notebooks/draw_moon.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import csv\n",
    "import ndjson\n",
    "import os\n",
    "import sys"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "root_path: C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\n"
     ]
    }
   ],
   "source": [
    "base_path = os.getcwd().split('sliver-maestro')[0]\n",
    "root_path = os.path.join(base_path, \"sliver-maestro\")\n",
    "print(\"root_path: {}\".format(root_path))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "sys.path.insert(1, root_path)\n",
    "from src.utils.model_utils import *\n",
    "from src.draw_model import *\n",
    "from src.generate_images import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "categories_file = os.path.join(root_path, \"src\", \"data\", \"categories.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(categories_file, newline='') as f:\n",
    "    reader = csv.reader(f)\n",
    "    categories = {rows[0]:rows[0] for rows in reader}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Select a category to download images from "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of categories: 345\n"
     ]
    }
   ],
   "source": [
    "print('Number of categories: %d' % len(categories))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['aircraft carrier',\n",
       " 'airplane',\n",
       " 'alarm clock',\n",
       " 'ambulance',\n",
       " 'angel',\n",
       " 'animal migration',\n",
       " 'ant',\n",
       " 'anvil',\n",
       " 'apple',\n",
       " 'arm']"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list(categories.keys())[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Selected category from the list above: moon\n"
     ]
    }
   ],
   "source": [
    "category = input(\"Selected category from the list above: \") or 'cat'\n",
    "# Checking to see if the entered category is in the categories dictionary\n",
    "assert categories.get(category)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Creating directories to save images, weights and the model ...\n",
      "folders are created...\n"
     ]
    }
   ],
   "source": [
    "print(\"Creating directories to save images, weights and the model ...\")\n",
    "src_list, dst_list, paths_dict = create_folders(root_path, category)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Select a download method ... \n",
      " 1. Use Google Cloud credentials stored in a json file \n",
      " 2. Use gsutil in the notebook \n",
      " 3. Use gsutil in the terminal \n",
      "2\n"
     ]
    }
   ],
   "source": [
    "option = input(\"Select a download method ... \\n 1. Use Google Cloud credentials stored in a json file \\n 2. Use gsutil in the notebook \\n 3. Use gsutil in the terminal \\n\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "gs://quickdraw_dataset/full/numpy_bitmap/moon.npy C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\\src\\data\\input\\moon\\moon.npy\n",
      "gs://quickdraw_dataset/full/raw/moon.ndjson C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\\src\\data\\raw\\moon\\moon.ndjson\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Copying gs://quickdraw_dataset/full/numpy_bitmap/moon.npy...\n",
      "/ [0/1 files][    0.0 B/ 91.0 MiB]   0% Done                                    \n",
      "/ [0/1 files][264.0 KiB/ 91.0 MiB]   0% Done                                    \n",
      "-\n",
      "\\\n",
      "\\ [0/1 files][  8.2 MiB/ 91.0 MiB]   9% Done                                    \n",
      "|\n",
      "/\n",
      "/ [0/1 files][ 16.8 MiB/ 91.0 MiB]  18% Done                                    \n",
      "-\n",
      "- [0/1 files][ 27.6 MiB/ 91.0 MiB]  30% Done                                    \n",
      "\\\n",
      "|\n",
      "| [0/1 files][ 38.4 MiB/ 91.0 MiB]  42% Done                                    \n",
      "/\n",
      "-\n",
      "- [0/1 files][ 48.2 MiB/ 91.0 MiB]  53% Done                                    \n",
      "\\\n",
      "|\n",
      "| [0/1 files][ 57.8 MiB/ 91.0 MiB]  63% Done                                    \n",
      "/\n",
      "/ [0/1 files][ 64.2 MiB/ 91.0 MiB]  70% Done                                    \n",
      "-\n",
      "\\\n",
      "\\ [0/1 files][ 72.2 MiB/ 91.0 MiB]  79% Done                                    \n",
      "|\n",
      "/\n",
      "/ [0/1 files][ 79.4 MiB/ 91.0 MiB]  87% Done   6.8 MiB/s ETA 00:00:02           \n",
      "-\n",
      "- [0/1 files][ 87.9 MiB/ 91.0 MiB]  96% Done   7.4 MiB/s ETA 00:00:00           \n",
      "\\\n",
      "\\ [1/1 files][ 91.0 MiB/ 91.0 MiB] 100% Done   7.4 MiB/s ETA 00:00:00           \n",
      "\n",
      "Operation completed over 1 objects/91.0 MiB.                                     \n",
      "Copying gs://quickdraw_dataset/full/raw/moon.ndjson...\n",
      "/ [0/1 files][    0.0 B/315.6 MiB]   0% Done                                    \n",
      "/ [0/1 files][    0.0 B/315.6 MiB]   0% Done                                    \n",
      "-\n",
      "\\\n",
      "\\ [0/1 files][  1.3 MiB/315.6 MiB]   0% Done                                    \n",
      "|\n",
      "| [0/1 files][  9.8 MiB/315.6 MiB]   3% Done                                    \n",
      "/\n",
      "-\n",
      "- [0/1 files][ 18.6 MiB/315.6 MiB]   5% Done                                    \n",
      "\\\n",
      "|\n",
      "| [0/1 files][ 26.6 MiB/315.6 MiB]   8% Done                                    \n",
      "/\n",
