msjjz+d+n1095gH+XH9Zi7qN4DXXnkBgOSUZNasWsHET6bwyaQveOKxh9i4IWc+Fpd/279/P0MG9uVh/3EZNmI0K39Yz9dLVlC+fHnuuv3mY5ZfvuxbChUqTL36wdF7HEqSk5NZvWoll426nCXLV1G4SBGe/A/0ssp/Q0AlucAh51y0c64BkAiM9jogAPPJs/dq8eJFTJs2hbq1qjHkkgHMnzeHYUMvoXmLlsya8zVfL/qW1m3accYZtfIqpByVlJTEgL4X0G/AIHr17uN1OCe0bfd+wNeKMPmbjbRtGEWV8sVZ+vIQfn7nMiLLnMbiFy+hXMl0/w/j7Rk/0PrK8Zx740QS9h9mQ9y/K1iBoESJErQ7qz3fLlnM3oSE1JMa4+K2UjEy0uPoTiwpKYkhAy/ion4D6N6z97+ev6j/QCZ//hkAFStGck7HThQpUoTTS5emVeu2/PD9d3kd8n/C0eNyYZrjUrZcOcLDwwkLC2PwpSNYufzYcxg+/WgiF/Tt50W42VKxYuQxn6TFxW0lMkB/TjISGRVFZFQUzZr7Ku69L7iQ1atWehxV9oTScckMQyeeZSTQkty0FgA1zay9mU09Omhm48xsaHovNLMb/NXgH8zsOv/Yo2Z2ZZplxpjZTf77N5vZMjP7zszu849VNbN1ZvYO8ANQ6QSbyhX3P/gIG37dwtr1m3j73Q84q/05vPHWu+zcuROAv//+m6effJzhl43Kq5ByjHOO0ZcNp3adulx7/Q1eh3NChQvmo2ihfKn3O8ZUYcW6HVTp+yJ1Br9GncGvEbdrHy2veJc/9hxMd11lSviS4EplTqNnmzOYOGdtrsefWbt27SIhIQGAQ4cOMWf2V9SuU5d2Z53NZ598DMD4d9+ma/ceHkZ5Ys45rr78MmrVrsuV11yfOv7Lxg2p96dPnUyt2rUB6NKtB0u+WURycjIHDx5k+bKl1KpdJ8/jDnXOOa45wXHZsX176v2pkz+nbv36qY+PHDnCpE8/ps+FwZPkNmnalI0bN7B50yYSExP5aOIEunYLvJ+TzChfvjxRUZVYv24dAPPmzKZO3eA8gS6UjovkjIC84pmZRQDnAzOy8dpY4FKgOb5/dL41s/nAROBZ4AX/on2BzmbWCTgDaOZffrKZtQN+948Pcc4tOcF2RgIjASpVrpzVMLPl2aefYMYX0zhy5AgjRo6m/dnn5Ml2c9I3ixbx/nvv0qDBmTSPjQbgvgcf5rzzu3gbWBplSxRm4r09AYgID2Pi3LXMWr75pMuXK1mYReMu4bTC+TniHFf1jqXxZW+y72AiH9zdg1LFCpGUnMJ1z89m74G/82gvMrZj+3YuGz6UIykpHDlyhD4XXkSXrt2oW7cegy8ewH1j7qZRo8apJz8GkiWLFzHxg/HUq38mbVv42nruHvMA4995kw3r1xMWFkalypV5euyLANSuU5cO53amTfPGmIUxeOiwoPhofPDFA1gwfx7x8fHUqBrF3ffcl3qCUCD6Ns1xaZfmuHzy0US+/24NZkblKlV4euw/Z4R/s/BrKkZFUbVada/CzrKIiAieeW4c3bt2JiUlhSFDh1EvTeIebJ5+9nkuHTyIxMREqlavzquvv+l1SNkSasdFTl1ATSFmZinA9/6HC4AbgVbATc65bv5lxgHLnXNvmdk8/3PLzWwz0AQYBJzunLvHv/wDwC7n3FgzWwt0AMoALzrnWpvZk8CFQIJ/u0WBR4DZwFznXLWM4s6tKcS8kNtTiOWlU5lCLNDkxhRiXsmJKcQCRW5PIZaXcmIKsUCSm1OIiYD3U4gViazt6l7+slebT7Xi7nMCdgqxQKvkHnLORacdMLNkjm2rKHgK6/8IX0JbHl9lF3zV20ecc68ct92qQHCeYioiIiLyHxfIPblH/QbUM7MCZlYCXyU2PQuAXmZW2MyKAL39Y+BLbPvjS3Q/8o99CQwzs6IAZhZpZsExiaaIiIhIEDCzcDNbdfQ8KzOrZmbfmtlGM5toZvn94wX8jzf6n6+aZh23+8fXmVnnjLYZ8Emuc24L8CG+k78+BFZlsPxK4C1gKfAt8LpzbpX/uR+B04A459x2/9hM4H1gsZl9D3zsX0ZEREQkYHk9s0IWZ1e4Fkh79vVjwDPOuZrAHuDoCQfDgT3+8Wf8y2Fm9fAVKusD5wEvmlm6fUkB1a7gnCt6kvFbgFtOMN4+zf2qae4/DTx9knWdeYKx54DnTrB44J+ZIiIiIhLAzCwK6Ao8BNxgvuz4HGCgf5G3gTHAS0BP/33wFR7H+ZfvCUxwzv0NbDKzjfgmDVh8su0GVJIrIiIiIpkTINPUljaz5Wkev+qce/W4ZZ7FV6w8+kn56UCCcy7Z/3grcHRS40hgC4BzLtnM9vqXjwTSznaV9jUnpCRXRERERLIrPr3ZFcysG7DTObfCzNrnWVQoyRURERGR3NMa6GFmXfDNkFUMX4toCTOL8Fdzo4A4//Jx+C7AtdV/3YTiwO4040elfc0JBfyJZyIiIiJyHAuOE8+cc7c756L85071B+Y45wYBc/HNdgUwBJjkvz/Z/xj/83Oc76IOk4H+/tkXquG7YNfS9LatSq6IiIiI5LVbgQlm9iC+mbP+5x//H/Cu/8SyP/ElxjjnfjSzD4GfgGTgSudculexUZIrIiIiIrnOOTcPmOe//yu+2RGOX+YwcNFJXv8QvhkaMkVJroiIiEiQMQJmdoWApZ5cEREREQk5SnJFREREJOSoXUFEREQk6GT5srr/OarkioiIiEjIUSVXREREJAipkJs+VXJFREREJOQoyRURERGRkKN2BREREZEgpBPP0qdKroiIiIiEHCW5IiIiIhJy1K4gIiIiEmxMsytkRJVcEREREQk5quTKMZxzXoeQY/6cdqPXIeSYUmff7XUIOWbnV/d7HYKcQIEI1TxEgomhE88yot9qIiIiIhJylOSKiIiISMhRu4KIiIhIEFK7QvpUyRURERGRkKMkV0RERERCjtoVRERERIKQuhXSp0quiIiIiIQcVXJFREREgpBOPEufKrkiIiIiEnKU5IqIiIhIyFG7goiIiEiwMZ14lhFVckVEREQk5CjJFREREZGQo3YFERERkSBjmGZXyIAquSIiIiIScpTkioiIiEjIUbuCiIiISBBSt0L6VMkVERERkZCjJDdA1a1VjaYxDWnRtDFtWjY95rnnnnmKIgXCiI+P9yi6zDt8+DBtWzWneWw0sY0a8MB99wLw0ovjaFD3DArnD479OCoY9+fnj25k2dtXseTNK1n4+uWp45df0ILV713Linev5qHLO6eON6hRjnkvj2TFu1ez7O2rKJDf94HPpKcG8+1bV7Li3asZe1MPwsK8LSFcMWo41SuXp3lsw3899/yzT1OsUDi70xyLBV/Po3XzGJrFnMn5556dl6GekplfzqBh/drUr1OTJx5/1OtwsiwhIYFB/S+i8Zl1iWlYj2+XLAbgpReep/GZdWkS3YA7b7/F4yizLtiPS1pjn32GmEb1iY1uwOCLB3D48GGvQ8qWUSOGUbliWWKjG3gdSp4JM/P8FsjUrhDAps+cQ+nSpY8Z27plC7O/mkWlypU9iiprChQowPSZsylatChJSUl0aN+WzuedT8uWrenSpRudgyjZgODdn/OueYPdew+mPm7XuBrd2tal2dBxJCalUKZEEQDCw8N44+6LGP7gx3y/cQelihUiKTkFgIvvnsi+g38D8MGDA7jg7AZ8NPv7vN8Zv0GXDGHk6CsZNWLoMeNbt2xh9uyZVKr0z89IQkICN1x7FZ9O+oJKlSuza+fOPI42e1JSUrjumiuZNn0WkVFRtGnRlG7delC3Xj2vQ8u0m2+8jnM7dea9CR+RmJjIwYMHmT9vLlOnTGbJ8tUUKFCAnUFyPI4KheNyVFxcHC++MJZV3/1EoUKFGDSgLx9NnMAlQ4Z6HVqWXTJkKKOvuIoRwwZ7HYoECFVyg8ytN9/Ag488FjTThpgZRYsWBSApKYmkpCQwI7pxY6pUreptcNkQKvszsncznhz/NYlJvgR2V8IBADo2rckPv+zg+407APjzr0McOeIAUhPciPAw8uULxznnQeT/aN2mHSVLlfrX+O233MADDx37M/LRxA/o3rN36j+HZcqWzbM4T8WypUupUaMm1apXJ3/+/FzUrz9Tp0zyOqxM27t3L4sWfM2QS4cDkD9/fkqUKMHrr77MjTffSoECBQAoGyTH46hgPy7HS05O5tChQ76vBw9SoWJFr0PKljZt21HqBL8T5L9LSW6AMoweXTvTukUT3nj9VQCmTp5EhYoVadiwkcfRZU1KSgrNmzSmSmQ5OnToSLNmzb0O6ZQE2/44B1OeHsqi/13OsB5NAKhZqTStG1bh61dHMfP54cTWiQTgjEqn4xxMfmoI3/zvCm4Y2OaYdU1+agi/T72d/Qf/5tN5P+b5vmRk2pRJVKgYyZnH/Yxs3LCehIQ9dOl0Du1aNeX9997xKMKs2bYtjqioSqmPIyOjiIuL8zCirNm8eROly5Rh1GXDaNkshitGj+DAgQNs2LCebxYt4Kw2LejcsT0rli/zOtQsCfbjklZkZCTXXX8TtapXplqlChQrVpyO53byOizJJDPvb4EsqNsVzCwF+B7IByQD7wDPOOeOeBpYDvhq7gIqRkayc+dOunfpRK3adXji8UeYPO1Lr0PLsvDwcL5dvoqEhAT6X9SHH3/4gfoNgrdnKtj2p8MVr7Itfh9lShRh6rNDWfdbPBHhYZQqVoh2I1+hSd1Ixt/fn7p9nyIiIoxWDavQ5rKXOHg4ienPXcrKdduYt+JXAHrc+DYF8kfw1j0X0T6mOnOW/+Lx3v3j4MGDPPn4o3w+dca/nktOTmb1ypVMmT6Lw4cO0aF9a5o2a8EZZ9TyINL/jpTkZFavWslTz4ylabPm3HTDtTz1xKMkJyez588/mbdgMSuWL+OSgf34cd0vQfMJVSjZs2cPU6dMYu2GTZQoUYKB/S/ig/fGM2DQxV6HJnLKgr2Se8g5F+2cqw+cC5wP3Hv8QmYWdMl8xUhfZa1s2bL06NmLhQvms3nzJlo0jaZurWrEbd1K6xax7Nixw+NIM69EiRK0O6s9s2b+OwkJRsGyP9vi9wG+loTJX6+lab1I4nbt5fP5PwGwfG0cR5yjdInCxO38i4VrNrN770EO/Z3EjMXraVzr2I8u/05MZsrCtXRvWzfP9yU9m379hd9+20TrZo1pULs6cXFbaduyCX/s2EFkZBQdzu1EkSJFOL10aVq3acsP363xOuQMVawYydatW1Ifx8VtJdL/uyEYVIyMIjIqiqb+Tzt697mQ1atWERkZRY9efTAzmjRtRlhY4J2wmZ5gPy5pzZn9FVWrVqNMmTLky5ePXr36sGTxN16HJZIjgj3JTeWc2wmMBK4yn6FmNtnM5gCzzayImb1hZkvNbJWZ9QQws/r+sdVm9p2ZneFfdpqZrTGzH8ysX17uy4EDB9i3b1/q/dlfzSImtim/bf2Dtes3sXb9JiKjoli0ZAXly5fPy9CybNeuXSQkJABw6NAh5sz+ilq163gb1CkItv0pXDAfRQvlT73fsWlNfvx1J1O+XstZMdUBqFnpdPJHhBOfcJBZSzdQv3o5ChXIR3h4GG0bV2Pt5p0UKZSf8qf7epHDw8M4v2Vt1v22y7P9OpH6Dc7k19938MO6X/lh3a9ERkaxYPFyypUvT9fuPVjyzSKSk5M5ePAgy5ctpXadwErST6RJ06Zs3LiBzZs2kZiYyEcTJ9C1Ww+vw8q08uXLExVVifXr1gEwb+5s6tStS/cePfl6/lwANqxfT2JS4r9Osg1kwX5c0qpUqTJLly7h4MGDOOeYO2d2UPxsyNF2AfP8FsiCrsKZHufcr2YWDhw9iyEGaOic+9PMHgbmOOeGmVkJYKmZfQWMBp5zzr1nZvmBcKALsM051xXAzIofvy0zG4kvqc7xmQ52/vEH/fv2AXwf9/XtP4BOnc/L0W3klR3bt3PZ8KEcSUnhyJEj9LnwIrp07caL48by9FNP8MeOHTSLbUTn887npVde9zrcDAXb/pQtVZSJDw8EfCeMTZz1HbO+3UC+iHBeub03y9+5msSkFEY89AkACfsOM3biIha+Phrn4MvF65mxeD1lSxbh40cvJn++CMLCjK9X/sprk7zto7x08EAWLpjP7vh46tSozB1338vgocNPuGztOnXpeG5nWjaNJiwsjMFDh1OvfuC2mBwVERHBM8+No3vXzqSkpDBk6DDq1a/vdVhZ8uQzYxk29GISExOpVq06L7/2BkWKFGH0yOE0aXwm+fPn59XX3wr4P5ZphcJxOapZ8+b07nMhLZvFEBERQaNGjRl+2Uivw8qWwRcPYMH8ecTHx1OjahR333MfQ4ed+HeC/DeY12dInwoz2++cK3rcWAJQG1/rwlnOuUv948uBgvh6dwFKAZ2BxsCd+Pp5P3XObTCzWsBMYCIw1Tm3IL04YmKbuIWLg+vEiZMJor8z/ymlzr7b6xByzM6v7vc6hByTLyJkPgxLnUUjVHg9j7OEvkL5bIVzrolX2y9epa5rddtbXm0+1YwrWnj6PqQndH5DA2ZWHUgBjk66eCDt08AF/h7eaOdcZefcWufc+0AP4BDwhZmd45xbj68K/D3woJndk4e7ISIiIiKnKGSSXDMrA7wMjHMnLk9/CVxt/s/EzKyx/2t14Ffn3FhgEtDQzCoCB51z44En8CW8IiIiIhIkgr0nt5CZreafKcTeBZ4+ybIPAM8C35lZGLAJ6Ab0BS4xsyRgB/Aw0BR4wsyOAEnA5Sdco4iIiIhHgqmX3QtBneQ658LTee4t4K00jw8Bo06w3KPA8Rce/9J/ExEREZEgFDLtCiIiIiIiRwV1JVdERETkv0rdCulTJVdEREREQo4quSIiIiJBxgBDpdz0qJIrIiIiIiFHSa6IiIiIhBy1K4iIiIgEIV29On2q5IqIiIhIyFGSKyIiIiIhR+0KIiIiIsHGTJf1zYAquSIiIiIScpTkioiIiEjIUbuCiIiISBBSt0L6VMkVERERkZCjSq6IiIhIkDEgTKXcdKmSKyIiIiIhR0muiIiIiIQctSuIiIiIBCF1K6RPlVwRERERCTmq5OYAA8LC9O9UoNl7MMnrEHLMrtn3ex1Cjqk4dLzXIeSYne8O9jqEHLP/72SvQ8hRxQrl8zoEEfGYklwRERGRIKTL+qZP7QoiIiIiEnJUyRUREREJMmY68SwjquSKiIiISMhRkisiIiIiIUftCiIiIiJBSJf1TZ8quSIiIiIScpTkioiIiEjIUbuCiIiISBBSs0L6VMkVERERkZCjJFdEREREQo7aFURERESCkC7rmz5VckVEREQk5KiSKyIiIhJkDAhTITddquSKiIiISMhRkisiIiIiIUftCiIiIiLBxkwnnmVAlVwRERERCTlKcoNE7ZpVaRJ9Js1jo2ndvInX4ZySlJQUWjRpTJ+e3bwOJdNSUlLo2KYpF/ftBcAVIwbTOrY+Z7WI5rorLyMpKQmADet/pmvHtlQuU5QXxz7tYcQndvnI4VSrVJ5mMQ1Txx4Ycw8tmkTTqlkMPbt2Zvu2bQA457j5hmtpVK8WLZpEs3rVSq/CTvX92D4sfqw7Cx/pxryHugBQskh+Pr+jI6ue7sXnd3SkRJH8qcs/PqQpq5/pxTePdadR1VKp4/cNiGHJ491Z8nh3+rSomte7kSUzv5xBw/q1qV+nJk88/qjX4WRKSkoKHdo0ZdBFvQD43ysv0rxRXcoVy8/u3fHHLLtowXzOad2Eds0a0ev8Dh5Emz3BeFyOGjViGJUrliU2ukHq2IP3j6F6lUiax0bTPDaaGdO/8C7AUxDMx0VynpLcIDLjq7l8u2I1i75d7nUop2Tc2OeoXbeu12FkyWsvPc8ZteukPu7TdwALl//AvMWrOHzoEO+9/QYAJUqW4sHHnuHyq6/3KtR0DbpkCJ9NPvaP17U33MSS5av5ZulKzuvSjUcffgCAmV9O55eNG1j94zrGvvAy119zpRch/0vXB2fS5vaptL/Ttx/X92zA/B920PiGz5n/ww6u7+H7w90pOpIa5YsRff3nXPvaYp4Z3hyAzo0jaVStFK1vm8o5d0/nmm71OK1QPs/2Jz0pKSlcd82VTJoynVXf/cRHEz5g7U8/eR1Whl576XnOqPXPz0uzFi35aPJ0KlWucsxyexMSuO2Gq3lnwqd8vXQNr73zQV6Hmi3BelyOumTIUCZNnfGv8auvvZ5vV6zm2xWrOe/8Lh5EdmqC/bhkh5n3t0CmJFfy1NatW5kxfRqXDhvhdSiZti1uK199OZ1Bg4eljnXsdD7m74dqHNuU7du2AlCmTFkaxzYhIl9gJk1t2rajZMlSx4wVK1Ys9f6BAwdSe7ymTZnMgEGXYGY0a96ChIQEdmzfnqfxZkbX2Eq8//UvALz/9S90a1IJgC6xlfhggW982cZ4ihfOT7kShagdWYJv1v5ByhHHwb+T+eH3PXRsVNGz+NOzbOlSatSoSbXq1cmfPz8X9evP1CmTvA4rXdvitjLry+kMGvLPz8uZjRpTuUrVfy376UcT6NK9F1GVKgO+n59gEIzHJa02bdtRqlSpjBcMMsF+XCTnKckNEmZG9/M70apZLP977VWvw8m2m2+8joceeZywsOD51rv7thu5+/5HsBPEnJSUxMcT3uPsjp09iCzn3HfPXdSpUYUPJ7zPnffcB8C2bXFERlVKXSYyMopt2+K8ChHwtVB8fntH5j/UlaHnnAFAmeKF+CPhEAB/JByiTPFCAFQsVZituw+mvjbuz4NULFWYH377k46NIimUP5xSpxWgbb3yRJ5eJO93JhO2bYsj6rhjEBfn7THIyN233cg99z+SqZ/xXzZuYG9CAr27dOTcds358P138yDCUxeMxyUzXn5xHE0bN2TUiGHs2bPH63CyLFSPS3qOFlu8vAWy4Mk0ssDMypvZBDP7xcxWmNkXZlYri+soYWZX5FaMWTV73kIWL1vJ51On88pLL7Bwwddeh5RlX0ybStkyZYmJjfU6lEybOWMapcuUpVHjmBM+f9sNV9OidVtatGqTx5HlrHvvf5Cff/mNvv0H8upLL3gdzkl1HjODdndM44LHZnNZp9q0qvPvyp9zLt11zPl+OzNXxzHrvvN54+q2LNuwiyNH0n+NZM7M6dMoXfrkPy/HS0lOZs3qlYz/aBITPpvG048/wi8b1udylHIil426nJ/W/cK3K1ZTvkIFbrv5Rq9DEjllIZfkmu/fis+Aec65Gs65WOB2oFwWV1UCCJgkNzIyEoCyZcvSo1dvli1b6nFEWbf4m0VMnTqZ2jWrMnhQf+bNncOlgy/2Oqx0LVvyDTOnT6XJmWcwetjFLPp6LldeNgSAJx99gN27d3Hfw094HGXO6dd/IJM+/xSAihUjidu6JfW5uLitVKwY6VVoAGzf46vYxv91mKnLthBbozS79h6iXAlf9bZciULE/3UYgG1/HiTq9MKpr40sVZhtf/oqu09+/j1tbp9Kr4e/wszYuP2vPN6TzKlYMZKtxx2Do78LAtHSb7/hy+lTadLgDEZd6vt5uWLEkJMuXyEykrM7nEuRIkU4/fTStGjdhh9/+C4PI86eYDsumVGuXDnCw8MJCwtj2PDLWL48+P7GhOJxkVMTckkucDaQ5Jx7+eiAc24NsNDMnjCzH8zsezPrB2BmRc1stpmt9I/39L/sUaCGma02M0+zmAMHDrBv377U+1/Nmkn9+g0yeFXgeeChR/hl81bWbdzMO+9NoP3Z5/DmO+O9Ditdd455iFVrN7H8+w28/MZ4Wrc7mxdee5v33n6DebNn8dL/xgdV68WJbNy4IfX+tKmTqVW7NgBdunXng/fexTnH0m+XULx4ccpXqOBVmBQuEEHRghGp989pWIG1WxP4YsVWBrarAcDAdjWYtsL3R276yi0MaOsbb1qzNH8dTOKPhEOEmVGqaAEA6lcuQf3KJZj93TYP9ihjTZo2ZePGDWzetInExEQ+mjiBrt16eB3WSd015iFW/7yJ5T9s4JU3fT8vL77+9kmXP69rd75d8g3JyckcPHiQlcuXHnOCZ6AKtuOSGdvT9NtP+vwz6gXh35hQPC7pOXpZX69vgSwULwbRAFhxgvE+QDTQCCgNLDOzr4FdQG/n3F9mVhpYYmaTgduABs656BNtxMxGAiMBKlWunNP7cIydf/xBvwt7A5Cckky//gPp1Pm8XN2mpO+W668kqlIVup3bFoAu3Xtx4613sfOPHXRu35J9+/4iLCyM1156nq+/XcNpaU7u8tKllwxkwYL57I6Pp3aNytxx173M/HI6G9avJywsjEqVK/Pc8y8B0Pm8LsycMZ1G9WpRqHBhXnr1f57GXrZ4Qd67oT0AEeFhfLRoE1+t2cbKX3bz1rXtGNy+Jr/HH2Doc/MB+HJVHJ2iI1nzbG8O/p3MFa98A0C+CGPGvb4e6n2HkrjshYWkBGi7QkREBM88N47uXTuTkpLCkKHDqFe/vtdhZdlrL43jheeeYucfOzi7ZSwdOp3HM+NeoVbtupzTsRNnt4zBwsIYNHgYdesFfnIV7Mdl8MUDWDB/HvHx8dSoGsXd99zH1/Pn8d2a1ZgZVapW5fkXX/E6zCwL9uMiOc8y6l8LNmZ2DVDNOXf9cePPAN87597wP34X+AiYDjwDtAOOALWBakBBYKpzLsPfuLGxTVywT+sVivYeTPI6hBxTpEC41yHkmIpDA7t6nxU73x3sdQg55q9DofPzAlAsQKeFk9BRKJ+tcM55NnF96er1XbeHvJ927+2BjTx9H9ITipXcH4ELs7D8IKAMEOucSzKzzfgSXBEREZGAFeizG3gtuJsJT2wOUMDfTgCAmTUEEoB+ZhZuZmXwVW6XAsWBnf4E92zg6Gzl+4DT8jRyEREREckRJ63kmtnzwEl7GZxz1+RKRKfIOefMrDfwrJndChwGNgPXAUWBNfj26xbn3A4zew+YYmbfA8uBn/3r2W1mi8zsB2C6c+7mvN8bERERkRNTHTd96bUrBG2TqXNuG9D3BE/d7L+lXTYeaHmS9QzM+ehEREREJLedNMl1zh0z74uZFXbOHTzZ8iIiIiIigSLDnlwza2lmP+H/GN/MGpnZi7kemYiIiIickBmEmXl+C2SZOfHsWaAzsBtSL6zQLhdjEhERERE5JZmaXcE5t+W4oZRciEVEREREJEdkZp7cLWbWCnBmlg+4Flibu2GJiIiISHoCvFvAc5mp5I4GrgQigW34Lo17ZS7GJCIiIiJySjKs5Pqn2BqUB7GIiIiIiOSIzMyuUN3MppjZLjPbaWaTzKx6XgQnIiIiIidmZp7fAllm2hXeBz4EKgAVgY+AD3IzKBERERGRU5GZJLewc+5d51yy/zYeKJjbgYmIiIjIyZl5fwtkJ+3JNbNS/rvTzew2YALggH7AF3kQm4iIiIhItqR34tkKfEnt0Tx9VJrnHHB7bgUlIiIiInIqTprkOueq5WUgIiIiIpI5RuBfVhfAzAoCXwMF8OWdHzvn7jWzavi6BE7HV1i9xDmXaGYFgHeAWHxX2+3nnNvsX9ftwHB8FyW7xjn3ZXrbzszFIDCzBkA90vTiOufeycpOioiIiMh/zt/AOc65/f6Lii00s+nADcAzzrkJZvYyvuT1Jf/XPc65mmbWH3gM6Gdm9YD+QH18EyF8ZWa1nHMnvQpvZqYQuxd43n87G3gc6HEKOysiIiIi/wHOZ7//YT7/zQHnAB/7x98Gevnv9/Q/xv98B/PNVdYTmOCc+9s5twnYCDRLb9uZmV3hQqADsMM5dynQCCieideJiIiISG4IgJkV/N0Spc1seZrbyH+FahZuZquBncAs4BcgwTmX7F9kK74r6+L/ugXA//xefC0NqeMneM0JZaZd4ZBz7oiZJZtZMX+AlTLxOhEREREJbfHOuSbpLeBvKYg2sxLAZ0CdvAgsM0nucn9Qr+FrDN4PLM7NoEREREQkfYF+xbHjOecSzGwu0BIoYWYR/mptFBDnXywOXzF1q5lF4Ose2J1m/Ki0rzmhDNsVnHNXOOcSnHMvA+cCQ/xtCyIiIiIiJ2VmZfzFUsysEL5cci0wF19LLMAQYJL//mT/Y/zPz3HOOf94fzMr4J+Z4QxgaXrbTu9iEDHpPeecW5nBfol4qnjhfF6HICew893BXoeQY0o2u9rrEHLMnqXPex1CjvL9TQwNwVatS8+RI6FzXCTTKgBvm1k4vuLqh865qWb2EzDBzB4EVgH/8y//P+BdM9sI/IlvRgWccz+a2YfAT0AycGV6MytA+u0KT6Xz3NGz4kRERETEA5mZPcBrzrnvgMYnGP+VE8yO4Jw7DFx0knU9BDyU2W2ndzGIszO7EhERERGRQBIM/wSIiIiIiGRJpq54JiIiIiKBwwitfu3coEquiIiIiISczFzW18zsYjO7x/+4spmlexk1EREREREvZaZd4UXgCL7ZFO4H9gGfAE1zMS4RERERSUeYuhXSlZkkt7lzLsbMVgE45/aYWf5cjktEREREJNsyk+Qm+SfwdeC7cgW+yq6IiIiIeESV3PRl5sSzscBnQFkzewhYCDycq1GJiIiIiJyCDCu5zrn3zGwF0AHfjBW9nHNrcz0yEREREZFsyjDJNbPKwEFgStox59zvuRmYiIiIiJyYmebJzUhmenKn4evHNaAgUA1YB9TPxbhERERERLItM+0KZ6Z9bGYxwBW5FpGIiIiIyCnK8mV9nXMrzax5bgQjIiIiIpmj2RXSl5me3BvSPAwDYoBtuRaRiIiIiMgpykwl97Q095Px9eh+kjvhiIiIiEhm6Lyz9KWb5PovAnGac+6mPIpHREREROSUnfRiEGYW4ZxLAVrnYTxyAjO/nEHD+rWpX6cmTzz+qNfhnLJQ2p9Q2ZdRI4ZRuWJZYqMbeB1KjgiG4/Lz1DEsm3g7Sz64lYXjbwagYa1I5r99Q+pYk/pVAChxWiEmPjmCpRNvY8E7N1GvRoV01xOoguG4nMzhw4dp26o5zWOjiW3UgAfuuxeAkcMvpW6t6jRv0pjmTRqzZvVqbwPNhto1q9Ik+kyax0bTunkTr8PJsoSEBAb1v4jGZ9YlpmE9vl2yOPW55555iiIFwoiPj/cwQvFKepXcpfj6b1eb2WTgI+DA0Sedc5/mcmwCpKSkcN01VzJt+iwio6Jo06Ip3br1oG69el6Hli2htD+htC+XDBnK6CuuYsSwwV6HcsqC6bicN2osuxNSf63y0LU9eeiVGcz85ic6t67HQ9f2pPPIsdwyvBNr1sfR76bXqVW1HM/edhFdRo876XoCUTAdlxMpUKAA02fOpmjRoiQlJdGhfVs6n3c+AA8/8ji9L7jQ4whPzYyv5lK6dGmvw8iWm2+8jnM7dea9CR+RmJjIwYMHAdi6ZQuzv5pFpcqVPY4wdxgQpn6FdGXmsr4Fgd3AOUA3oLv/q+SBZUuXUqNGTapVr07+/Pm5qF9/pk6Z5HVY2RZK+xNK+9KmbTtKlSrldRg5IpiPiwOKFS0IQPGihdi+ay8AdapVYP6y9QCs3/wHVSqUomyp0062moAUzMcFfJPuFy1aFICkpCSSkpLUEBkA9u7dy6IFXzPk0uEA5M+fnxIlSgBw68038OAjj+mCCf9h6SW5Zf0zK/wAfO//+qP/6w95EJsA27bFERVVKfVxZGQUcXFxHkZ0akJpf0JpX0JJsBwX52DKC1ey6L2bGdanFQA3P/kJD1/bkw1f3M8j1/finnGTAfh+Qxw9z2kEQJP6VahcoRSR5UqcdD2BKFiOS3pSUlJo3qQxVSLL0aFDR5o1882mOeaeu2gW04hbbrqev//+2+Mos87M6H5+J1o1i+V/r73qdThZsnnzJkqXKcOoy4bRslkMV4wewYEDB5g6eRIVKlakYcNGXocoHkqvXSEcKIqvIn48lzvhnJyZ3QkMBFKAI8Ao59y3p7jOecBNzrnlp7KMiEhWdRj2DNt27aVMyaJMfekq1m3+gz4dornlqU/5fM4aLji3MS/dM4iul4/jyTdn8eTNF7Dkg1v5ceM21qzbSkrKkZOuZ9HKXzzeu9AUHh7Ot8tXkZCQQP+L+vDjDz9w34MPU758eRITE7nq8lE89cRj3HHXPV6HmiWz5y0kMjKSnTt30u28c6ldpw5t2rbzOqxMSUlOZvWqlTz1zFiaNmvOTTdcy0MPjGHRwgVMnval1+Hlusx8HP9flt77s905d79z7r4T3O7PswgBM2uJr0UixjnXEOgIbMnLGLxSsWIkW7f+s6txcVuJjIz0MKJTE0r7E0r7EkqC5bhs87ci7Nqzn8lz19C0fhUGdWvO53PWAPDJrFU0qe/rJdx34DCjxrxHiwGPMfzudyldsiib4nafdD2BKFiOS2aUKFGCdme1Z9bMGVSoUAEzo0CBAlwyZCjLly/zOrwsO3ocypYtS49evVm2bKnHEWVexcgoIqOiaOqvqvfucyGrV61i8+ZNtGgaTd1a1YjbupXWLWLZsWOHx9FKXksvyQ2kJpYKQLxz7m8A51y8c26bmd1jZsvM7Acze9X8jTdmNs/MHjOzpWa23sza+scLmdkEM1trZp8BhY5uwMxeMrPlZvajmd3nxU6eSJOmTdm4cQObN20iMTGRjyZOoGu3Hl6HlW2htD+htC+hJBiOS+GC+SlauEDq/Y4t6vDjL9vZHr+XtrE1AWjfrBYbt+wCfP25+SLCAbi0dysWrvyFfQcOn3Q9gSgYjkt6du3aRUJCAgCHDh1izuyvqFW7Dtu3+95v5xxTJn9O/Xr1PYwy6w4cOMC+fftS7381ayb16wfPLCvly5cnKqoS69etA2De3NlEN27Mb1v/YO36Taxdv4nIqCgWLVlB+fLlPY4255l5fwtk6bUrdMizKDI2E7jHzNYDXwETnXPzgXFHq8pm9i6+au8U/2sinHPNzKwLcC++6u/lwEHnXF0zawisTLONO51zf/rnBp5tZg2dc9+dLCAzGwmMBHL1zM2IiAieeW4c3bt2JiUlhSFDh1GvfnD9Ek0rlPYnlPZl8MUDWDB/HvHx8dSoGsXd99zH0GHDvQ4rW4LhuJQ9/TQmPnUZABHhYUycsZxZ36zlyoMf8MTNFxARHs7ffydx1YMTAKhTvRyv3XcJzjnW/rqD0fe9l+56AlEwHJf07Ni+ncuGD+VISgpHjhyhz4UX0aVrN87v1IH4XbtwztGwUTRjX3jJ61CzZOcff9Dvwt4AJKck06//QDp1Ps/jqLLmyWfGMmzoxSQmJlKtWnVefu0Nr0OSAGHO5Xl7bbb4k8+2wNnAKOA2YB9wC1AYKAU875x71N9He6dzbpGZlQMWOedqmtnnwFjn3Bz/OlcCI51zy81sNL6kNQJf5fhq59yEzPTkxsY2cYu+VcuuyH9NyWZXex1Cjtmz9HmvQ8hRwfK3LTNCaXaAI0dC57gUKRC2wjnn2cTCFc5o4IaN9X4214e71Pb0fUhPZi7rGxD8F6aYB8wzs+/xJboNgSbOuS1mNgbfdGdHHT3FNYWMr+xWDbgJaOqc22Nmbx23LhEREZGAYWaaJzcDQXFinpnVNrMz0gxFA+v89+PNrCiQmZm4v8Y3QwNm1gBfkgxQDN+FLvb6K7/n50TcIiIiIuKNYKnkFgWeN7MSQDKwEV9rQQK+OXt3AJk5pfUl4E0zWwusBVYAOOfWmNkq4Gd8szYsyuH4RURERCQPBUWS65xbAZxolvO7/Lfjl2+f5n48UNV//xDQ/yTbGHqS8fYnGhcRERHxkroV0hcU7QoiIiIiIlmhJFdEREREQk5QtCuIiIiIyLHC1K6QLlVyRURERCTkqJIrIiIiEmQMNE9uBlTJFREREZGQoyRXREREREKO2hVEREREgpC6FdKnSq6IiIiIhBwluSIiIiISctSuICIiIhJsTPPkZkSVXBEREREJOarkioiIiAQhQ6Xc9KiSKyIiIiIhR0muiIiIiIQctSuIiIiIBBnfZX29jiKwqZIrIiIiIiFHSa6IiIiIhBy1K4iIiIgEIbUrpE+VXBEREREJOark5oCkFMcfew97HUaOKFe8oNch5Ji/DiV5HUKOKZgv3OsQ5AT2LH3e6xByTK3rJ3sdQo5a/0wPr0OQEwhT6TFHmen9TI8quSIiIiIScpTkioiIiEjIUbuCiIiISJDRPLkZUyVXREREREKOklwRERERCTlqVxAREREJNgaaXCF9quSKiIiISMhRkisiIiIiIUftCiIiIiJBKEz9CulSJVdEREREQo4quSIiIiJBRvPkZkyVXBEREREJOUpyRURERCTkqF1BREREJAjpvLP0qZIrIiIiIiFHSa6IiIiIhBy1K4iIiIgEHSMM9SukR5XcALJ3bwKXXzqAc1o0okPLaFYsW8KVwy/m/PbNOb99c1o3rs357ZsDsOX336gdVTL1uTtuvNrj6DNn/bp1NI+NTr2VLVWM55971uuwMpSSkkKHNk0ZdFEvAH7bvInzzm5N80Z1uWzoQBITEwH4+++/uWzoQJo3qst5Z7fm9982exf0CVw5ajg1KpenRWzD1LGhF/enTfMY2jSP4cza1WnTPAaAP3fvplvnDlQsXYybrgu8768T7ct3a1bToV0r2jSP4azWzVixbCkAC76eR6VyJVP387GHH/Aq7Cyb+eUMGtavTf06NXni8Ue9Duekwgy+uOUs3hzVDICnLo5m4ZgOTL/1LKbfehb1Iosds3zDyiX49dludImukDpWsWQhxl/Rgtl3ns3sO84mqlShPN2HrAiW43Iio0YMo3LFssRGNzhm/MVxz9OoQR1iGtXnjttu8Si6UxPMx0Vyniq5AeS+O27irHM68dKbH5CYmMihQwd54X/jU59/8O5bOa1Y8dTHVapWZ/q8b70INdtq1a7NtytWA77EsUaVSHr06u1tUJnw2kvPc0atOuzbtw+AB++9g1FXXkPvC/tx83VX8v47bzJ0xCjef+dNSpQoybdr1vLZxxN54N47eO2t9z2O/h8DLxnCZaOvZPSIoaljb42fkHr/zltvolhx3/dYgYIFufOe+/jppx9Y++OPeR1qhk60L/fceSu33Xk353Y+n5kzvuCeO29j2sw5ALRs3YYPP53iUbTZk5KSwnXXXMm06bOIjIqiTYumdOvWg7r16nkd2r8Ma1+djX/s47SC//xZefjzn/hi9fZ/LRtmcHvPunz9865jxp+5pDHjvtzAgnW7KJw/nCMu18POlmA6LidyyZChjL7iKkYMG5w6Nn/eXKZOmcTSFWsoUKAAO3fu9DDC7An245JVhk48y4gquQHir7/2snTxQvpdPBSA/PnzU7x4idTnnXNMm/QJPfr09SbAXDB3zmyqVa9BlSpVvA4lXdvitjLry+kMGjIM8B2LhfPn0b3XBQD0HXAJ06dOBmDGtCn0HXAJAN17XcDCeXNxLnD+Urdu046SpUqd8DnnHJ998hEX9u0PQJEiRWjZug0FCxbMyxAz7UT7Ymb89ddfAPy1dy/lK1Q40UuDxrKlS6lRoybVqlcnf/78XNSvP1OnTPI6rH8pX6IgHeqXY8Li3zO1/KVnVWf66u3s3v936tgZ5YsSEWYsWOdLfA8mpnA4KSVX4j1VwXJcTqZN23aUOu5n59VXXuKmW26jQIECAJQtW9aL0E5JsB8XyXlKcgPElt82c/rppbnp6pF0ObsFt157OQcPHEh9funiRZQuU45qNWr+85rfN9Pl7Bb07X4uSxcv9CLsU/LRxAn07TfA6zAydPdtN3LP/Y8QFub7cfnzz90UK16CiAhfxapiZCTbt8cBsH17HJFRUQBERERwWrHi/Pnnbm8Cz6JvFi2gTLly1Kh5htehZNujTzzDPXfcSr2aVbjr9lu49/6HU59b+u0SWjdrzAU9u7D2p8CrTJ/Itm1xREVVSn0cGRlFXFychxGd2Jg+DXh40k8cOa70enO3unx5W3vu6VOf/BG+n59yxQvSuWF53l24+Zhlq5Utyl+HknhlRFO+uOUs7uhZL2Cv5hQsxyUrNq5fz6KFC2jbqjnnnnMWy5ct8zqkLAvF4yKnJiSTXDO708x+NLPvzGy1mTU3s81mVvoEy/Yws9tOsp72ZtYq9yOGlORkfvhuNRdfehlfzF1CoSKFeWnsk6nPT/70Q3r0uSj1cdly5flm9Xq+mLuEux94jGtHDWXfvr/yItQckZiYyLSpk+lz4UUZL+yhmdOnUbp0WRo1jvE6lFz38YcTuPCi/l6HcUr+9+rLPPz4U/y08Tcefvwprrr8MgAaRcfww7pNLFq6ilGXX8XAvn08jjR0dKhfjvj9f/P9lr3HjD82eS1nPziH7k9+TYnC+bm8o+8f9DEXNOCRyWs5/gOOiDCjaY3TeeizH+n+5NdULl2Yi5pXzqvd+M9LTknmzz//5OtFS3j40Se4eGDfgPoUSk7AfK0/Xt8CWcj15JpZS6AbEOOc+9uf2OY/2fLOucnA5BOsJwJoD+wHvsmdaP9RvmIk5StG0jjWd9JGl+69eem5pwBITk7my2mTmDJ7UeryBQoUSP1Y6czoGCpXrc6mjRto2Dg2t0PNEV/OmE504xjKlSvndSjpWvrtN3w5fSqzZ83g8OHD7N/3F3fdcgN/7U0gOTmZiIgItsXFUaFCJAAVKkQSt3UrFSOjSE5OZt9feylV6nSP9yJjycnJTJn0GfMXBV/1Jq0P3nuHx556FoDeF1zENVeMBKBYsX9Oeup0XhduvPYqdsfHc3rpf/3fG1AqVoxk69YtqY/j4rYSGRnpYUT/1qR6Kc5tUJ6z65WjQL4wTisYwbODY7junZUAJCYf4cMlvzOqQw0AzqxcnHFDfb+nShXNz9n1ypF8xLE94TA/bd3L77sPAjDzux00rlqSiUu82a/0BMNxyarIyCh69e6DmdG0WTPCwsKIj4+nTJkyXoeWaaF4XOTUhGIltwIQ75z7G8A5F++c2+Z/7mozW2lm35tZHQAzG2pm4/z33zKzl83sW+BDYDRwvb8a3DY3gy5brjwVI6P4ZcN6ABZ9PY8zatcBYOH8OVSvWYsKFaNSl98dv4uUFF+/2u+bN7H5141UrlotN0PMUR9O/CAoWhXuGvMQq3/exPIfNvDKm+Np3e5sXvrfO7RudxZTPv8EgA8/eJfzunYHoHOXbnz4wbsATPn8E9qc1R4LgjMD5s35ilq16qS2WgSr8hUqsnDBfADmz5tDdX/rxR87dqRWpVYsW8qRI0codXrg//PRpGlTNm7cwOZNm0hMTOSjiRPo2q2H12Ed47Epa2l+zyxaj/mKq95cwTfr47nunZWULVYgdZnODcuzbrvvpM02Y2bTesxXtB7zFV+s3sZdH37HzO92sOa3PRQrnI9SRX01iVa1SrNhxz5P9ikjwXBcsqp7j17MnzcXgA3r15OYmEjpAP8n8HiheFzk1IRcJReYCdxjZuuBr4CJzrn5/ufinXMxZnYFcBMw4gSvjwJaOedSzGwMsN859+TxC5nZSGAkQGSaHqBTMeaRp7lu9KUkJSVSqUpVnnz+VQCmfPbRv044W7p4IU8/+gAR+fIRZmE89OTzlCh54hOKAs2BAweY89Usxr34itehZNtd9z3MqEsv5tEHxnBmo0YMHHwpAAMHX8pVI4fSvFFdSpQsyStvjs9gTXlr2OCBLFwwn93x8dStUZnb776XwUOH88lHE7mgb79/LX9m7er8te8vkhITmTZlEp9NnUGduoFxpvKJ9mXsC69w683Xk5KcTIECBXlu3MsATPrsE/732stERERQsGAh3njn/aD45yMiIoJnnhtH966dSUlJYcjQYdSrX9/rsDLluSExnF60AAb8GPcXd0xYk+7yRxw89NlPfHBVK8zg+y0JfPDNb3kTbBYF83EBGHzxABbMn0d8fDw1qkZx9z33MeTSYYwaMYzY6Abkz5ef1994Oyh+RtIK9uOSHWFBdozymoViz42ZhQNtgbOBUcBtwBigtXMuzsyaAw855zqa2VCgiXPuKjN7C5jrnHvbv54xnCTJTathdKxL20oQzMoVD8wz6bPjr0NJXoeQYwrmC/c6BDmBoydThYJa1/+rayuorX9GFTzJXYXy2QrnXBOvtl+lbkN355veT4s4qmVVT9+H9IRiJRfnXAowD5hnZt8DQ/xPHZ2vJoWT7/uBk4yLiIiISJAIuSTXzGoDR5xzG/xD0cBvwJnZWN0+oFiGS4mIiIjkIV0MImOh81nbP4oCb5vZT2b2HVAPX6tCdkwBeufFiWciIiIiknNCrpLrnFsBnGhu26ppllmOb3ownHNvAW/57w89bl3rgYa5EaeIiIjIqdCJZ+kLxUquiIiIiPzHKckVERERkZATcu0KIiIiIv8F6lZInyq5IiIiIhJylOSKiIiISMhRu4KIiIhIkDFUqcyI3h8RERERCTmq5IqIiIgEGwPTmWfpUiVXREREREKOklwRERERCTlqVxAREREJQmpWSJ8quSIiIiIScpTkioiIiEjIUbuCiIiISJAxIEyzK6RLlVwRERERCTmq5IqIiIgEIdVx06dKroiIiIiEHCW5IiIiIhJy1K4gIiIiEoR03ln6VMkVERERkZCjSm4OyBdulCte0Osw5DjFCuXzOgQJcSlHnNch5Jj1z/TwOoQcVfLse7wOIcfsmXu/1yHkmEOJKV6HIP8hSnJFREREgo5h6ldIl9oVRERERCTkKMkVERERkZCjdgURERGRIGOoUpkRvT8iIiIiEnJUyRUREREJQjrxLH2q5IqIiIhIrjCzSmY218x+MrMfzexa/3gpM5tlZhv8X0v6x83MxprZRjP7zsxi0qxriH/5DWY2JKNtK8kVERERkdySDNzonKsHtACuNLN6wG3AbOfcGcBs/2OA84Ez/LeRwEvgS4qBe4HmQDPg3qOJ8ckoyRUREREJQhYAt4w457Y751b67+8D1gKRQE/gbf9ibwO9/Pd7Au84nyVACTOrAHQGZjnn/nTO7QFmAeelt2315IqIiIhIdpU2s+VpHr/qnHv1RAuaWVWgMfAtUM45t93/1A6gnP9+JLAlzcu2+sdONn5SSnJFREREJLvinXNNMlrIzIoCnwDXOef+SnvSnHPOmVmOXydd7QoiIiIiwcZ8syt4fctUqGb58CW47znnPvUP/+FvQ8D/dad/PA6olOblUf6xk42flJJcEREREckV5suE/wesdc49neapycDRGRKGAJPSjA/2z7LQAtjrb2v4EuhkZiX9J5x18o+dlNoVRERERIJMEF3xrDVwCfC9ma32j90BPAp8aGbDgd+Avv7nvgC6ABuBg8ClAM65P83sAWCZf7n7nXN/prdhJbkiIiIikiuccws5+UQMHU6wvAOuPMm63gDeyOy2g+SfABERERGRzFMlV0RERCQI6bK+6VMlNwjM/HIGDevXpn6dmjzx+KNeh3PKQml/tC+BKZj35YXnn6Np4zNpEt2AF8Y+C8Cnn3xEk+gGnFYwnJUrlqe/ggAWDMfl5w+vZ9lbV7LkjctZ+Nqo1PHLL2jO6vFXs+Kdq3jo8k4AnNOkBoteH82yt65k0eujOSumWuryfTucybK3rmTpW1cw6clLOL144Tzfl8yqXbMqTaLPpHlsNK2bZzgTlOe2bt1Cj/M70CL2TFo2acjLL4wFYNjgAbRrEUu7FrE0qluDdi1iU1/z4/ff0ens1rRs0pDWTaM5fPiwV+FLHlIlN8ClpKRw3TVXMm36LCKjomjToinduvWgbr16XoeWLaG0P9qXwBTM+/Ljjz/w1huvM3/Rt+TPn59e3c7nvC7dqFevAe9P/IRrrhrtdYjZFkzH5bxr32T33oOpj9s1rka3NnVodumLJCalUKZEEQB27z3Ahbe+x/bd+6hXrSxTnhpMjT5PEh4exhPXnk/MJePYvfcgD13eidF9mvPQm3O92qUMzfhqLqVLl/Y6jEyJCI/ggYefoFHjGPbt28c5bZrR/pyOvPHOB6nL3HXbTRQrXhyA5ORkRg0fwsuvv0WDho34c/du8uXL51X4kodUyQ1wy5YupUaNmlSrXp38+fNzUb/+TJ0yKeMXBqhQ2h/tS2AK5n1Z9/NamjZrRuHChYmIiKBNu3ZM/vxT6tStS63atb0O75QE83EZ2aspT45fQGJSCgC7Eg4AsGbDDrbv3gfAT5t2UrBABPnzhfsud2pGkYK+ROq0wgXYHv+XJ7GHovIVKtCocQwAp512GrVq12H7tn+mS3XO8fmnH3PBRf0BmPvVTOo3OJMGDRsBUOr00wkPD8/7wHOB15f0DfRmCSW5AW7btjiiov6Z+zgyMoq4uHTnPg5oobQ/2pfAFMz7Uq9eA75ZuJDdu3dz8OBBZs6YztatWzJ+YRAIluPiHEx5ejCLXh/NsO6+j7trVjqd1o2q8PUrI5n5/DBi61T81+t6t6/H6vXbSUxKITnlCNc+NYVlb1/Jr5/fTN2qZXhr2sq83pVMMzO6n9+JVs1i+d9rJ7waa8D6/bfNfLdmNbFNm6eOLV60gLJly1Gj5hkAbNy4ATPjgh7n075VU8Y+/YRX4UoeC7h2BTO7ExgIpABHgFHOuW9zaN3