      "/ [0/1 files][ 34.3 MiB/315.6 MiB]  10% Done                                    \n",
      "-\n",
      "\\\n",
      "\\ [0/1 files][ 43.0 MiB/315.6 MiB]  13% Done                                    \n",
      "|\n",
      "/\n",
      "/ [0/1 files][ 52.8 MiB/315.6 MiB]  16% Done                                    \n",
      "-\n",
      "- [0/1 files][ 61.4 MiB/315.6 MiB]  19% Done   8.4 MiB/s ETA 00:00:30           \n",
      "\\\n",
      "|\n",
      "| [0/1 files][ 70.9 MiB/315.6 MiB]  22% Done   8.7 MiB/s ETA 00:00:28           \n",
      "/\n",
      "-\n",
      "- [0/1 files][ 80.7 MiB/315.6 MiB]  25% Done   9.1 MiB/s ETA 00:00:26           \n",
      "\\\n",
      "\\ [0/1 files][ 87.4 MiB/315.6 MiB]  27% Done   8.6 MiB/s ETA 00:00:26           \n",
      "|\n",
      "/\n",
      "/ [0/1 files][ 95.9 MiB/315.6 MiB]  30% Done   8.4 MiB/s ETA 00:00:26           \n",
      "-\n",
      "- [0/1 files][106.2 MiB/315.6 MiB]  33% Done   8.7 MiB/s ETA 00:00:24           \n",
      "\\\n",
      "|\n",
      "| [0/1 files][115.5 MiB/315.6 MiB]  36% Done   8.7 MiB/s ETA 00:00:23           \n",
      "/\n",
      "-\n",
      "- [0/1 files][124.5 MiB/315.6 MiB]  39% Done   8.5 MiB/s ETA 00:00:22           \n",
      "\\\n",
      "\\ [0/1 files][133.8 MiB/315.6 MiB]  42% Done   9.2 MiB/s ETA 00:00:20           \n",
      "|\n",
      "/\n",
      "/ [0/1 files][142.8 MiB/315.6 MiB]  45% Done   9.2 MiB/s ETA 00:00:19           \n",
      "-\n",
      "\\\n",
      "\\ [0/1 files][151.6 MiB/315.6 MiB]  48% Done   8.9 MiB/s ETA 00:00:18           \n",
      "|\n",
      "| [0/1 files][159.8 MiB/315.6 MiB]  50% Done   8.7 MiB/s ETA 00:00:18           \n",
      "/\n",
      "-\n",
      "- [0/1 files][168.4 MiB/315.6 MiB]  53% Done   8.6 MiB/s ETA 00:00:17           \n",
      "\\\n",
      "\\ [0/1 files][175.8 MiB/315.6 MiB]  55% Done   8.3 MiB/s ETA 00:00:17           \n",
      "|\n",
      "/\n",
      "/ [0/1 files][184.6 MiB/315.6 MiB]  58% Done   8.2 MiB/s ETA 00:00:16           \n",
      "-\n",
      "\\\n",
      "\\ [0/1 files][193.6 MiB/315.6 MiB]  61% Done   8.3 MiB/s ETA 00:00:15           \n",
      "|\n",
      "| [0/1 files][200.1 MiB/315.6 MiB]  63% Done   7.9 MiB/s ETA 00:00:15           \n",
      "/\n",
      "-\n",
      "- [0/1 files][209.3 MiB/315.6 MiB]  66% Done   8.0 MiB/s ETA 00:00:13           \n",
      "\\\n",
      "|\n",
      "| [0/1 files][219.1 MiB/315.6 MiB]  69% Done   8.5 MiB/s ETA 00:00:11           \n",
      "/\n",
      "/ [0/1 files][228.7 MiB/315.6 MiB]  72% Done   8.7 MiB/s ETA 00:00:10           \n",
      "-\n",
      "\\\n",
      "\\ [0/1 files][237.7 MiB/315.6 MiB]  75% Done   8.8 MiB/s ETA 00:00:09           \n",
      "|\n",
      "/\n",
      "/ [0/1 files][246.2 MiB/315.6 MiB]  78% Done   9.1 MiB/s ETA 00:00:08           \n",
      "-\n",
      "- [0/1 files][253.2 MiB/315.6 MiB]  80% Done   8.6 MiB/s ETA 00:00:07           \n",
      "\\\n",
      "|\n",
      "| [0/1 files][259.6 MiB/315.6 MiB]  82% Done   8.0 MiB/s ETA 00:00:07           \n",
      "/\n",
      "-\n",
      "- [0/1 files][268.7 MiB/315.6 MiB]  85% Done   7.9 MiB/s ETA 00:00:06           \n",
      "\\\n",
      "\\ [0/1 files][278.0 MiB/315.6 MiB]  88% Done   8.0 MiB/s ETA 00:00:05           \n",
      "|\n",
      "/\n",
      "/ [0/1 files][287.5 MiB/315.6 MiB]  91% Done   8.1 MiB/s ETA 00:00:03           \n",
      "-\n",
      "- [0/1 files][296.5 MiB/315.6 MiB]  93% Done   8.6 MiB/s ETA 00:00:02           \n",
      "\\\n",
      "|\n",
      "| [0/1 files][306.6 MiB/315.6 MiB]  97% Done   9.3 MiB/s ETA 00:00:01           \n",
      "/\n",
      "-\n",
      "- [0/1 files][315.6 MiB/315.6 MiB]  99% Done   9.1 MiB/s ETA 00:00:00           \n",
      "- [1/1 files][315.6 MiB/315.6 MiB] 100% Done   9.1 MiB/s ETA 00:00:00           \n",
      "\n",
      "Operation completed over 1 objects/315.6 MiB.                                    \n"
     ]
    }
   ],
   "source": [
    "if option == \"1\":\n",
    "    download_data(src_list, dst_list)\n",
    "elif option == \"2\":\n",
    "    for (src_file, dst_file) in zip(src_list, dst_list): \n",
    "        print('gs://quickdraw_dataset/' + src_file, dst_file)\n",
    "        !gsutil -m cp {'gs://quickdraw_dataset/' + src_file} {dst_file}\n",
    "elif option == \"3\":\n",
    "    print('Type the following in the terminal to download the files manually:')\n",
    "    for (src_file, dst_file) in zip(src_list, dst_list):\n",
    "        print(\"gsutil -m cp \" + \"gs://quickdraw_dataset/\" + \"{}\".format(src_file) + \" \" + \"{}\".format(dst_file))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_path = os.path.join(paths_dict['input_path'], category, category)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "npy_data = np.load(paths_dict['base_input'] + '.npy', allow_pickle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load from file-like objects\n",
    "with open(paths_dict['base_raw'] + '.ndjson') as f:\n",
    "    ndjson_data = ndjson.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_bitmap():\n",
    "    print(\"Bitmap image:\")\n",
    "    plt.imshow(npy_data[int(index)].reshape(28,28))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_raw():\n",
    "    print(\"Raw image:\")\n",
    "    i = 0\n",
    "    for stroke in ndjson_data[int(index)]['drawing']:\n",
    "        stroke = np.array(stroke)\n",
    "        plt.plot(stroke[0], stroke[1]*-1)\n",
    "        i += 1\n",