tgZucc91yYn0iIjmpTt26XH/TLfTs2pnCRYpwZsNGIVNxChYdrnydbfH7KFOiCFOfGcK63+OJCA+jVLFCtBv1Kk3qRjL+vn7U7fdM6mvqVi3Dg6M70e2GtwGICA/jsl7NaDHsJTZt28Mz13Xl5ovb8dg7873arXTNnreQyMhIdu7cSbfzzqV2nTq0advO67AytH//foYM7MvDjz9NsWLFUsc/+WgifS7ql/o4OTmZJYsXMfvrJRQqXJheXc+lUeMYzjr7X7NXSYgJqEqumbUEugExzrmGQEcgIMoYZubJPwQVK0YeU8mJi9tKZGSkF6HkiFDaH+1LYAr2fRly6XAWLlnOzNnzKVmyJDXPqOV1SDkiWI7Ltnhf+8GuhANM/notTetGEbfrLz6fvxaA5WvjOOIcpUv4TiSLLFOMiQ8PYMRDn7Jp2x4AGp1RHiD18cdzf6BFg0rHbypgHD0OZcuWpUev3ixbttTjiDKWlJTEkIEXcWG/AXTv2Tt1PDk5mamTPqP3hX1TxypGRtGqdVtOL12awoULc27n81mzepUXYec4M+9vgSygklygAhDvnPsbwDkX75zbZmabzew+M1tpZt+bWR0AMytiZm+Y2VIzW2VmPf3jVc1sgX/5lWbW6vgNmVlT/2tqmFmsmc03sxVm9mWaaynPM7NnzWw5cG3evQ3/aNK0KRs3bmDzpk0kJiby0cQJdO3Ww4tQckQo7Y/2JTAF+77s3Om7fPuW339n0uef0bf/QI8jyhnBcFwKF8xH0UL5U+93bFqDH3/9gykL1qbOnFCz0unkjwgnPuEgxYsW5NPHL+bul2ex+PvfU9ezbdc+6lQtk5oId2hSg3W/7cr7HcqEAwcOsG/fvtT7X82aSf36DTyOKn3OOa65/DJq1a7Llddcf8xz8+Z8xRm1axMZGZU61qFjJ3768QcOHjxIcnIy3yz4mjp16+Z12OKBQGtXmAncY2brga+Aic65o5/vxDvnYszsCuAmYARwJzDHOTfMzEoAS83sK2AncK5z7rCZnQF8AKTOi+JPep8HegLbgXeBns65XWbWD3gIGOZfPL9z7l9zqpjZSGAkQKXKlXP0TUgrIiKCZ54bR/eunUlJSWHI0GHUq18/17aX20Jpf7QvgSnY92VQ/wtTz/5++rlxlChRgsmTPuOm668hftcuLujVjYYNo5k0bYbXoWZJMByXsiWLMvHhAYCv5WDirO+YtXQj+SLCeeX2Xix/+0oSk1MY8fCnAIzu05wakaW4fWh7bh/aHoDuN7zD9t37ePjNucx6fjhJKSn8vmMvI/2vCTQ7//iDfhf6KqHJKcn06z+QTp3P8ziq9H27eBETPxhPvfpnpk4TdveYBzj3vC589vGHqSecHVWiZEmuuPo6OrRrgWGc2/k8Op3X1YvQc5Tvsr4BXkr1mPmunhY4zCwcaAucDYwCbgPGAK2dc3Fm1hx4yDnX0V9hLQgk+19eCugMbAPGAdH4entrOecK+3ty/wccAjr5q8QNgG+AX/3rCAe2O+c6mdk84N40ifYJxcY2cYu+Dd65K0Uke1KOBNbvz1MRHhZafyxLnn2P1yHkmD1z7/c6hBxzKDHF6xByTKkiEStOVATLK2fUb+SenjDTq82n6tGwvKfvQ3oCrZKLcy4FmAfMM7PvgSH+p/72f03hn7gNuMA5ty7tOsxsDPAH0AhfS0baWZ+340uMG+NLhg340TnX8iQhHTiF3RERERERDwRUT66Z1fa3FxwVDfyWzku+BK42/3XtzKyxf7w4vmrsEeASfNXZoxKArsAj/sruOqCM/6Q3zCyfmQXWZ2giIiIix/H6pDOdeJY1RYG3zewnM/sOqIevVeFkHgDyAd+Z2Y/+xwAvAkPMbA1Qh+Oqsc65P/DN4vACvoruhcBj/uVXA/86UU1EREREgkdAtSs451Zw4gSzappllgPt/fcP4evbPX49G4CGaYZu9Y/Pw9cKgXPudyBtxfZfkwI659pnJX4RERERCQwBleSKiIiISGYYptkV0hVo7QoiIiIiIqdMlVwRERGRIBToJ355TZVcEREREQk5SnJFREREJOSoXUFEREQkyOiyvhlTJVdEREREQo6SXBEREREJOWpXEBEREQk2QXBZXa+pkisiIiIiIUeVXBEREZEgpEpu+lTJFREREZGQoyRXREREREKO2hVEREREgpBpntx0qZIrIiIiIiFHSa6IiIiIhBy1K4iIiIgEGQPC1K2QLlVyRURERCTkKMkVERERkZCjdgURERGRIKTZFdKnJFdEJJvCQ6gh7sgR53UIOWrP3Pu9DiHHlGx+rdch5Jg93z7ndQjyH6IkV0RERCQI6bK+6VNProiIiIiEHCW5IiIiIhJy1K4gIiIiEoR04ln6VMkVERERkZCjJFdEREREQo7aFURERESCjC7rmzFVckVEREQk5KiSKyIiIhJ0TCeeZUCVXBEREREJOUpyRURERCTkqF1BREREJNiYLuubEVVyRURERCTkKMkVERERkZCjdgURERGRIKRuhfSpkisiIiIiIUeVXBEREZEg47vimWq56VElNwBt2bKFzh3PpnHDesQ0qs+4sc8B8MnHHxHTqD6F84exYvlyj6PMvrHPPkNMo/rERjdg8MUDOHz4sNchZUqoHZdRI4ZRuWJZYqMbpI79+eefdD3vXBrUPYOu553Lnj17PIwwe9avW0fz2OjUW9lSxXj+uWe9DitbDh8+TJuWzWgW04iYRvV54L57vQ4py+rWqkbTmIa0aNqYNi2bpo6/9MLzND6zLk2iG3Dn7bd4GGH21K5ZlSbRZ9I8NprWzZt4Hc5J/TzlHpZNvJUl79/MwndvBKBhrUjmv3V96liT+pVTl28bW5Ml79/Mig9vY+arV6eOXz2wPSs+vI3lE2/j7YcGUyB/4NXITvQ7Tf7bAu+7VIiIiODRx5+icUwM+/bto1XzWDp0PJf69Rsw4cNPueqKUV6HmG1xcXG8+MJYVn33E4UKFWLQgL58NHEClwwZ6nVoGQq143LJkKGMvuIqRgwbnDr25OOP0v6cDtx8y2088fijPPn4ozz0yGMeRpl1tWrX5tsVqwFISUmhRpVIevTq7W1Q2VSgQAFmzJpD0aJFSUpK4pyz2tCp8/k0b9HC69CyZPrMOZQuXTr18fx5c5k6ZTJLlq+mQIEC7Ny508Posm/GV3OP2a9Add6ocexOOJD6+KFre/DQqzOY+c1aOreux0PX9KDzqHEUL1qI5267iJ5Xv8yWHXsoU7IoABXLFOeK/u1ofNEjHP47ifGPDuWizjGMn7LUq106oRP9TpP/NlVyA1CFChVoHBMDwGmnnUadOnXZti2OOnXrUqt2bY+jO3XJyckcOnTI9/XgQSpUrOh1SJkSaselTdt2lCpV6pixqVMmcfElQwC4+JIhTJn8uQeR5Zy5c2ZTrXoNqlSp4nUo2WJmFC3qSzSSkpJITkrCQuDjyddffZkbb76VAgUKAFC2bFmPI/pvcc5RrEhBAIoXLcj2+L8A6Hd+LJPmrGHLDt8nOLv27E99TUR4GIUK5CM8PIxCBfOzfdfevA88Ayf6nRbqLABugUxJboD7bfNmVq9eRdNmzb0OJUdERkZy3fU3Uat6ZapVqkCxYsXpeG4nr8PKslA7Lkft/OMPKlSoAED58uXZ+ccfHkd0aj6aOIG+/QZ4HcYpSUlJoXlsNJUrluWcjufSrHlwfc8ZRo+unWndoglvvP4qABs2rOebRQs4q00LOndsz4rlyzyOMuvMjO7nd6JVs1j+99qrXodzUs7BlBcuZ9H4mxjWuyUANz/5GQ9f15MN08bwyHU9uef5KQCcUbkMJYoV5stXrmLR+JsY2NXXXrJt116eHT+X9dPGsOnLB/hr/yFmL1nn2T6JZFbItyuYWQrwPb5/OFKAq5xz33gbVebs37+fAX0v4ImnnqVYsWJeh5Mj9uzZw9Qpk1i7YRMlSpRgYP+L+OC98QwYdLHXoWVaKB6XEzGzoK4aJiYmMm3qZO5/6BGvQzkl4eHhfLtiNQkJCfS7sDc//vAD9RsET8/hV3MXUDEykp07d9K9Sydq1a5DcnIye/78k3kLFrNi+TIuGdiPH9f9ElTfb7PnLSTSv1/dzjuX2nXq0KZtO6/D+pcOw59j2669lClZlKkvXsG6zTvp07ERtzz1GZ/PWcMF50bz0j0D6HrFi0SEhxFTtxLnj36BQgXzMe/N61n6/Wbi9+yn21kNqNv9PhL2H+L9xy6l//lNmDA9eM5BkP+m/0Il95BzLto51wi4HQiKv3hJSUkM6HsB/QYMolfvPl6Hk2PmzP6KqlWrUaZMGfLly0evXn1Ysjgo/ucAQve4HFW2XDm2b98OwPbt2ykTxB8jfzljOtGNYyhXrpzXoeSIEiVKcFb7s5k5c4bXoWRJxchIwNeS0KNnL5YvW0pkZBQ9evXBzGjStBlhYWHEx8d7HGnWRKbdr169WbYssPpTj9rmbyvYtWc/k+d+R9MGlRnUrRmfz1kDwCezVtOkvq+dJ27nXmYt/pmDhxPZnXCAhSt/oWGtSM5pXpvNcX8Sn3CA5OQjfD7nO1o0qubZPkkaXvcqBPj/pf+FJDetYsAeADMramazzWylmX1vZj2PLmRmd5vZOjNbaGYfmNlNeRmkc47Rlw2ndp26XHv9DXm56VxXqVJlli5dwsGDB3HOMXfObGrXqet1WJkSysflqK7dejD+3bcBGP/u23Tr3jODVwSuDyd+EPStCrt27SIhIQGAQ4cOMfurWdSuXcfboLLgwIED7Nu3L/X+7K9mUa9+A7r36MnX8+cCsGH9ehKTEoPiBK6jjt+vr2bNpH79wKuuFy6Yn6KFC6Te79iiDj9u3M72XXtpG1sTgPZNa7Fxyy4Apsz7nlbR1f19t/lo2qAKP2/6gy079tDszCoUKpgPgLOb1WLdph3e7JRIFoR8uwJQyMxWAwWBCsA5/vHDQG/n3F9mVhpYYmaTgSbABUAjIB+wElhx/ErNbCQwEqBS5crHP31Kvlm0iPffe5cGDXzT0wDc9+DD/P3339xw3dXE79pFn55dadgomilffJmj285tzZo3p3efC2nZLIaIiAgaNWrM8MtGeh1WpoTacRl88QAWzJ9HfHw8NapGcfc993HTLbdx8YC+vP3m/6hcuQrjP/jQ6zCz5cCBA8z5ahbjXnzF61BOyY7t27ls2BBSUlI44o5wwYV96dK1m9dhZdrOP/6gf1/fJx4pycn07f//9u47Ts6q+uP455tCR4oEhEiRDqEHCEjvLfSSUJQOClFAWn4oVaUJUhQUVKRL70ioIgIJhCAgVZAgEIogAQwJJCTn98e5C8Oy2RSy+8xMvm9eebH77OzOmZ3ZZ85z77nn7sqmm23O2LFj+d4B+7LqysszwwwzcOHvL26oUoX/vP02/XbKjh2fjv+Ufv13Y9PNNq84qi+b9+uzc/UZ+wK5cOzqQcO4e/DzHPyzq/nFETvQrWsXPhk7jgE/uwqAF155m7sffo6hVx3NhAnBxTcN5tl/5czOjfc+yeArjuTTTyfw5Auv84cb6m8Grq1z2l777Ft1WFYhRUTVMXQoSaMiYrby8ZrA74HlyAT/LGBdYAKwFPAtoD8wV0QcX77nl8AbEXHGxO6jd+9