    "    print('There are', i, 'number of strokes in this raw image file!')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are  121662  images in selected category moon\n",
      "Select an index value from 0 to 121661 to display and compare bitmap and raw images:\n",
      "Bitmap image:\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Raw image:\n",
      "There are 1 number of strokes in this raw image file!\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX8AAAD4CAYAAAAEhuazAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy86wFpkAAAACXBIWXMAAAsTAAALEwEAmpwYAAA4+0lEQVR4nO3dd3gc1b3/8feRVr333t1tWS5yAzc6mN5LKJdqCBBufiGFkHtvctNIgNwAIQQIvfdux4DBNrhhuVfZqrZ6byttP78/dm0EkWyDtFpJ8309zz47OzvyfscrfXb2zJlzlNYaIYQQxuLn6wKEEEIMPQl/IYQwIAl/IYQwIAl/IYQwIAl/IYQwIJOvCzhe8fHxOjs729dlCCHEiLJ58+YmrXXCt9ePmPDPzs6mqKjI12UIIcSIopSq7Gu9z5p9lFJnKqWKlVIlSqlf+KoOIYQwIp+Ev1LKH3gUOAuYBFyplJrki1qEEMKIfHXkPxso0VqXaa1twKvA+T6qRQghDMdX4Z8GHOr1uMqzTgghxBDwVfirPtb92yBDSqlblFJFSqmixsbGIShLCCGMwVfhXwVk9HqcDtR8eyOt9RNa60KtdWFCwr/1VBJCCPE9+Sr8NwFjlVI5SqlA4ArgfR/VIoQQhuOTfv5aa4dS6g5gBeAPPK213u2LWoQQg8vhdFHR3M2B+k5q2i1cPCON6NBAX5clvsVnF3lprZcBy3z1+kKIgemxOSlvMlPW1EV5o5mSxi6K6zopazRjc7qObJcZG8ppk5J8WKnoy4i5wlcIMfR6bE6q27o51NJDRbOZssavw76m3fKNbVOjghmXHMGicQmMTYrgq/JmXi+qoiAjykfVi6OR8BfCwGwOF1Wt3VS19nDIc1/V2sOhFvdyU5f1G9tHBJnITQhjTm4cOfFh5CaEkRPvvoUGfjNO3t1azfikCBIjgodyl8RxkvAXwgC01jR2Wtlb18m+2g721nawr66TkoYuHK6ve1kH+CtSo0NIjwnh1ImJpMeEkB4TSnpMCFlxYcSHB6JUXz213VwuzZoDjTy/vpK1pU3cND9nKHZPfA8S/kKMMha7k5KGLvbWdrC3tpN9de6gbzHbjmyTGhXMhJRITp6QSF5COBmxoWTEhpAYEYy/X//h3p8Oi523Nlfx/PpKypvMJEQE8aOTx3LLwtzB3DUxiCT8hRihtNZUtfZQXNdJcX3nkaP58iYzTs/RfHCAH+OTIzl9UhITkiOYkBLJxORIokIDBqWGkoZOnltXydtbqjDbnEzPjOahK6Zx1pQUAk0yXchwJuEvxAjQ3m1nX10HxfWd7KvrdAd+XSddVseRbdJjQpiYEsmSKclMSIlkQnIEWXFh3+tI/mgsdicf76nntU0HWVvSTKC/H+cWpHLdCVlMTY8e1NcS3iPhL8QwYnU4KW0wu4O+7uugr+v4umdNZLCJCcmRXDQjjfHJEUxIjmBcUgQRwYNzNN+fXdXtvFF0iHe31dDeYyctOoSfnjGeK2ZlEBce5NXXFoNPwl8IH7E6nOyr7WRndTu7qtvZWd3O/vpO7E53k02gvx95ieHMy4tjfHLEkaBPjgw+6knXwdTebee97dW8tukQu2s6CDT5ccbkZC4vzOCEvDj8BvlbhRg6Ev5CDAGL3Ulx3TeDvriu80hPm6iQAPLTorhxfi6TUt1NNjnxYQT4D327ucXuZFVxAx/sqOWTPfXYHC4mpUTym/Mmc/60VLlad5SQ8BdikPUO+p1VXx/R9w76qelR3Lwwl/y0KPLTokiPCRmyo/n+al5V3MhHO2tZubeebpuTuLBArpiVwWWFGUxJkwu1RhsJfyEGwOXSlDZ2sfVQG9sPtbG9qo19tV8HfXSo+4j+5vHDJ+gPs9idrNnvDvxP99RjtjmJCQ3g/GlpnDM1hTk5sZh88M1DDA0JfyG+g4YOy5Gg33aojR1V7Ud63EQEmZia4T6in5oWxZRhFPSHdVrsrNnfxCd76vh0bwNdVgfRoQGcW5DK2VNTmJsb55OmJjH0JPyF6IfZ6mBndTvbeoV9rWc8G5OfYmJKJBdMT2VaRgzTMqLIjQ8flidAD7V08+neelbubWBjeTN2pyY6NIAl+cmcPTWVE/Ik8I1Iwl8I3GPcFNd1sr2qjZ1V7WyvamN/fSeHRz7IjA1lVnYsBRnRTMuIZnJqJMEB/r4tuh9Ol2bboVY+3dvAyr317K/vAiAvIYwbTszhlIlJzMiMliYdg5PwF4bjcLooaexiR1U7Ozxhv7e288gwxDGhAeSnR3P65GSmZ0QzNT1q2PdjdzhdrC9rZtnOWj7eXU+z2Ya/n2J2diy/OjuDUycmkR0f5usyxTAi4S9GNZdLU9Fs9gS9O+x313TQY3cCEB5kIj8tiutPzGZqujvoh1s7fX96B/6K3fW0mG2EBfpz8sQkTpuUxKJxCUSFePfCLzFySfiLUaXFbKOoooXNB1uPdLPstLhPyAYH+DE5NYorZmcwNT2KqenR5MSFDct2+v70F/inTExiSX4Ki8cnDNvmKDG8SPiLEUtrzaGWHjZVtFBU2cKmilZKGtzt2wH+7hOy5xWkUpAeTX56FGMTw0dkO7fLpdlQ3swH22sk8MWgkfAXI4bTpdlb20FRRQubKlspqmihvsM92UhksInC7FgumpHGrOxY8tOiRnQgaq3ZXdPB+9treH9bDXUdFgl8Magk/MWw1WNzsu1Q25Gw31LZeqRPfWpUMHNz4yjMjmVWdgzjEiNGVPNNfw61dPPetmre3VZDSUMXJj/F4vEJ3Hv2RE6dmERIoAS+GBwS/mLY6LY5KKpoZX1ZMxvKmtlV3Y7dqVEKxidFcMH0VGZlx1KYHUtadIivyx00LWYbH+2o4d1tNWyubAVgVnYMv7tgCmfnpxATJmPpiMEn4S98xmJ3sqXSHfbrS5vZXtWG3akx+SkKMqK5aUEus7JjmJkZO2iTjwwXXVYHK/fW8962Gtbsb8Th0oxPiuBnZ47nvIJU0mNCfV2iGOUk/MWQsTqcbDvYdiTstx5qw+Zw4acgPz2aG+fnMi8vjsKsGMKCRt+vZofFzsq99SzbWcfq/Y3YHC5SooK5cUEOF0xLY2JKpK9LFAYy+v7CxLBhd7rYUdXG+tJm1pc1s7myFYvdhVIwOTWS6+ZlMS8vjlnZsV6fiMRX2nvsfLqnnmU7a/niQBM2p4vkyGB+MCeTJfkpzMyMGRXnKsTII+EvBo3Wml3VHXxZ0sT6smaKKlrotrkvppqQHMEVszI5IS+OOTlxo64Zp7e2bhsfewJ/bUkTdqcmLTqEa+dlcVZ+CtMzoiXwhc9J+IsBcbk026raWL6zlmU766hu6wFgTGI4F89IZ15eHHNyYof98AgD1WK28fHuOpbtqmNdSRMOlyY9JoQbTszhrPwUCtKjRsRVw8I4JPzFd+ZyabYcbGXZzjqW76qltt1CgL9iwdgE7jp1LIvHJ5AYEezrMr1Oa8260mZe3FDJx3vqcbo0mbGh3LQgl7PzU5iSFimBL4YtCX9xXJwuTVFFC8t21rJ8Vx0NnVYC/f1YOC6Bn54xnlMmJhlmHJn2Hjtvba7ixY2VlDWaiQkN4Mb5OZxXkMrkVAl8MTJI+It+OZwuvvIE/r921dPUZSXI5Mfi8QksyU/h5AmJo/ZEbV92VbfzwvpK3ttejcXuYkZmNH+5rIAl+Slyta0YcST8xb/ZUdXGq5sOsWJXHc1mG8EBfpw8IZGzpqRw0oREwkdhN8z+WOxOPtxRy4sbKtl2qI2QAH8unJ7GD+Zkyby2YkTz2l+xUurXwM1Ao2fVL7XWyzzP3QPcCDiBH2mtV3irDnF87E4Xy3fV8ezacrYcdIfcKRMTj4wjExponMAH93SNz6yr4JWvDtLWbScvIYxfnzuJi2amE2mgbzti9PL2X/T/aa0f6L1CKTUJuAKYDKQCnyqlxmmtnV6uRfShsdPKK18d5KWNldR3WMmOC+W/z5nEJYXGDLnSxi6eXFPG21uqcbhcnDE5mWvnZTM3N1ba8sWo4ovDufOBV7XWVqBcKVUCzAbW+6AWw9pZ1c4z68r5cHstNqeLheMSuO+ibBaNSzBkH/TNla08vrqUT/bWE+jvx2Wz0rlpfq7MfiVGLW+H/x1KqWuBIuAnWutWIA3Y0GubKs+6f6OUugW4BSAzM9PLpY5+dqeLf+2q49l1FWyubCU00J8rZmdw7bxsxiSG+7q8IedyaT7b18Dja0rZVNFKVEgAd540hmtPyCZ+lF+XIMSAwl8p9SmQ3MdT9wKPAb8FtOf+QeAGoK/DSt3Xv6+1fgJ4AqCwsLDPbcSxNXe5m3Ze2OBu2smMDeW/zpnEpQZt2rE6nLy3rYYn1pRR0tBFWnQI/3PuJC4rzBiVYwoJ0ZcB/aZrrU89nu2UUk8CH3oeVgEZvZ5OB2oGUofo267qdp5ZW8EHO2qwOVwsGBvPHy7MZ/H4RPwN2LTTYbHz8saDPLO2nPoOKxNTInnoimksyU8hYATO8CXEQHizt0+K1rrW8/BCYJdn+X3gZaXUX3Cf8B0LfOWtOozG7nSxYncdz66toMjTtHN5YQbXnZDFmMQIX5fnE/UdFp7+spyXNh6ky+rgxDFx3H9JAQvGxstJXGFY3vyO+2el1DTcTToVwFIArfVupdTrwB7AAdwuPX0Grsfm5Ln1FTy7toK6DguZsaH86uyJXFqYYZgrb7+tqrWbv68q5Y2iQzhdmiX5KSxdmEd+uvTPF8Jr4a+1vuYoz/0e+L23XttInC7NW1uq+MvH+6nrsHDimDh+d8EUTppgzKYd+GboA1xWmMHShXlkxskEKUIcJme3RiitNauKG7lv+T6K6zspyIjmoSumMSc3ztel+cy3Q//yWRnctnjMqJryUYjBIuE/Au2oauOPy/axvqyZrLhQHr1qBkvykw3bfi2hL8R3J+E/ghxs7ub+j4v5YHsNsWGB/Oa8yVw5O5NAkzF7qkjoC/H9SfiPAC1mG498doAXN1Ti76e48+Qx3LIw11AjavYmoS/EwEn4D2MWu5On15bz2OelmG0OLp+VwX+eOo6kyNE/UUpfmrqsPLLyAC9/dRCQ0BdiICT8h6HDPXj+75P91LZbOHViIj8/cwJjk4zZT7/b5uCpL8r5x+pSLA4XlxVmcMfJEvpCDISE/zDSVw+ev15u3B48DqeL14uq+Oun+2notHLG5CR+esYEQ45DJMRgk/AfJiqazPzynZ2sK20m2+A9eLTWfLKnnj/9ax+ljWZmZsXw9x/MoDA71telCTFqSPj7mNaa14sO8ZsP9mDyU4bvwbPlYCt/XLaXTRWt5CaE8fg1Mzl9UpIhPwSF8CYJfx9qNdu45+2d/Gt3HfNy43jwsgJSDdqOXdbYxf0rilm+q46EiCB+f+EULi/MwCQDrgnhFRL+PvLFgUbufmM7LWYb95w1gZsX5BpyEpVum4O/fnqAp74sJ9jkx/87bRw3zs+RoZWF8DL5CxtiFruT+1cU89SX5YxJDOep62YZdiLwVcUN/OrdXVS19nDl7Ax+cvp4mURFiCEi4T+Eius6uevVreyr6+TaeVncc9ZEQgL9fV3WkGvstPLbD/fw/vYaxiSG88at85glJ3OFGFIS/kPA5dI8u66C+/61j8hgE8/8xyxOmpDo67KG3OGT239Yto8em5MfnzqOWxfnEmQy3gegEL4m4e9lDR0W7n5zB2v2N3LKhET+dMlUQzZtlDZ28cu3d7KxvIXZObH84cJ86a8vhA9J+HvRJ3vq+dmb2+mxO/ndBVP4wZxMw3VZtDqc/GNVGY9+XkJwgB9/ujifS2dmGPLkthDDiYS/lzy7tpxff7CHKWmR/PXy6YY8yt1T08Fdr27lQEMX5xak8l/nTCQxwpjjEgkx3Ej4e8Gjn5dw/4piTp+UxCNXTTdcm7bWmhc3HuS3H+4hOiTAsOc4hBjOJPwHkdaaBz4u5tHPSzl/WioPXFpAgMEuUmrvsXPP2ztYtrOOReMS+MtlBcQZ8ByHEMOdhP8gcbk0//vhHp5dV8GVszP43QX5hptDd9uhNu54eQt17RZDX7gmxEgg4T8InC7NPW/v4PWiKm6cn8Ovzp5oqBO7Wmue+rKc+5bvIykymNeWzmNmVoyvyxJCHIWE/wDZnS5+/No2PtxRy49OGcuPTx1rqOBvMdu4+43tfLavgdMnJXH/JQVEhRpzhjEhRhIJ/wGw2J3c8fIWPt3bwD1nTWDpojxflzSkvipv4UevbKXFbOPX507iuhOyDfXBJ8RIJuH/PXXbHNz8fBFrS5r57fmTuWZetq9LGlLLd9ZyxytbyYgJ4e0fnmDY8YmEGKkk/L+HDoudG57ZxJaDrTxwaQGXzEz3dUlD6uPdddz5ylamZUTz7PWzDDuRvBAjmYT/d9RitnHt0xspruvkb1fNYEl+iq9LGlKf7avn9pe3MDktSoJfiBFMwv87aO+xc8UT66ls7uaJawoNd+HS6v2N3PrCFsYnR/D8DbMl+IUYwST8j5PTpfnRK1spazTz3A2zOXFMvK9LGlJrS5q45fki8hLDefHGOUSFSPALMZJJ+B+nPy7by+r9jfzxonzDBf+GsmZufG4T2XFhvHTTHKJDA31dkhBigAY09oBS6lKl1G6llEspVfit5+5RSpUopYqVUmf0Wj9TKbXT89zDagT0DXyj6BD//LKc6+ZlceXsTF+XM6R2Vbdzw7ObSI8J5aWb5xAbJsEvxGgw0IFndgEXAWt6r1RKTQKuACYDZwJ/V0odHt3sMeAWYKznduYAa/CqzZUt3PvOLk4cE8d/nTPJ1+UMqeYuK0tf2ExUSAAv3zTHkPMQCDFaDSj8tdZ7tdbFfTx1PvCq1tqqtS4HSoDZSqkUIFJrvV5rrYHngQsGUoM31bT1sPSFLaREB/PoVTMwGWiQNrvTxR0vb6Wxy8rj18wkMVKGYhZiNPFWmqUBh3o9rvKsS/Msf3t9n5RStyilipRSRY2NjV4ptD89Nie3vFCExe7kn9cWGq6d+w/L9rK+rJk/XpjP1PRoX5cjhBhkxzzhq5T6FEju46l7tdbv9fdjfazTR1nfJ631E8ATAIWFhf1uN9i01tz95nZ213Tw9HWzGJsUMVQvPSy8tbmKZ9ZWcP2J2VxssAvYhDCKY4a/1vrU7/HvVgEZvR6nAzWe9el9rB9W/vZZCR/tqOWesyYYri//jqo27nlnJ/Ny4/jlkom+LkcI4SXeavZ5H7hCKRWklMrBfWL3K611LdCplJrr6eVzLdDftwef+NeuOh78ZD8XTU/jloW5vi5nSDV5TvAmhAfxt6umG24iGiGMZKBdPS9USlUB84CPlFIrALTWu4HXgT3Av4DbtdZOz4/dBvwT90ngUmD5QGoYTHtrO/h/r29jWkY0f7go33AjVD6wopjmLhuPXzNTZt8SYpQb0EVeWut3gHf6ee73wO/7WF8ETBnI63qDxe7khy9tISLYxBPXzCQ4wFjz7lY2m3ljcxXXzM2SETqFMAC5wtfjyTVllDeZeeHG2Ybs1vjIZyWY/BS3LTbWnARCGJU06gJVrd08uqqEs6Yks2Bsgq/LGXLlTWbe3lLF1XOzSDLgB58QRiThD/z2wz0oFL8y2BW8hz288gCBJj9uNdhMZEIYmeHDf/X+RlbsrueOk8eQFh3i63KGXJfVwXvbqrlqdhYJEXKSVwijMHT4Wx1OfvP+bnLiw7hpQY6vy/EJq92JS0N2fKivSxFCDCFDh/9TX5ZT1mTmf86dRJDJWL17hBDGZtjwr2nr4ZGVJZw+KYnF4411Fa8QQhg2/H//0V5cWhtumGYhhACDhv/akiY+2lnLDxePISPW2G3dgSb3r0Bbt93HlQghhpLhwt/mcPE/7+8mMzaUpYuMNXZPXyKCA5iYEsnakiZflyKEGEKGC/9n15VT0tDF/5w7yXBDOPRn4bh4Nle20mV1+LoUIcQQMVT4250uHvmshJMnJHLKxCRflzNsLBqbgMOl2VDa7OtShBBDxFDhv/VgG50WB5fPyjj2xgYyMzuGkAB/1hwY2tnShBC+Y6jw/+JAI/5+inl5cb4uZVgJMvkzNzeWz4sbcLqGbMI0IYQPGSr81xxoYlpGNJHBAb4uZdi5eGY6h1p6eGtL1bE3FkKMeIYJ/7ZuGzuq2pg/Jt7XpQxLZ+enMD0zmgdWFGOWE79CjHqGCf91pc1o7e7ZIv6dUopfnT2Rhk4rT6wp83U5QggvM0z4f3GgkYggEwXp0b4uZdiamRXL2VNTeGJNGXXtFl+XI4TwIkOEv9aaNfubmJcXh0kmJT+qX5w5AadL88DHxb4uRQjhRYZIwormbqrbelgwznizdH1XGbGhXH9iNm9tqWJXdbuvyxFCeIkhwv8LT//1BXKy97j88KQxxIYG8tM3d2CxO31djhDCCwwR/mv2N5ERG0JWnLEHcTteUSEBPHBpAXtrO/jth3t8XY4QwgtGffhrrdlQ1sz8MQkopXxdzohx0oREli7M5aWNB/lwR42vyxFCDLJRH/4dPQ66rA7yEsJ8XcqIc/cZ45mRGc0v3tpJRZPZ1+UIIQbRqA//ZrMVgNiwQB9XMvIE+PvxyFUz8PdT3P7yFmn/F2IUGfXh39ptAyT8v6+06BAevLSA3TUd/GHZXl+XI4QYJKM+/Ju73OEfFxbk40pGrlMnJXHT/ByeX1/Jsp21vi5HCDEIRn34t5jd4R8TJoO5DcTPzpzAtIxofv7mDiqbpf1fiJFu9Id/txz5D4ZAkx9/u2o6SsHNzxfRYZE5f4UYyQYU/kqpS5VSu5VSLqVUYa/12UqpHqXUNs/tH72em6mU2qmUKlFKPay83P+ypctGSIA/IYEyZeNApceE8o+rZ1LWaOa2Fzdjc7h8XZIQ4nsa6JH/LuAiYE0fz5Vqrad5brf2Wv8YcAsw1nM7c4A1HFWL2SYnewfRCWPiue/iqawtaebed3aitUz+IsRIZBrID2ut9wLHffGUUioFiNRar/c8fh64AFg+kDqOptlsIy5cwn8wXTIznUMt3Ty08gBZcaHccfJYX5ckhPiOvNnmn6OU2qqUWq2UWuBZlwb0niqqyrOuT0qpW5RSRUqposbG7ze/bGu3jZhQCf/B9p+njuWi6Wk88PF+3t1a7etyhBDf0TGP/JVSnwLJfTx1r9b6vX5+rBbI1Fo3K6VmAu8qpSYDfX1F6LfdQGv9BPAEQGFh4fdqX4gINtHYaf0+PyqOQinFfRdPpaa9h5+9uYPkqGDm5srcyEKMFMc88tdan6q1ntLHrb/gR2tt1Vo3e5Y3A6XAONxH+um9Nk0HvDpwzOzsOPbWddDeLb1TBlugyY/Hry4kIzaEpS9sprSxy9clCSGOk1eafZRSCUopf89yLu4Tu2Va61qgUyk119PL51qg3w+RwTA3Nxat4auKFm++jGFFhQbw7PWzCfBXXP/MJpq65FuWECPBQLt6XqiUqgLmAR8ppVZ4nloI7FBKbQfeBG7VWh9O39uAfwIluL8ReO1kL0BBRjSBJj82ljV782UMLSM2lH9eN4uGTgs3PVckYwAJMQIMKPy11u9ordO11kFa6ySt9Rme9W9prSdrrQu01jO01h/0+pkiT7NRntb6Du3lvoLBAf7MyIxmQ7mEvzdNy4jmoSums72qjTtf2YrdKdcACDGcjforfAHm5MSxp6aD9h5p9/emMyYn87/nTeaTPfXc+bJ8AAgxnBki/OfmxuHSUCTt/l53zbxs/vucSfxrdx0/km8AQgxbhgj/6ZnRBPr7sbFcwn8o3DA/h1+dPZHlu+r4z1e34ZAPACGGnQFd4TtSBAf4My0zWk76DqGbFuQC8LuP9qIU/PXyaZj8DXGsIcSIYJi/xrk5seysbqdTRqMcMjctyOXeJRP5cEctP359u3wDEGIYMU74e9r915fK0f9QunlhLvecNYEPttfwkze243TJQHBCDAeGaPYBmJUTS0xoAO9vr+H0yX2NViG8ZemiPFwa/vSvfSjgwcum4e/n1ZG8hRDHYJjwD/D349yCVF7bdIhOi52IYJnZayjdtjgPl9bcv6IYP6W4/9IC+QAQwocM0+wDcOH0NKwOF8t31fm6FEO6/aQx3H36ON7eWs1tL26my+rwdUlCGJahwn9aRjQ58WEyBLEP3XHyWP77nEms3NfA+X/7kpIGGQxOCF8wVPgrpbhgWhrry5qpbe/xdTmGdcP8HF64cTZt3XYueHQtK3bLNzEhhpqhwh/ggumpaA3vbfPqSNLiGE7Ii+eDO+eTlxDG0hc2c/+KfdITSIghZLjwz4oLY0ZmtDT9DAOp0SG8tnQelxdm8Ojnpdzw7Cbaum2+LksIQzBc+ANcOCOdfXWd7Knp8HUphhcc4M+fLpnKHy7MZ11pE+f+7Ut5X4QYAoYM/3PyUwjwV7y7TY7+h4ur5mTy2tJ52BwuLnpsLe/JeyOEVxky/GPCAlk8PpH3tlVLO/MwMiMzhg/unM/UtGjuenUbP3l9Oy1maQYSwhsMGf4AF89Io77DysfS02RYSYwI5qWb53D7SXm8t62akx9cxWubDuKSD2khBpVhw/+0ScnkxofxyGcleHkyMfEdBfj78dMzJrDsrgWMTQzn52/t5PIn1lNc1+nr0oQYNQwb/v5+ih+eNIY9tR2s3Nvg63JEH8YlRfDaLfP48yVTKWno4uyHv+C+5fvotsmVwUIMlGHDH+D8aalkxIbwyGcH5Oh/mPLzU1xWmMHKnyzmohlp/GN1Kaf9ZQ0r99b7ujQhRjRDh3+Avx+3Lx7D9qp2Vu9v9HU54ihiwwL58yUFvHbLXEID/bnxuSKWvlAkV2oL8T0ZOvwBLpqRTlp0iLT9jxBzcuP46EcL+NmZ41m9v5FTH1zN46tLsdidvi5NiBHF8OEfaPLj1kW5bK5slYleRohAkx8/XDyGT368iDm5cfxx+T4W37+KlzcelAnjxajicmmvzYCnRsrRbmFhoS4qKvLKv22xO1n458/JiQ/jtaXzvPIawnvWlTZx/4pith5sIzsulB+fNo5zp6biJ/MFiGHO5dLsqe1gU0ULdR0Wmjpt7K3tYE/tN69yr7jv7O/9GkqpzVrrwm+vN8xkLkcTHODPrYvy+N8P9/BVeQuzc2J9XZL4Dk7Ii+ft2+JYubeBBz4u5q5Xt/HYqlLuPn08p0xMRCn5EBDDR32HhS8ONPHFgUa+PNBE8zEuZPzZmeO9Uocc+Xv02Jws+PNnTEyJ5IUb53jtdYR3uVyaD3bU8JdP9lPZ3M2MzGh+esYE5uXF+bo0YVA2h4sNZc28XnSID3fU9rtdkMmP2TmxnJAXz4lj4picGjUos93Jkf8xhAT6c/OCXP64fB9bDrYyIzPG1yWJ78HPT3H+tDSW5KfwRlEVD688wJVPbmDB2HjuPn08BRnRvi5RjFJOl6amrYfK5m4qms28u7WaosrWfrefmRXDuKRwxiRGMCklkumZ0QQH+A9ZvXLk34vZ6mD+nz5jano0z14/S5oLRgGL3cmLGyp59PMSWrvtnD4piStnZzJ/bDwB/obv7yCOQWuN2eakuctKU5eVpi6b+77TRrP563WNnVaqWruxO/vO06y4UH4wJ5P8tGjGJoUTFxY4ZPkiR/7HISzIxG2L8/jDsn28vaWai2em+7okMUDBAf7ctCCXy2dl8PSXFTy9tpyP99QTExrA2VNTOK8gjcKsGDk5PMppremxO+myOujosdPabafVbKOtx05bt43Wbvd9W7edVs/94WWro+/eNtGhAcSFBRIc4E95k/kbz+XGh3HrojzOyk8mIjhgKHbxOxvQkb9S6n7gXMAGlALXa63bPM/dA9wIOIEfaa1XeNbPBJ4FQoBlwF36OIoYiiN/cH91u/KJDeyuaWfZXQvIigvz+muKoWN1OFmzv4n3t9fwyZ46LHYXqVHBnDstlfML0piYEiHf+IYBrTUWuwuzzYHZ6sBsddJtc9BlddBtc4d4t9WB2eb0PO+g0/r1tl1W97bmXvdHGxvQ5KeIDg0kOjSAmNAA93JIADFhgcSGBRIfHkR8+OH7IKJDAyiqaOXNzYf4127379GYxHAunpHOBdNTSYkKGbr/rGPo78h/oOF/OvCZ1tqhlPoTgNb650qpScArwGwgFfgUGKe1diqlvgLuAjbgDv+HtdbLj/VaQxX+ANVtPZz51zXkJYTzxq3zpHlglDJbHXyyp573tlXzxYEmHC7NmMRwzi9I5bxpqfLBP0BWh5P2HjsdPXbae9+67bT3OOiwfHP94e26LA7MtqOHdW8mP0V4sImwQBPhQSbCgvwJDw4gPMifsEATYUEmIoLd92FBJqJCPAEf4gn7sEDCAv2P60O/stnMm5ureGtzFTXtFiKDTZw3LZVLZmZQkB41LA8cvBL+33qBC4FLtNY/8Bz1o7X+o+e5FcCvgQrgc631BM/6K4HFWuulx/r3hzL8AT7cUcMdL2/lzpPH8JPTvdPVSgwfLWYby3bW8v62Gr6qaAGgICOa8wpSmZsby9jECAJNxjwIsNid32gKae/xLPd867FnXbunGaXnGFddhwX6ExUSQKTnFuW5RQS7Qzw00ER4kD+hgYeD29+zzkRooL/7PsifIJN3T5L22Jws31XL60WH2FDWgp+CBWMTuLQwnVMnJg3pSdrvYyja/G8AXvMsp+E+sj+syrPO7ln+9vo+KaVuAW4ByMzMHMRSj+2cqamsKm7k0c9LmD8mnjm50lVwNIsNC+TquVlcPTeL6rYePtxew3vbavjth3sACPBXjEmMYHJqJJNSIpmUGsnElEiiQoZne+7RWB1OWsw2mrtsNJtttJitNHfZaOr6ernZ7D6h2dJlw2zrP8QD/f2ICg0gOiSA6NAA0qJDmJwaSbQnyKNDvxnsh2+RIQHD/ht1SUMXz64r592tNXRZHWTFhfLTM8Zz0Yy0YdWs830dM/yVUp8CyX08da/W+j3PNvcCDuClwz/Wx/b6KOv7pLV+AngC3Ef+x6p1sP36vMkUVbTw49e2sfyuhUSFjrw/dPHdpUWHsHRRHksX5VHZbGZndTu7azrYU9PBquJG3tz89fFLRmyI+8MgJYpJqZFMTo0kJSrYJ1//LXYnjZ1W6jos1HdYqO+w0tBhOfK4ocNKY5eVTkvfQ2IH+CviwoKIC3e3c+fEhxEXFkhseCCxnvbwqJDD9+5gDwk4vuaSkUJrzRcHmnh6bTmrihsJNPlxztQULi/MYHZO7Kja12OGv9b61KM9r5S6DjgHOKXXidsqIKPXZulAjWd9eh/rh6XwIBMPXTGdix9bxy/f3cnfrpw+qt58cWxZcWFkxYVxztTUI+saOi3sqXFfgr+7poO9NR18vKeew7/9kcEm4sODjhzxuu9N/370G/z1UbHWYHE4sdidWB0uLHYnFrsLq8N933u91e7E4nDRarZR32mlvt1CfaeFtm77v9UfaPIjOTKYpMggJqZGstBz4jLWE/K9lyOCTIb9/bbYnbyztZqnvyznQEMX8eFB/L/TxnHVnEziw4N8XZ5XDKjZRyl1JvBzYJHWurvXU+8DLyul/oL7hO9Y4CvPCd9OpdRcYCNwLfDIQGrwtoKMaH582jjuX1HM4nEJXFqYcewfEqNaYkQwieODWTw+8cg6s9XBvjr3t4Pi+k5au+2eLoU2KprNR05mDtZslMEBfkSHBJIUGURWXCizc2JJigwiMTLYE/buwI8KCTBsoB+PunYLL2yo4OWNB2nttjMpJZIHLy3gnIIUr59L8LWBtvn/DQgCPvH8gm3QWt+qtd6tlHod2IO7Oeh2rfXhhsPb+Lqr53LPbVi7dVEeXxxo5H/e382s7Fiy46UXiPimsCATM7NimZnV/7hQWmu6rI5evVvcyx0WO/5KERTgR7DJn+AA/17LfgQdvg9w3wf6+0mgD9COqjae+rKcj3bU4tSa0yclcf2JOcwZZU07RyNX+B6n2vYezvzrF2THhfLmbScM+5NVQohv0lqzen8jj60qZWN5C+FBJi6flcF187LJjAv1dXleI1f4DlBKVAj3XZTPbS9t4S+f7OfnZ07wdUlCiOPgcLr4aGct/1hdxt7aDlKigvnV2RO5fFbGsL36dihI+H8HZ+WncOXsDB5bVUpmbChXzh7a7qdCiONnsTt5Y3MVT6wp5VBLD2MSw3ng0gLOK0g17DUbvUn4f0e/OW8Kde0WfvnOTkIC/Llger+XKQghfMBsdfDSxkqeWFNOU5eVaRnR/NfZkzh1YpKM4dSLhP93FGjy47GrZ3L9M5v4yRvbCQ7w58wpfV0GIYQYSh0WO8+vq+CpL8tp7bYzf0w8t580nbm5xjmJ+11I+H8PwQH+/PO6Qq55aiN3vrKFJ68t/Ea3PyHE0GnrtvH02gqeXVtOh8XBSeMTuPOUsTInxzFI+H9PYUEmnrl+Nlc9uYGlL2zm2etny2xRQgyhpi4rT31ZzgvrK+myOjh9UhJ3njyW/PQoX5c2IkhXzwFqMdu4/PH1VLf18OJNc+RoQwgvq23v4fHVZby66SBWh4sl+SnccdIYJqZE+rq0Ycnro3p623ANf4CGDguXPb6eFrONV26Zy+RUOfIQYrBVNJn5x+pS3tpShUvDBdPSuG1xHmMSw31d2rAm4e9l1W09XPaP9fTYnby+dC5jEiN8XZIQo0JxXSd/X1XCB9trMPn7cXlhBrcszCUjdvRemDWYJPyHQEWTmUsfX48C3rh1nkwGIsQAbD/UxqOfl/DxnnpCA/25em4WN83PITEy2NeljSgS/kNkf30nlz++ntBAE2/cOo/U6JE/7rcQQ8Xp0ny+r4Fn1pWztqSZyGAT/3FiDtefkE1MWKCvyxuRJPyH0K7qdq58cgPhQSaevLaQKWlyDkCIo+mw2HmjqIrn1lVwsKWb5Mhgrjshm6vnZhp6CIbBIOE/xHbXtHPzc0W0dNt48NJpnD01xdclCTHslDR08fz6Ct7cXEW3zUlhVgz/cWI2Z0xOlsETB4kM7DbEJqdG8d4d81n6QhG3v7yF/fVjueuUsXJ5uTA8l8s9uuYz6ypYs7+RQH8/zi1I5T9OyJY++kNIwt+LEiKCeOWWudz7zi4eWnmA/fWdPHhZAaGB8t8ujKfDYuetzVU8v76S8iYziRFB/OS0cVw5imfLGs4khbwsyOTP/ZdMZUJyBH9YtpfKx7p58rpC0uREsDCI4rpOnl9fwTtbq+m2OZmeGc3DV07nzMnJMrqmD0n4DwGlFDctyCUvMZwfvbyV8//2Jf+4eiaF2f3P+iTESGZ3uvhkTz3PratgY3kLgSY/zi9I5dp50rQzXMgJ3yFW0tDJTc8VUd3Ww+8vzOcymRNYjCINnRZe/eoQL288SF2HhfSYEK6Zm8VlhRnSVdNH5ITvMDEmMYJ3bz+RO17eys/e3MH+uk7uWTIRfzkRLEYorTVbDrbx/PoKlu2sxe7ULByXwO8umMJJExLld3uYkvD3gejQQJ69fha/+2gv//yynAMNXTx0xTSiQ+XISIwcde0W3t1WzVubqzjQ0EVEkImr52ZxzdwschNkvJ3hTpp9fOyVrw7yX+/uIjYskD9elM8pE5N8XZIQ/eqxOVmxu463tlSxtqQJl4aZWTFcPCOd86elEhYkx5PDjTT7DFNXzs4kPy2Ku9/Yzo3PFXHxjHT++9xJRIXIVY1ieHC5NF9VtPD2liqW7ayjy+ogLTqEO04aw4Uz0smJlzGsRiIJ/2FgSloU798xn0c+O8DfV5XyZUkj9100lZMmyOxgwncqmsy8vaWKt7dWU9XaQ1igP0vyU7h4Zjqzs2PlgsURTpp9hpmdVe385I1t7K/v4tKZ6fzqHPkWIIZOXbuFj3bW8uGOGrYebEMpmD8mnotnpHPG5GRCAv19XaL4jmRsnxHE6nDy8MoD/GN1GQnhQdx3cb7MESy8pqHDwvJddXy4o4ZNFa0ATEqJ5NyCVC6cnkZylAyhPJJJ+I9A2w+1cfcb2znQ0MXlhRnce85EImWEQzEImrqsLN9Vx0c7athY3oLWMD4pgrOnpnD21BTypLfOqCHhP0JZ7E4eWnmAx1eXkhQZzJ8unsrCcQm+LkuMQPUdFj7b18BHO2pZV+ruqZOXEMY5U1M5Z2oKY5Nk9rnRSMJ/hNt2qI2fvL6N0kYzV87O4JdLJso45+KorA4nRRWtrNnfyOr9jeyr6wQgOy7UHfgFKYxPikApOXE7mkn4jwIWu5P/+3Q/T64pIyY0kJsX5nLN3