V46FHXJtkZo1rwoTmei/o0qVxkuZJmavPIVWHMM2MfOScqkOYZmburmERUVmT5GWWXzkuvun+qu7+M2ssPmelv4f2TA8juZ+JiMFl1LYHsGX5f++IGCfpFXK018zMzMwa3HRVkytpaaAr8F9gDuA/JcHdAGhppPkQsLWkmSTNBjTO3KCZmZlNN6Tq/9Wz6WEkt6UmF3Id4J4RMV7SFcCtkv4BPAY8DxARQ0tt7lPA22T7sfrrem1mZmZmE9X0SW5EdJ3I8XeBNSfybWdExAmSZgEeoI2FZ2ZmZmZWv5o+yZ1KF0palqzRvSQiHq86IDMzM7NadV4tUDknuW2IiN2qjsHMzMzMpt50tfDMzMzMzKYPHsk1MzMza0SuV2iXR3LNzMzMrOl4JNfMzMyswQiQh3Lb5ZFcMzMzM2s6TnLNzMzMrOm4XMHMzMys0TTAtrpV80iumZmZmTUdJ7lmZmZm1nRcrmBmZmbWgFyt0D6P5JqZmZlZ03GSa2ZmZmZNx+UKZmZmZo3I9Qrt8kiumZmZmTUdj+SamZmZNRx5W99J8EiumZmZmTUdJ7lmZmZm1nRcrmBmZmbWgLytb/uc5JqZGV26NNe75fgJUXUI08zIR86pOoRpZq7VBlQdgk1HXK5gZmZmZk3HI7lmZmZmDUa4Te6keCTXzMzMzJqOR3LNzMzMGpGHctvlkVwzMzMzazpOcs3MzMys6bhcwczMzKwBeVvf9nkk18zMzMyajpNcMzMzM2s6LlcwMzMza0De1rd9Hsk1MzMzs6bjkVwzMzOzBuSB3PZ5JNfMzMzMmo6TXDMzMzNrOi5XMDMzM2s0wvUKk+CRXDMzMzNrOk5yzczMzKzpOMltAHfdOYgVei1Fr6UX5xenn1p1OF9ZszyeA/fbh4UWmJfeKy1XdShf2ccff8zaa67O6qusyCor9uKnJx5fdUhfSbO8xgDOPfssVlmxF71XWo7v7rErH3/8cdUhTbVGfl7++cILrLnayp/9m3+eOTjv3LP58cAjWXn5ZejTe0X677wD77//ftWhTpFGOo89f/uJDL3mGIZcNZAHrzgKgOWX7Mn9lxzO0GuO4bqzD2T2WWcCYO45ZmXQhT/knYfO5Kyjd/7Cz7nzd4fw5I3HMuSqgQy5aiA95pqt0x/LtKI6+K+eOcmtc+PHj+fQHx7Mzbfewd+fepZrr/oTzz37bNVhTbVmejzf2XMvbr5tUNVhTBMzzjgjg+6+j0cff5JHHnuCu+4cxCNDhlQd1lRpptfYiBEjOP+8c3loyGMMe+Jpxo8fz7VXX1V1WFOl0Z+XJZdaisFD/87goX/nwSGPMfMss7D1ttuz4UabMPTv/+CRYU+yxBJLcObpp1Qd6hRptPPY5gecwxr9T2Xt3U8H4DfH7cZPzr2Z1XY5mVv+8iSH7bkRAB9/Mo6Tzr+N/zvrxjZ/zt4/voQ1+p/KGv1P5Z2RozotfutcTnLr3NBHH2WxxRbnW4suygwzzMDO/fpz2603Vx3WVGumx7P2Ousy99xzVx3GNCGJ2WbL0Yxx48bx6bhxqEG30mmm1xjAp59+ypgxY/L/o0cz/wILVB3SVGmm5+X+++5l0UUXY6GFF2ajTTalW7dcw71anzUYMWJExdFNmUY/jy2+0Lw8OOwlAO4b8jzbbbQSAKM/HsvDT7zMx5+MqzA6q5qT3Dr3xhsj+OY3F/zs8549v9lwJ9FazfZ4msn48ePp03slFlpgXjbceBNW79On6pCmSjO9xnr27Mmhhx3BkosuxLcWnJ+vfW0ONt5k06rDmirN9Lxcd+1V7LRL/y8dv+ziP7LpZptXENH0ISK49fwBPHTFUeyzw1oAPPfym2y9/goA7LDJKnxzvrkm62ddcMIeDLlqIAP3b9znS+S2vlX/q2cdmuRK2k5SSFp6Mm//iqR52jg+RXMJU3r7dn7OXpIac9jEbAp17dqVR4Y9wUuvvM5jQx/lmaefrjqk6d7IkSO57dabee7F4bz86ht8NPoj/nTF5VWHNV0bO3Yst992K9vv+MU6z9NP/Tldu3Wj3667VxRZ89to77P49m6nsd2A8zmw3zqstcpiHHjCFRywyzo8dMVRzDbLjIwdN36SP2fvYy5mtV1OZuN9zmKtlRdjt76rd0L00zdJF0n6j6Sna47NLeluSS+W/89VjkvSuZJekvSUpFVqvmfPcvsXJe05qfvt6JHcXYEHy/8b0V5ApUnuAgv05PXXX/vs8xEjXqdnz54VRvTVNNvjaUZzzjkn662/AXfd1Th1erWa6TV23733sMgi36JHjx50796d7bbbgSGDH646rKnSLM/LXYPuYKWVVmG++eb77Njll17MoD/fzkWXXN6wZT6N4I13PgDgnZGjuOW+p1it1yL885W32fqg81hr99O5ZtAwhr/+zmT/nFGjP+HqOx5jtV4Ld2jcHUl18G8yXQy0HjYfCNwbEUsA95bPAbYAlij/DgB+A5kUA8cDfYDVgeNbEuOJ6bAkV9JswNrAvkD/muPrS7pf0nWSnpd0hVqdFSTNLOkOSfu38XOPlDS0ZPcntnP/Z0l6RtK9knqUYytJGlK+98aaq4YvHZe0E7AqcIWkJyTNPE1+MVNo1dVW46WXXuSV4cMZO3Ys1159FVv13aaKUKaJZns8zeKdd975bFX4mDFjuPeeu1lqqcmagKk7zfQaW3DBhXj00SGMHj2aiOAv993LUksvU3VYU6VZnpdrr7mKnft9Xqpw952DOOvMX3D19TczyyyzVBhZc5tlphmYbZYZP/t44zWX5pl/vfFZZwRJDNx/M3533YPt/pyuXbvw9TlnBaBbty5sue5yPPOvNzs2eCMiHgDea3V4W+CS8vElwHY1xy+NNASYU9L8wGbA3RHxXkSMBO7my4nzF3TkjmfbAoMi4p+S/iupd0QMK19bGegFvAE8BKxFjvgCzAZcRT7AS2t/oKRNycx+dfIC4hZJ65ZfXq1Zgcci4jBJx5GZ/wDgUuAHEfFXSSeV44e2dTwiDpU0ADgiIh5r/eAkHUBeYbDgQgtN7e9okrp168ZZ5/yarbfajPHjx7PnXvuwbK9eHXZ/Ha2ZHs9399iVv/31ft59910WW+SbHHvciey1z75VhzVV3nrzTfbfZ0/Gjx/PhJjAjjvtwpZb9a06rKnSTK+x1fv0YfsddmLN1VehW7durLjiyuy7/wFVhzVVmuF5+eijj/jLvXdz7nm//ezY4Yf+gE/GfsI2W2at9Gqr9/nC1+tdo5zH5v367Fz9yxz36ta1K1ff8Rh3P/wcB++6Pgf2WxeAm+97gktv/rwrzPO3n8jss87EDN27sfUGK9D3oPN49Y33uOW8g+nerStdu3bhL488z0U3PFTJY2oi80iqzZMujIgLJ+P75ouIliuMt4CW6ZGewGs1t3u9HJvY8YlSRExGHFNO0m3AORFxt6QfAgtFxBGS1gd+HBGblNv9BngoIi6X9ArwAXB6RFxR87NGRcRsks4AdgLeL1+aDTglIv7Q6r7HAzNGxKeSFgVuANYD/hERC5XbLAZcC2zQ1vGIWEXS/Uwkya3Vu/eq8dAj7d7EzMw60fgJHfPeVoWuXZqnBGKu1QZUHcI08/ET5w2LiFWruv/lVlwlrh30t6ru/jPLLjDbZP0eJC0C3BYRy5XP34+IOWu+PjIi5ir546kR8WA5fi9wNLA+MFNE/KwcPxYYExFnTOw+O2Qkt9RNbAgsLymArkBIOrLc5JOam49vFcdDwOaSrowvZ+Aik9oLpjCk5jnbmZmZmTW+tyXNHxFvlnKE/5TjI4AFa273zXJsBJno1h6/v7076Kia3J2AyyJi4YhYJCIWBIYD60zG9x4HjATOa+NrdwL7lHpfJPWUNG8bt+tSYgDYDXgwIj4ARkpqieE7wF8ndrx8/D9g9smI2czMzMwm3y1AS4eEPYGba45/t3RZWAP4oJQ13AlsWtZNzQVsWo5NVEfV5O4KnNbq2PXl+NWT8f2HABdJOj0ijmo5GBF3SVoGGFzWqo0C9uDz7L/FR8Dqkn5SvtavHN8T+K2kWYCXgb0ncfzicnwMsGZEjJmM2M3MzMw6XL1vq9tC0p/IUdh5JL1Orok6FbhG0r7Av4Fdys3/DGwJvASMpuRkEfGepJ8CQ8vtToqI1ovZvni/HVWTOz1xTa6ZWX1xTW59ck3utLPciqvEdYPa7ybRGZZZYNZKfw/t6cjuCmZmZmbWQdyWuX3e1tfMzMzMmo6TXDMzMzNrOi5XMDMzM2tArlZon0dyzczMzKzpOMk1MzMzs6bjcgUzMzOzRuR6hXZ5JNfMzMzMmo6TXDMzMzNrOi5XMDMzM2swonG29a2KR3LNzMzMrOl4JNfMzMys0cjb+k6KR3LNzMzMrOk4yTUzMzOzpuNyBTMzM7MG5GqF9nkk18zMzMyajpNcMzMzM2s6LlcwMzMza0SuV2iXR3LNzMzMrOl4JNfMzMys4cg7nk2Ck9xp4PHHh707c3f9uxPuah7g3U64n87gx1Kf/Fjqkx9LffJjqU+d9VgW7oT7sK/ASe40EBE9OuN+JD0WEat2xn11ND+W+uTHUp/8WOqTH0t9aqbHYl+Nk1wzMzOzBuRtfdvnhWdmZmZm1nSc5DaWC6sOYBryY6lPfiz1yY+lPvmx1Kdmeiz2FSgiqo7BzMzMzKbACiv1jlvueajqMPhWj5mH1WsNtEdyzczMzKzpeOGZmZmZWSPywrN2eSTX6o7k9aJmk0vSbP6bMfvqJHUp//ffU5Nwkmt1RZKiFIpL2kxSz6pjmhYa9aTZqHHXkjR3zcdLVRnLtCZpCeAyYKWKQ5kqkpp+NrH2b0jSDFXGMiWa4W9/SkiaC/ha+XSlCkOxachJbpNqOUFJmlnSLFXHM7lqEtwNgOOA/1Ub0VfXKnHfRtKcFYc0WVrFvaGkZauOaUqVkZkNJZ0r6XvA0ZK+NqnvaxQR8SIwHBgoaYWq45kSkuYAVisfb9KIr69JafU3tDewa8toYT2qed9YAvhuIyXl08A65PnhROCyRpkhUR38V8/q9o/NvpqICEnbATcD90j6jqRZKw5rskjaEbgCOC0iPmz0E23Nm9xWwI9pkFr4mrgPBU4DxlQa0FSIiAkRcR35BnYycHx5TXWvOLSvRKkLQET8iEx0j2+wRHc+YF1JNwO/Ajpja/ROVfM3tCbQD7ghIiZUG1XbWhJySeuTf+8HA9tJmrHSwDpJRNwCrA4cAhwUEaNanj9rXE5ym5SkZYAfAccCPwe+A+xevlZXl15txHML8F/ghwARMVZS104PbBqStBbwB+CUiHi3Ud44JG1Ivm7WjojhklaWtHHVcU1KqyniGYF7gQeBUyR1i4hxlQX3FbUkIxExQdLXASJiIPAscGK9J7otz01E/BPoSSYW1wCfVBlXR5G0CnAB8DbwUcXhTFRJcNcBfgtcBDwJrAv0a/SBholp473nXPK1uIekpSsIyaYxJ7lNSNLiwInAOxHxSETcDhwP/ETSWvV0ddpGDe7GwNfJmqj5JF0CEBHjGynRbePk+QTwGHnRQUR8Uo+Pp424XwOeAk6WdDrwS+BHkvp3enCTqdVrahlgnog4IiK2AboCfypfW1/SZhWGOlVqHtsA4GxJJ0taLCKOJROTY0tiVXdaPTf7AnMDx5CzG4dKWqB8bZ56uxifXK3jjojHyZHqxYBV6/xxrQlcHRG3AQPIC6fvANs3+uxHa61eiztK6geMj4gDgA/I98sekg6UtGelwbZDqv5fPXOS2yRanThfAf4OzFlq3WaJiMHkm/s8VcQ3MTUnmSOAgcDm5EKaJYBVgOUk3VRuO76iMKdIq5PnxpL6AnMBewBPSrpZUtd6S9xbxf19STsAo8lR0AWA64DtgAeo45KLmsdwGDmCdqmkC8r0/gFAV0mPA2cBL1UX6dSTtD+wC/B/5f8nS1onIk4g//4Pq7fZAklfq3lu1ga2JaeF/wg8AnwT2EnSMWQ9fl3FP7lqHuN+ko4t57bLgGvJi9xVVb91uS8Aa0haNiI+iYjzyQvDtYGmGtmseZ5+CBwFrAjsI+niiDgS+A95jjiSfD+1BlSvf2g2BWpqqdaWtCuwZUScAtwB7Az8UNIm5LTzf6uMtS1l5HmtiNiAnLIcDQwvU8prAF+XtECdj4B8pubkeTj5prY5cDmwDFmC8R/gXkld6ilxbzVCuA/wVESMiIgrI2LXiHgU6EvWFj5WYaiTJGkPYIeIWJdMoPYAfh8R/4uIHcj63O0j4l9Vxjm5JK0maVtJM0qaDViSTG53IJPa4eQI7jrlDfqwiKib6X9JiwEHS5pJufDyCLJUYVmAiLgZuBuYA9iafK4+rijcr0zSIUB/4CHgu2Qyfw5ZMvNL6mD1fsv5VFJvSRtJWggYBAwjR25XV3Yj6QIsSF6UNBVJMwEbALtHxDHkc9Zd0tGl1v1nwLcj4qkq47Sp5yS3CZQEdzPgN8BywCGSboiIM4CngR3JN8T9IuLBqkcR2khWxwH/k3QOeTW9S5nO34GcPlonIt6opzKLSZG0JJm4r0dO+Y8hk63RwOHAM+SbfF2RNA85WrsX8Jak/pIOK1P7fcjk97sR8XyFYX5JG6+pp4DdJR0M9CJHoteTdIOk7hFxXUS80tlxfgVLkIsWN46IUeXjWYGtImLjUpO7GLBFmbl5t8JY2zKBPD8tTc4mHUL+DXy7JMCUsqqfk4+xoZKK2nNq+XgRYBNyNuo14Lwye3MacDXwThVxtqgZGNmCnOFbExgKLE/O2HxK1qdeCnyfvEj/Wj3NPE2NNs4TXciSmaUg13+Qj3W+8vnzEfGfTg1yCqkO/tWzup1ytMlXTqp7AD+LiKvLsUGSzo+Ig5T9/+YBPixv8JUtumk1Jd6PTEb+BXxIroDfpiS4+5En1wfJkc+GIakH8CbwuqSLgG+Qj2tCSdzvAQbUQ9Je+3wU7wPPk9PF3YCRwCzAeDJJ2TEi3u/kMNvV6jU1B3nd91T5u+gD/DYiPpB0JZl4zEnFScbkKqP9EyLiSkm9gP8rSey1ksYBC0tag3yjfg44LyJGVxp0jZbnJnLR4uzkzNI3gJ+SCe1PgJB0W0S8GNl5oG4XZ01MiZsy2PAIOfJ5N/n3tH3k4tnvSfpXRPy6ukhTSXB7kTMaWwCLly9dDhwcEaeVc9en5MLAk8jBh7qZeZpSrc4Ta5Kzmm8Bp5AXISMj4mGybGaxMsr7ST2cp23qeSS3AUmaoUwjIWkRMoF9H6j9YzyQHOmBPJGNI0foKl08UHOSOZh8g/ukXD3fDNxF9ic8GTiUHDFstAR3JXIqtuX3vATwg4gYp+yTeQIwSz2cOFud9LctpS6LAb8n6wePiIj9gL+Sq6yj3hJc+FJd9yXAA5J2J0dv/wH0lXQSsDKwc0Q0RIILX0ieBpDlLq8CZ0naPiJeBc4GziSnVQdGxIiqYm2t9QVURPyPXLn/Mll//wmZ7K4HbKoG3BhC0irKjgSUJP7E8jfyG+Bb5CKuscqFS4eQF/RVxbqYpB0kbQsQEc+QpUffAH4eEfMBfwTukLRh+TsZT5bE7BwRT1cV+7RQc574ATlK/WOyk88E4AfAtZLOBw4D/i8iPq6H83S76mDRWb0XETbcSWV6p2wZ1AtYStI+wGbAWsCjwDmSniknr0XJq9EeEfGOpKOAuaoa5WmVUC1PTntvHhEjJK1HnkyvBwaTCeIfGqVespXu5GjhRWTLsIOAMyS9AaxPjoa8VV14n6t5Pr5PrqC+hCxvWT8irlPai1xl3S8iPq0s2DZI6k0uivknOS28C7AxWQO9EflcPACMLccHRsRr1UQ79cp0/t7kiOCr5WLkmDL9/VtJNwDU0wWhpEUj4uXy8aFkGdJwciHgBeTfxZFkgn4EMLreXl+TUpLyPsDOko6NiIckdVFuNvIQcDRwurI/9vLATi2/kwpiXRK4AbgTWFPS/BHx24j4p6R1yfcPyFHoh8myKiJ7Sg+ocvZvWlJucrEXsBU5e/ht4HQy2d+QPJ+c0ojnCWubk9wGouxV2J+cYlkR2B84OSI+IkdA5wauk3Q3+aZ+ZElwu5Q3kMpGsFpdEb9MnkiPLSVSywDvkqMe11QQ3lcmaX7g7YgYKulScnpvd7Kp+qJkd4UzImJ4hWECX7rgmBfYlJyy3JFMCoeUm/Yga7z7RcSzVcQ6MZI2J0cBfwW8Qcb6QkR8CFwj6X3gHLJm9ZxSutMQb9RtlJC8RibyC0l6MyL+pOzhebGkURExqJpI21YuxO+QdBk5O7MjmdguD/yOnGU6j0xyBwBHNVqCCxARn0q6hrxAHyjpbODekhh2LReKT5LlF59WdRGi3EnuCnJ08lbloszZJa0UEU+QHRU2Va6J2AA4ICKG1JSaNMTfTVtayn1q/qbGA2/VDDTcI+kWYL2I+H11kVpHcZLbQMq0F+SK5F+RV6IzKHuWXl/ezB8je/z9MSL+Xv6462KHnVKvdkxErCfpLnL07cZSP3k80JtMUFq/ydc15YKs/cjawiPIhRvzAwuX0ehKRm/a0irB3Y0cwRkCnEHWEW4e2drscLI+b2C9JSBl5P9XwG4RMbQcewHYUlKfyN7Qd0kaTJYsvEzWFta9Vs/PIsC4MtvxGpmAjCBHRB8H/kKOvNcNZbu8XmQ9/S/JOvvTIuI2Za36weRU/sHkCFqXent9TUrtcxQR/5V0FfleegrZHqwn0FPSW+S25AdXXMs6N7BiRNxaPj+KfB0dqGyltz/Zrm0t4OiIGAJfGphoVC2PYUFJb0XEy5LGSbokIlp633YnFwo2qDqvF6iYk9zGcz7wZ+DdiDhG2apmDXJR2YfkyNuFLYltlSeqNpLVu4GDJF0bETsDt5bb7UbWC+8K9X9ybZWIdCcb8J9DTilfB9xETpnPSo5U1Y2auHcmp4wfJWsH1yRb5YwrX9uTvHCqxwSkN/DrMmrercQ4nHzj3r7USL5L1noeB/X/moIvva5+BOwLvCzpr+RmLr8Bflpec0uTiwBfryzgVkqC+3PgpxFxX5mmv5lsPXVbmVX6FVmPeybZtqmhFjK1eo42Aj4GXoqI8yUFOc1/PzmC3Q3oWvVjjOyos5Wkl8kLvusi4qQyM/g0cGhEnEkuiG3rvN1wynOjiLhH2S97L+Afkt4uH/9B0p3kwua+ZKmTNSEnuXVOucKze0T8T9JCpSbvB8D+Zar5j+Sb4dbktODe9TJyW/NmsCQwIiI+UnYXuFLSnyNiS+WOVNtSh22pJqbmcX2fHP0YA1weEYcrt8HtSXYk2FDSXBExsrpov0zSamS/3j9FxEuSjgOuBM4tr7fFyVHSVyoM80tq3ny/Rc5WAIwvU5IfSPoFOYK4TPnaNvWUBE5KzeuqDznL0ZccYbuCTJb2krQyeSF7TNRRzbqkb5Ct8fYrFx+zlhHow8jNOAZExK/LyOcp5AhuQyW48KVe0t8lL9QvlrRRRPymJLp7k6UzD1UY6hdExB3KDUTuJMuTWmYGTyO7jdTetqET3GJe4AplPfgyZJlfV3JR2blkl4/9yXr9fhHxQkVxWgdzklv/NgCWlDSSbB+0LtlJYTzQp9RYXQDMDpwdES/Uy5W4soXTgmQfxl9LujoiRitXGv9F2Taor6R9IuuKG4ak7cmR0P5k+7ZtJC0YEZeXrz8KjK2HBFe52OLrZOI9DHidbDe1k6QHI+JJSTuSNZMzk2/Qr1YW8ETUvKZvJBde9Y6IYUrdImJkmc04DXglsmtHw1DWIq1A1q0+BrxaSkd2IFd+LxARh1Cfuy99QnZw+bhcKB0paX3gbbKm+GjlItjjI+K96sL86pRbj+9KnpsPJi9EnpK0SuRCwPFkF4y6EhH3StqGrO9eXLkJz5HkBW/TKO9/f5L0CVlydX1EPFdmQA4jFwX3jogLKg10GhD1392gam4hVv/uJKfyzwOOjYh3Ihu93wUcV97oR0fE2y1Xo1WXKLR8HNnf899k65w9yKRq1sidjG4DZpPUsxESXEnrKhunt1gCuCqyk8VPyFZVO0iaBSAiXoj6WGS2Fdl8/mhyC9inyIT3JHKR2X6SlouIDyLiwYi4ux4T3FYeIacZ+5XX/4TIRUD9ya4dYxolwW319xIR8SRZb7w40EfZ1/rf5OrvFSTNV/s9deR98lx1BrlV8iJkgnEuucj0HGAjZV1uQ2nj9/04sFP5t2lELESO6D5fLnR/F3W6Oj8i/gwMkDSaLCU5NCLuqjisaa4kujeQXWN2l7ReRIyLXJg6luyFa9MBj+TWoVYjsb3IkoTRwMqShgJvRMT9Zaqpr6Tn6yVRrJnO25tsz/I2cA3Z9/ZsoIeyYf9qNFbP0vmA8yV9J3I1+zPkwo0VIndnulhZW7wYmfBWTtmB4Fhyi9e/lmMnkBcYW5AXTt8HjpB0SqNM2ZWyl9+RZTq/VC62HENJPOo1wWhLzd/LTuQCoUci4ndlFuQE4CRJj0RuprBJ1GeNdMvmAheQCe2CwM1RthWWdCBwL3BmPcwwTYnac7HKFrcR8Vz5fEFyVgHygnFuclS3rkXEIElbA1+LOuvMMS2U16LKc3e9pO+QXRROIDe6WZmscW8K9XjFW0+c5NahVrWsZ5NXo5eT0/6Hk2UL3yZHE+ch93uvLMlteSOo+f/+ZJugX5Cr268ADiAbbm9DjlId3QgJrrIX62yRO0yNJxPdg8g3tTWBXSUtTM6KzEO2d6ucsp3cn8m61L9KmimyufkJZWTqFnJq/Eay1/IH7fy4ulPqPX8B3Ee2y3uTfKz/rDayyaPctWx0+fhQsp7+HuBsSVdFxAXl9XYmeYE4uF4T3BaR2w0PLv+AzxY49gKGN1qCC184F/+I7K06rpTEfI/sbrOicjHdGuTr783Kgp0CEXEvNM0isy89hvI+1EUSEXGFpLHkjNYlwGYNMFtl04iT3DqiXLyxIFk3uShZknBjRLxRvr4vWU90DrAlucDh4ZavV2hhsgYyJM1GdhU4rkyNIWk4uaJ6t4g4VaV3YYXxTon1yDKE/4uIG5R7t59H/u5/T3ZROAgYRS76q4vEPSLeK6M1p0oaHLngZ8aI+CQijle24Vo6ss3cCy2jbo0kIsaQZQsPVh3LlCglJJuUJP2bwJoRsY6y/dwcQG9JB5ZEdyzZNaKhKPtG9yMX9/SLiJcqDmmqSdoE2DgiNpL0M2C18vd1OXmBuCKwT6MkuLWaKcEt5WSzk91uXo2IMTUjutdKGgW87AR3+uIkt770Bf4GzBi56v1yYEdJZ0XEqxExSrloa2WyhdITLfViVV2RK9sG/VK5i9m+ZJI+N7m715