CzpWy0A95W2Fc3dR8J+fWkzPXYnAf6KwqxYFo1PYNG4BCYkS+AbiYT/KLL9UBsPfFzMFweaiA0L5OYFuVw7Tz4EjMhsdbCutJnV+xtYs7+Jgy3dAGTFhbJoXAILxyYwLy9OfjcMzCvhr5S6HzgXsAGlwPVa6zalVDawFyj2bLpBa32r52dmAs8CIcAy4C59HEVI+P+7zZWtPLTyAGv2NxITGsDNC3O5dl424fKHPmpprTnQ0MWq4gZWFTeyqaIFu1MTGujPvNw4Fo13B362XHglPLwV/qcDn2mtHUqpPwForX/uCf8PtdZT+viZr4C7gA24w/9hrfXyY72WhH//thxs5eGVB1hV3Eh0aAA3L8jluhPkQ2C06LTYWVvSzOr9jawubqCm3QLAuKRwFo9PZNG4BAqzYwgySR988e+83uyjlLoQuERr/YP+wl8plQJ8rrWe4Hl8JbBYa730WP++hP+xbTvUxkOf7udzz4fATfNzuO6EbDkx7GMdFvtxddG1OpwcaummrNFMeZOZimYzB+q72HaoDYdLEx5k4sQxcSwen8jCcQkyIZA4LkMxts8NwGu9HucopbYCHcCvtNZfAGlAVa9tqjzr+qSUugW4BSAzM3MQSx2dpmVE88z1s9l+qI2HVx7ggY/38+QX5Zw/LZUzpyQzOzsWk0yKPaQcThfT//cTgk1+BJj8OHyaVSnVa9l932K24ep1LBYXFkhOfBg3Lshh8bhEZmbFyMxXYtAc88hfKfUpkNzHU/dqrd/zbHMvUAhcpLXWSqkgIFxr3exp438XmAyMB/6otT7V83MLgJ9prc89VqFy5P/d7ahq4/HVZazcV4/F7iI2LJDTJyVx5pRkTsiLlyAZAha7kxc3VFLd1oPTk+xag6b3svs+ISKInPhQcuLDyYkLIypUvrGJgfNas49S6jrgVuAUrXV3P9usAu4GqpFmnyHXbXOwqriR5bvq+GxvPWabk4hgE6dNdH8QLByXQHCAtBePdDaHi9LGLoID/GWkTXGEV5p9lFJnAj8HFvUOfqVUAtCitXYqpXKBsUCZ1rpFKdWplJoLbASuBR4ZSA3i2EIDTSzJT2FJfgoWu5O1JU0s31XHJ3vqeXtrNaGB/pw0IZEzJydz8oRE6RY4zGmtqWm3UFzXwd7aTorrOtlX10FZoxmHS3P13Ex+d0G+r8sUw9xA/8r/BgQBn3guGjncpXMh8L9KKQfgBG7VWrd4fuY2vu7qudxzE0MkOMCfUyYmccrEJOxOFxvKmlm+q46Pd9fx0Y5aAk1+LBybwGmTEinMjiU3PkwuCPIBrTWNndYjJ37Lm7qp8CxXNJux2F1Htk2LDmFCcgSnTkxiQkokBTJBujgOcpGXAMDp0myubGX5rlpW7Ko70p0wJjSAmVkxzMyKZWZWDFPTo6SJaBC1mG2UN3V9o4dPeVM3lc1mum3OI9sF+CsyYkPJiQsjOz6MnPgwJiRHMC45Qgb7E0clV/iK46a1prTRzObKFjZXtlJU2UpZoxlwh9Dk1CgKs2IozI5hRlYMiREy5O/RWOxOd6g3milrMlPWaKasqYvyJjNt3fYj25n83AGfHRd6JOCz4sLIiQsjNTpYemqJ70XCXwxIi9nGFs8HwebKFrZXtWNzuJseMmNDmZkVw/TMaHLiw8iMDSU1OoQAA4WV1pq6DgulDWZKG7soa+w6EvQ17T30/jNLigwiNz6cnIQwcuPDyE0IIyc+nPQYY/2fiaEh4S8Glc3hYldNO5srWo98O2jqsh553k9BSlQImbGh7ltcKOkxXz+ODQsckecSDh/FHw55d9C7l3s304QHmTyhHvaNoM+JD5MT6mJISfgLrzp85HuwuZuDLd0cavHct/ZwsKWbxk7rN7YPC/QnIzaUjNhQ4sMDiQgOICLIRESwyb3c6z7yyGPTgJs+tNZ025yYrQ66et3MVve6TqsDs+fWZXXQZXFgtjnotDiobO7mUGv3N47i06JDyE0IIy8hnLzEcPI8y4kRQSPyw02MPkNxha8wMKUUKVEhpESFMCc37t+e77Y5qGrtOfKhcOQDormbbYfa6LTYv9GDpT8hAf5EBJsID3b/6moNLq3dN5c73J1a49LuZdfh510ap0vTbXdyPMc7SkF4oImwIPdrhQWZmJoexYXT046EfE58GKGB8ickRib5zRVDIjTQxLikCMYdZbYom8NFl9VBp8VOp8VBh+fefbN/897qAMBfKfwU+CmF6rXs59drWeF5ThEW5E94kCfUe927l93PhQebCAnwlyN3MapJ+IthI9DkR6wpkNiwQF+XIsSoJ10LhBDCgCT8hRDCgCT8hRDCgCT8hRDCgCT8hRDCgCT8hRDCgCT8hRDCgCT8hRDCgEbM2D5KqUag0td1eMQDTb4uwoeMvP9G3ncw9v6P1H3P0lonfHvliAn/4UQpVdTXQElGYeT9N/K+g7H3f7TtuzT7CCGEAUn4CyGEAUn4fz9P+LoAHzPy/ht538HY+z+q9l3a/IUQwoDkyF8IIQxIwl8IIQxIwr8PSqmnlVINSqldvdbFKqU+UUod8NzH9HruHqVUiVKqWCl1hm+qHhz97PuvlVLVSqltntuSXs+Npn3PUEp9rpTaq5TarZS6y7PeKO99f/s/6t9/pVSwUuorpdR2z77/xrN+9L73Wmu5fesGLARmALt6rfsz8AvP8i+AP3mWJwHbgSAgBygF/H29D4O8778G7u5j29G27ynADM9yBLDfs49Gee/72/9R//4DCgj3LAcAG4G5o/m9lyP/Pmit1wAt31p9PvCcZ/k54IJe61/VWlu11uVACTB7KOr0hn72vT+jbd9rtdZbPMudwF4gDeO89/3tf39Gzf5rty7PwwDPTTOK33sJ/+OXpLWuBfcfCZDoWZ8GHOq1XRVH/4MZqe5QSu3wNAsd/uo7avddKZUNTMd9BGi49/5b+w8GeP+VUv5KqW1AA/CJ1npUv/cS/gOn+lg32vrPPgbkAdOAWuBBz/pRue9KqXDgLeA/tdYdR9u0j3Wjcf8N8f5rrZ1a62lAOjBbKTXlKJuP+H2X8D9+9UqpFADPfYNnfRWQ0Wu7dKBmiGvzKq11vecPwwU8yddfb0fdviulAnAH30ta67c9qw3z3ve1/0Z6/wG01m3AKuBMRvF7L+F//N4HrvMsXwe812v9FUqpIKVUDjAW+MoH9XnN4V9+jwuBwz2BRtW+K6UU8BSwV2v9l15PGeK972//jfD+K6USlFLRnuUQ4FRgH6P5vff1GefheANewf311o77E/5GIA5YCRzw3Mf22v5e3Gf7i4GzfF2/F/b9BWAnsAP3L33KKN33+bi/uu8AtnluSwz03ve3/6P+/QemAls9+7gL+G/P+lH73svwDkIIYUDS7COEEAYk4S+EEAYk4S+EEAYk4S+EEAYk4S+EEAYk4S+EEAYk4S+EEAb0/wErIDTfV3GhGQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "print('There are ', len(npy_data)+1, ' images in selected category', categories[category])\n",
    "print(\"Select an index value from 0 to\", len(npy_data), \"to display and compare bitmap and raw images:\")\n",
    "#index=input()\n",
    "index = '0'\n",
    "plot_bitmap()\n",
    "plt.show()\n",
    "plot_raw()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "# TRAIN TEST SPLIT\n",
    "filepath = paths_dict['input_path']\n",
    "x_train, x_test, y_train, y_test = split_data(filepath, categories[category], shuffle=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DRAW(\n",
       "  (encoder): LSTMCell(306, 256)\n",
       "  (mu_linear): Linear(in_features=256, out_features=10, bias=True)\n",
       "  (sigma_linear): Linear(in_features=256, out_features=10, bias=True)\n",
       "  (decoder): LSTMCell(10, 256)\n",
       "  (dec_linear): Linear(in_features=256, out_features=5, bias=True)\n",
       "  (dec_w_linear): Linear(in_features=256, out_features=25, bias=True)\n",
       "  (sigmoid): Sigmoid()\n",
       ")"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "DRAW(model_path, category)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_data = Dataset(x_train)\n",
    "test_data = Dataset(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = DRAW(category)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "training...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\\src\\draw_model.py:211: UserWarning: torch.nn.utils.clip_grad_norm is now deprecated in favor of torch.nn.utils.clip_grad_norm_.\n",
      "  torch.nn.utils.clip_grad_norm(self.parameters(), self.clip)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Phase: train | Epoch: 0 | Count: 100 \\ Start Time: 03:17:33 | Loss: 222\n",
      "Phase: train | Epoch: 0 | Count: 200 \\ Start Time: 03:18:35 | Loss: 191\n",
      "Phase: train | Epoch: 0 | Count: 300 \\ Start Time: 03:19:38 | Loss: 186\n",
      "Phase: train | Epoch: 0 | Count: 400 \\ Start Time: 03:20:42 | Loss: 184\n",
      "Phase: train | Epoch: 0 | Count: 500 \\ Start Time: 03:21:49 | Loss: 181\n",
      "Phase: train | Epoch: 0 | Count: 600 \\ Start Time: 03:23:02 | Loss: 173\n",
      "Phase: train | Epoch: 0 | Count: 700 \\ Start Time: 03:24:17 | Loss: 163\n",
      "Phase: train | Epoch: 0 | Count: 800 \\ Start Time: 03:25:28 | Loss: 153\n",
      "Phase: train | Epoch: 0 | Count: 900 \\ Start Time: 03:26:29 | Loss: 150\n",
      "Phase: train | Epoch: 0 | Count: 1000 \\ Start Time: 03:27:31 | Loss: 144\n",
      "Phase: train | Epoch: 0 | Count: 1100 \\ Start Time: 03:28:37 | Loss: 140\n",
      "Phase: train | Epoch: 0 | Count: 1200 \\ Start Time: 03:29:39 | Loss: 136\n",
      "Phase: train | Epoch: 0 | Count: 1300 \\ Start Time: 03:30:40 | Loss: 137\n",
      "Phase: train | Epoch: 0 | Count: 1400 \\ Start Time: 03:31:42 | Loss: 132\n",
      "Phase: train | Epoch: 0 | Count: 1500 \\ Start Time: 03:32:45 | Loss: 129\n",
      "Phase: train | Epoch: 1 | Count: 1600 \\ Start Time: 03:33:46 | Loss: 129\n",
      "Phase: train | Epoch: 1 | Count: 1700 \\ Start Time: 03:34:48 | Loss: 129\n",
      "Phase: train | Epoch: 1 | Count: 1800 \\ Start Time: 03:35:53 | Loss: 124\n",
      "Phase: train | Epoch: 1 | Count: 1900 \\ Start Time: 03:36:57 | Loss: 125\n",
      "Phase: train | Epoch: 1 | Count: 2000 \\ Start Time: 03:38:00 | Loss: 121\n",
      "Phase: train | Epoch: 1 | Count: 2100 \\ Start Time: 03:39:02 | Loss: 117\n",
      "Phase: train | Epoch: 1 | Count: 2200 \\ Start Time: 03:40:04 | Loss: 116\n",
      "Phase: train | Epoch: 1 | Count: 2300 \\ Start Time: 03:41:07 | Loss: 115\n",
      "Phase: train | Epoch: 1 | Count: 2400 \\ Start Time: 03:42:09 | Loss: 113\n",
      "Phase: train | Epoch: 1 | Count: 2500 \\ Start Time: 03:43:12 | Loss: 110\n",
      "Phase: train | Epoch: 1 | Count: 2600 \\ Start Time: 03:44:15 | Loss: 110\n",
      "Phase: train | Epoch: 1 | Count: 2700 \\ Start Time: 03:45:17 | Loss: 108\n",
      "Phase: train | Epoch: 1 | Count: 2800 \\ Start Time: 03:46:19 | Loss: 108\n",
      "Phase: train | Epoch: 1 | Count: 2900 \\ Start Time: 03:47:21 | Loss: 107\n",
      "Phase: train | Epoch: 1 | Count: 3000 \\ Start Time: 03:48:24 | Loss: 103\n",
      "Phase: train | Epoch: 2 | Count: 3100 \\ Start Time: 03:49:26 | Loss: 102\n",
      "Phase: train | Epoch: 2 | Count: 3200 \\ Start Time: 03:50:32 | Loss: 102\n",
      "Phase: train | Epoch: 2 | Count: 3300 \\ Start Time: 03:51:34 | Loss: 98\n",
      "Phase: train | Epoch: 2 | Count: 3400 \\ Start Time: 03:52:36 | Loss: 98\n",
      "Phase: train | Epoch: 2 | Count: 3500 \\ Start Time: 03:53:38 | Loss: 98\n",
      "Phase: train | Epoch: 2 | Count: 3600 \\ Start Time: 03:54:41 | Loss: 95\n",
      "Phase: train | Epoch: 2 | Count: 3700 \\ Start Time: 03:55:44 | Loss: 94\n",
      "Phase: train | Epoch: 2 | Count: 3800 \\ Start Time: 03:56:46 | Loss: 95\n",
      "Phase: train | Epoch: 2 | Count: 3900 \\ Start Time: 03:57:48 | Loss: 90\n",
      "Phase: train | Epoch: 2 | Count: 4000 \\ Start Time: 03:58:50 | Loss: 91\n",
      "Phase: train | Epoch: 2 | Count: 4100 \\ Start Time: 03:59:53 | Loss: 92\n",
      "Phase: train | Epoch: 2 | Count: 4200 \\ Start Time: 04:00:55 | Loss: 91\n",
      "Phase: train | Epoch: 2 | Count: 4300 \\ Start Time: 04:01:57 | Loss: 90\n",
      "Phase: train | Epoch: 2 | Count: 4400 \\ Start Time: 04:03:01 | Loss: 87\n",
      "Phase: train | Epoch: 2 | Count: 4500 \\ Start Time: 04:04:03 | Loss: 86\n",
      "Phase: train | Epoch: 3 | Count: 4600 \\ Start Time: 04:05:05 | Loss: 87\n",
      "Phase: train | Epoch: 3 | Count: 4700 \\ Start Time: 04:06:09 | Loss: 85\n",
      "Phase: train | Epoch: 3 | Count: 4800 \\ Start Time: 04:07:11 | Loss: 83\n",
      "Phase: train | Epoch: 3 | Count: 4900 \\ Start Time: 04:08:14 | Loss: 83\n",
      "Phase: train | Epoch: 3 | Count: 5000 \\ Start Time: 04:09:16 | Loss: 81\n",
      "Phase: train | Epoch: 3 | Count: 5100 \\ Start Time: 04:10:19 | Loss: 80\n",
      "Phase: train | Epoch: 3 | Count: 5200 \\ Start Time: 04:11:22 | Loss: 79\n",
      "Phase: train | Epoch: 3 | Count: 5300 \\ Start Time: 04:12:25 | Loss: 79\n",
      "Phase: train | Epoch: 3 | Count: 5400 \\ Start Time: 04:13:28 | Loss: 76\n",
      "Phase: train | Epoch: 3 | Count: 5500 \\ Start Time: 04:14:30 | Loss: 77\n",
      "Phase: train | Epoch: 3 | Count: 5600 \\ Start Time: 04:15:33 | Loss: 77\n",
      "Phase: train | Epoch: 3 | Count: 5700 \\ Start Time: 04:16:36 | Loss: 76\n",
      "Phase: train | Epoch: 3 | Count: 5800 \\ Start Time: 04:17:38 | Loss: 75\n",
      "Phase: train | Epoch: 3 | Count: 5900 \\ Start Time: 04:18:41 | Loss: 75\n",
      "Phase: train | Epoch: 3 | Count: 6000 \\ Start Time: 04:19:44 | Loss: 74\n",
      "Phase: train | Epoch: 4 | Count: 6100 \\ Start Time: 04:20:48 | Loss: 74\n",
      "Phase: train | Epoch: 4 | Count: 6200 \\ Start Time: 04:21:52 | Loss: 72\n",
      "Phase: train | Epoch: 4 | Count: 6300 \\ Start Time: 04:22:56 | Loss: 71\n",
      "Phase: train | Epoch: 4 | Count: 6400 \\ Start Time: 04:24:00 | Loss: 69\n",
      "Phase: train | Epoch: 4 | Count: 6500 \\ Start Time: 04:25:03 | Loss: 70\n",
      "Phase: train | Epoch: 4 | Count: 6600 \\ Start Time: 04:26:07 | Loss: 70\n",
      "Phase: train | Epoch: 4 | Count: 6700 \\ Start Time: 04:27:11 | Loss: 69\n",
      "Phase: train | Epoch: 4 | Count: 6800 \\ Start Time: 04:28:14 | Loss: 68\n",
      "Phase: train | Epoch: 4 | Count: 6900 \\ Start Time: 04:29:20 | Loss: 66\n",
      "Phase: train | Epoch: 4 | Count: 7000 \\ Start Time: 04:30:24 | Loss: 67\n",