/LzQaTLVq6w+dbldazlt9nRPxSuYDhOEk/LSdLkSUkB0QuNLmEbFdTyW5yExMRt0uaADwqadXIhVndI5u7f0AuXqQRE9xGpc/bbB1fRqPfA35UEt++wOpkjfcPyihUoy6OeR94Edi2kRPc4r/kJjs/I8us+pbjS0XEn8gZNutENefn2pH27chymcPJLXvvaVW6cEd1EXecuqzQryNeeFZHIndcGQmcJ2mDiPgJOa18raQFym3GRMTDkTvV1H5vFQnuZmRD963JqfrtyNZGBwJzSvpjGfncgBztnaGzY5xaNSfPBSPiPHLL4WMlrRu5K9tRwPWSNi3PSV0luC3KiX0A8Jiyndk4Sd8l96uvi9KK6YW+2GbrJkkzAxPIsrplgLtKScKr5Jv1jRP9YXWu/E3c3mgJbsugQavPPyXLE1YhN0sZJ2k/4BeS5mr9PdYpvg7ZwUfZJWL1iFiXPKeNBO6T1L11MmzTH4/kVqzlBNnyRxgR/5H0OlnrOS4ijpZ0MnB3Sai+MHVZ1R+vpE2BS8mp4oXJBUyzkG/Wz5KjHTeW2ywO7BsVbWs5tSStABwl6boyYisy0T0pckHDOOpoN7OJieyROQB4QNL5ZI13wz0fTaB1m62jgbXJJGpucresxYF1gb5+fjpfzcXtIWTN+itkh4hzyUT3IEnzUTaviTpoETg9KefgHsBwSbtGxC1lNuQdSVeS26f3jdzKdxdyV8C3KwzZKuYkt0ItNZLl42+Tf7yDI+K4Up+3T7kQPUbZ428R6qA+T7mbzK/JnoPfIEc4niSb9PeR9GFkE/5NJM1KNrL/sLKAJ0Pri43y8VOS/g5sJmlCfN7s/UxJh0bELVXFO6VKotsVuAFYObL1mXWu9/m8zVYvcqHZVWTP4i2Bi8kuKic32ghoo9MXFwKuTXbquIBMdC8mN374D7kRR1eyi0dDLHJsNmUgaB/gj5L2iuwV/xFZxndAZDvBPYEjyPK/pib3V2iXk9yKSJoTuEnS98iykUvJ/pJbKTdJOKPUGQ2Q1DUijqww3NY+BPaKiIeVO5b1J1s4vUm2DZsg6f6I+HfUSWuzydC1TBW3lGHMGxGXRcSZkn4IbCdpbBnRHUtuqNBQIuI2SXPWa2lFsyv1gRNrs3UA8HhEXF9ljNMjfb4Q8HRyFP0g4LTy99KD7ApzCU3aU7aR1Mx4Xi3pA+AaSduSXYjmBE6T9A7QB9il9cynTX9ck1uRiHgfuAO4jOzZ1zciNic7K2wqaZuI+CW5u9H7VcXZlogYWhLclp6RV5I7Zb1HjnZsAaxVRg7rnnL19GWSBip3l5oV2E3Z/oiIOLfc9GeStoiIi6JBV+g6wa1WRIyKiMERcU1NgrszOUL4RKXBTYfKQsCTgb9Edql5iEyW+gJEdks5hxyAOLPUefp9s5PV/s4l7Srp8Mgev/uQm1osHhHfA04hy+S2mG5mq1QH/+qYR3IrUFMMf5qkd8kdji4hG1VfCwSwbRnBPb3KWNsTpUtC5FbCV5Itg8aR069/iQbYm165YcJJ5MXGvGRLsOvIVm3fKc/VNWQ/1pnIixCzr0xfbrP1r4pDmq60Wgg4VLkb42vKPtiXSRoQEb+ObL93KrkRxLhqo57+SFqRbIW4U5kZ7Am8CxAR1yk7yNwsab+IaNjFmtYxnOR2spYEt9R9fRARf5A0DznN8nZEDJN0HVn31TA1XyXRvY7c7OF3EfHfqmOaFH2+YcK2pa5rIbJbxIxkogu549S2ZC/M7b0YyKah92meNluNqPVCwCMlrU8uVHoNOFpSj4g4PiLeqy7M6VtEPCnpU+BqSduTi8verfn6DWVNxbmS7gZGRwO0qbTOIXfW6HzKptXnAXtGxN/KsYPIqZeDI+IRSd2iznc4aos+78PaEEo93ulkQ/4PJV0BPBClP6mkXsA6ZM9FJyJmTaIkRj8CNuXzhYAPkjNRW5OJ1HbkxW1dbPIyPSnPT5eWGUFJ1wNjgX+V/z9HXqhAlpmMjtwgZrqx4sq9466/Dqk6DL4xxwzDImLVquNoi0dyO5mknsCpQP+IeLRMxXyNXPUO8IfSaWFUVTF+FY2U4MIXNkwYJulOsrb4Uvhs1P0ZYPqo7TKbjkxiIeCBwL3AmeGRoE7XMuMJjJfUMyJGRMSO5fk6hux8MQf53jkz8EQjzB5a53OS28EkLQssHxFXl0OjgLuBvpL2JnvITgCujYjzJd1R7+22mk1pr/V9chvlb0RuBzlTRHxcdWxm1nEiYhS5I+PglmNlIWAvYLgT3M5Xk+Ci7O+9u6RhwPkRcWApXVgwIrYpt5khIsZWGLLVMa8S7UCSlgSuIFfrAxARH5CbJcwA3BQRm5CjuL3L14dXEOp0LyLuAbYC/iJpXie4ZtMXSfNLOhQ4gSwlc3lSBWoS3O2ADckdGwUcKGnNiDgY6CKppRa34cr6phWpPv7VM4/kdhBJSwG3AddFxEXl2MylZugyYEJEjJfUh9wtbGB10Rp8NqI7AzBI0qp5yCM5ZtOJ9/FCwLogaWmy682VZTH2y8AhQP/SdaivpAXK+dnnaJsoj+R2gFKicDm5JeQHktaC3M9d0qLA74BvKLeNPRQ4LiIGlatSq1BE3AysGxETnOCaTT8iYkxE3O4Et/Mpt0qu9SE5w7l7Gb0dCZxJLjjbppSTvdHZcVrj8UjuNCZpZrJzwtnkSO7hwNaSxgOPkT1xH47ciWWEpMMi4q3aOiSrVqnTMzOzDlZGbZ+VdA7wXERcGBFvlN7E7wPHSDo5IgZLOhHweoka3ta3fU5yp7EyWrtrRLwFUFpS7U72j10AODoinlbuFjah5XZOcM3MbDo0iuxw8Rawk6R1gGuA+yLibEnjyD7yR0XEEBq085BVw+UKHaAmwe0SES+QNbifkhsKzFlu42bVZmY2XYuI14FHgVWALYHbyV0A/1zWRjwJ/BoYUVmQ9azqLX3rfCDZSW4His+3vX2RTHRnIuuJ5qo0MDMzs4rVrEMZSC4gm4cc0V2O3Ob+x0B/4M6IeK2SIK2huVyhk0TEi5J+Xz4eWXU8ZmZmVSobcrQkui+Si8t6Az+KiJtKG853SutNsynmJLcTlRFdMzMz47P1KGMlXQ78FTgvIm4qX/tnlbE1gjqvFqicyxXMzMysUmX9ykCgq6RZqo7HmoOTXDMzM6sHQ8gFaGbThMsVzMzMrHIR8byk/hExuupYGoW3kGqfR3LNzMysLjjBtWnJI7lmZmZmDUfe8WwSPJJrZmZmZk3HSa6ZmZmZNR0nuWbWUCSNl/SEpKclXftV2g1JuljSTuXj30tatp3bri/p21NxH69Immdyj7e6zagpvK8TJB0xpTGaWeMRufCs6n/1zEmumTWaMRGxUkQsB4wFvlf7RUlTtdYgIvaLiGfbucn6wBQnuWZmVg0nuWbWyP4GLF5GWf8m6RbgWUldJf1C0lBJT0k6EEDp15JekHQPMG/LD5J0v6RVy8ebS3pc0pOS7pW0CJlMH1ZGkdeR1EPS9eU+hkpaq3zv1yXdJemZspX3JMc6JN0kaVj5ngNafe2scvxeST3KscUkDSrf8zdJS0+T36aZWRNxdwUza0hlxHYLYFA5tAqwXEQML4niBxGxmqQZgYck3QWsDCwFLAvMBzwLXNTq5/YAfgesW37W3BHxnqTfAqMi4oxyuyuBsyLiQUkLAXcCywDHAw9GxEmStgL2nYyHs0+5j5mBoZKuj4j/ArMCj0XEYZKOKz97AHAh8L2IeFFSH+B8YMOp+DWamTUtJ7lm1mhmlvRE+fhvwB/IMoJHI2J4Ob4psEJLvS0wB7AEsC7wp4gYD7wh6b42fv4awAMtPysi3ptIHBsDy+rzorSvSZqt3McO5XtvlzRyMh7TDyVtXz5esMT6X2ACcHU5fjlwQ7mPbwPX1tz3jJNxH2Zm0xUnuWbWaMZExEq1B0qy91HtIeAHEXFnq9ttOQ3j6AKsEREftxHLZJO0PpkwrxkRoyXdD8w0kZtHud/3W/8OzMzsi1yTa2bN6E7g+5K6A0haUtKswANAv1KzOz+wQRvfOwRYV9K3yvfOXY7/D5i95nZ3AT9o+UTSSuXDB4DdyrEtgLkmEescwMiS4C5NjiS36AK0jEbvRpZBfAgMl7RzuQ9JWnES92FmTajqzgrurmBm1vl+T9bbPi7paeACcubqRuDF8rVLgcGtvzEi3gEOIEsDnuTzcoFbge1bFp4BPwRWLQvbnuXzLg8nkknyM2TZwquTiHUQ0E3Sc8CpZJLd4iNg9fIYNgROKsd3B/Yt8T0DbDsZvxMzs+mKIqLqGMzMzMxsCqy8yqpx/0OPVh0Gc87SdVhErFp1HG3xSK6ZmZmZNR0nuWZmZmbWdNxdwczMzKzRNMDCr6p5JNfMzMzMmo6TXDMzMzNrOi5XMDMzM2swKv9s4jySa2ZmZmZNxyO5ZmZmZo3IQ7nt8kiumZmZmTUdJ7lmZmZm1nRcrmBmZmbWgOR6hXZ5JNfMzMzMmo6TXDMzMzNrOi5XMDMzM2tA3ta3fR7JNTMzM7Om45FcMzMzswbkgdz2eSTXzMzMzJqOk1wzMzMzazouVzAzMzNrRK5XaJdHcs3MzMys6TjJNTMzM7Om43IFMzMzswbkbX3b55FcMzMzM2s6TnLNzMzMrMNI2lzSC5JekjSws+7X5QpmZmZmDUY0xra+kroC5wGbAK8DQyXdEhHPdvR9eyTXzMzMzDrK6sBLEfFyRIwFrgK27Yw79kiumZmZWYN5/PFhd87cXfNUHQcwk6THaj6/MCIurPm8J/BazeevA306IzAnuWZmZmYNJiI2rzqGeudyBTMzMzPrKCOABWs+/2Y51uGc5JqZmZlZRxkKLCHpW5JmAPoDt3TGHbtcwczMzMw6RER8KmkAcCfQFbgoIp7pjPtWRHTG/ZiZmZmZdRqXK5iZmZlZ03GSa2ZmZmZNx0mumZmZmTUdJ7lmZmZm1nSc5JqZmZlZ03GSa2ZmZmZNx0mumZmZmTWd/wekDYqL1i/6BAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 720x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from plotcm import plot_confusion_matrix\n",
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "# getting all predictions on train_set\n",
    "all_preds = torch.Tensor([]).to(device)\n",
    "for batch in predict(network, train_loader, device):\n",
    "    all_preds = torch.cat(\n",
    "        (all_preds, batch),\n",
    "        dim=0\n",
    "    )\n",
    "cm = confusion_matrix(train_set.targets, all_preds.cpu())\n",
    "plot_confusion_matrix(cm, classes)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}