      "Phase: train | Epoch: 4 | Count: 7100 \\ Start Time: 04:31:30 | Loss: 67\n",
      "Phase: train | Epoch: 4 | Count: 7200 \\ Start Time: 04:32:33 | Loss: 67\n",
      "Phase: train | Epoch: 4 | Count: 7300 \\ Start Time: 04:33:37 | Loss: 64\n",
      "Phase: train | Epoch: 4 | Count: 7400 \\ Start Time: 04:34:42 | Loss: 65\n",
      "Phase: train | Epoch: 4 | Count: 7500 \\ Start Time: 04:35:46 | Loss: 63\n",
      "Phase: train | Epoch: 4 | Count: 7600 \\ Start Time: 04:36:49 | Loss: 62\n",
      "Phase: train | Epoch: 5 | Count: 7700 \\ Start Time: 04:37:52 | Loss: 62\n",
      "Phase: train | Epoch: 5 | Count: 7800 \\ Start Time: 04:38:55 | Loss: 63\n",
      "Phase: train | Epoch: 5 | Count: 7900 \\ Start Time: 04:39:58 | Loss: 61\n",
      "Phase: train | Epoch: 5 | Count: 8000 \\ Start Time: 04:41:02 | Loss: 61\n",
      "Phase: train | Epoch: 5 | Count: 8100 \\ Start Time: 04:42:05 | Loss: 61\n",
      "Phase: train | Epoch: 5 | Count: 8200 \\ Start Time: 04:43:08 | Loss: 60\n",
      "Phase: train | Epoch: 5 | Count: 8300 \\ Start Time: 04:44:11 | Loss: 60\n",
      "Phase: train | Epoch: 5 | Count: 8400 \\ Start Time: 04:45:14 | Loss: 57\n",
      "Phase: train | Epoch: 5 | Count: 8500 \\ Start Time: 04:46:17 | Loss: 57\n",
      "Phase: train | Epoch: 5 | Count: 8600 \\ Start Time: 04:47:20 | Loss: 59\n",
      "Phase: train | Epoch: 5 | Count: 8700 \\ Start Time: 04:48:24 | Loss: 60\n",
      "Phase: train | Epoch: 5 | Count: 8800 \\ Start Time: 04:49:30 | Loss: 56\n",
      "Phase: train | Epoch: 5 | Count: 8900 \\ Start Time: 04:50:33 | Loss: 59\n",
      "Phase: train | Epoch: 5 | Count: 9000 \\ Start Time: 04:51:37 | Loss: 57\n",
      "Phase: train | Epoch: 5 | Count: 9100 \\ Start Time: 04:52:41 | Loss: 56\n",
      "Phase: train | Epoch: 6 | Count: 9200 \\ Start Time: 04:53:44 | Loss: 56\n",
      "Phase: train | Epoch: 6 | Count: 9300 \\ Start Time: 04:54:47 | Loss: 56\n",
      "Phase: train | Epoch: 6 | Count: 9400 \\ Start Time: 04:55:51 | Loss: 54\n",
      "Phase: train | Epoch: 6 | Count: 9500 \\ Start Time: 04:56:55 | Loss: 54\n",
      "Phase: train | Epoch: 6 | Count: 9600 \\ Start Time: 04:57:58 | Loss: 54\n",
      "Phase: train | Epoch: 6 | Count: 9700 \\ Start Time: 04:59:02 | Loss: 53\n",
      "Phase: train | Epoch: 6 | Count: 9800 \\ Start Time: 05:00:05 | Loss: 54\n",
      "Phase: train | Epoch: 6 | Count: 9900 \\ Start Time: 05:01:10 | Loss: 53\n",
      "Phase: train | Epoch: 6 | Count: 10000 \\ Start Time: 05:02:14 | Loss: 51\n",
      "Phase: train | Epoch: 6 | Count: 10100 \\ Start Time: 05:03:19 | Loss: 53\n",
      "Phase: train | Epoch: 6 | Count: 10200 \\ Start Time: 05:04:23 | Loss: 53\n",
      "Phase: train | Epoch: 6 | Count: 10300 \\ Start Time: 05:05:28 | Loss: 52\n",
      "Phase: train | Epoch: 6 | Count: 10400 \\ Start Time: 05:06:32 | Loss: 54\n",
      "Phase: train | Epoch: 6 | Count: 10500 \\ Start Time: 05:07:36 | Loss: 51\n",
      "Phase: train | Epoch: 6 | Count: 10600 \\ Start Time: 05:08:39 | Loss: 51\n",
      "Phase: train | Epoch: 7 | Count: 10700 \\ Start Time: 05:09:43 | Loss: 51\n",
      "Phase: train | Epoch: 7 | Count: 10800 \\ Start Time: 05:10:48 | Loss: 52\n",
      "Phase: train | Epoch: 7 | Count: 10900 \\ Start Time: 05:11:52 | Loss: 51\n",
      "Phase: train | Epoch: 7 | Count: 11000 \\ Start Time: 05:12:57 | Loss: 49\n",
      "Phase: train | Epoch: 7 | Count: 11100 \\ Start Time: 05:14:02 | Loss: 48\n",
      "Phase: train | Epoch: 7 | Count: 11200 \\ Start Time: 05:15:07 | Loss: 49\n",
      "Phase: train | Epoch: 7 | Count: 11300 \\ Start Time: 05:16:11 | Loss: 49\n",
      "Phase: train | Epoch: 7 | Count: 11400 \\ Start Time: 05:17:16 | Loss: 48\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Phase: train | Epoch: 7 | Count: 11500 \\ Start Time: 05:18:20 | Loss: 46\n",
      "Phase: train | Epoch: 7 | Count: 11600 \\ Start Time: 05:19:24 | Loss: 48\n",
      "Phase: train | Epoch: 7 | Count: 11700 \\ Start Time: 05:20:29 | Loss: 48\n",
      "Phase: train | Epoch: 7 | Count: 11800 \\ Start Time: 05:21:33 | Loss: 47\n",
      "Phase: train | Epoch: 7 | Count: 11900 \\ Start Time: 05:22:37 | Loss: 48\n",
      "Phase: train | Epoch: 7 | Count: 12000 \\ Start Time: 05:23:41 | Loss: 49\n",
      "Phase: train | Epoch: 7 | Count: 12100 \\ Start Time: 05:24:45 | Loss: 46\n",
      "Phase: train | Epoch: 8 | Count: 12200 \\ Start Time: 05:25:49 | Loss: 47\n",
      "Phase: train | Epoch: 8 | Count: 12300 \\ Start Time: 05:26:53 | Loss: 47\n",
      "Phase: train | Epoch: 8 | Count: 12400 \\ Start Time: 05:27:58 | Loss: 46\n",
      "Phase: train | Epoch: 8 | Count: 12500 \\ Start Time: 05:29:01 | Loss: 44\n",
      "Phase: train | Epoch: 8 | Count: 12600 \\ Start Time: 05:30:07 | Loss: 45\n",
      "Phase: train | Epoch: 8 | Count: 12700 \\ Start Time: 05:31:12 | Loss: 44\n",
      "Phase: train | Epoch: 8 | Count: 12800 \\ Start Time: 05:32:16 | Loss: 44\n",
      "Phase: train | Epoch: 8 | Count: 12900 \\ Start Time: 05:33:20 | Loss: 44\n",
      "Phase: train | Epoch: 8 | Count: 13000 \\ Start Time: 05:34:25 | Loss: 43\n",
      "Phase: train | Epoch: 8 | Count: 13100 \\ Start Time: 05:35:30 | Loss: 44\n",
      "Phase: train | Epoch: 8 | Count: 13200 \\ Start Time: 05:36:34 | Loss: 45\n",
      "Phase: train | Epoch: 8 | Count: 13300 \\ Start Time: 05:37:41 | Loss: 44\n",
      "Phase: train | Epoch: 8 | Count: 13400 \\ Start Time: 05:38:44 | Loss: 45\n",
      "Phase: train | Epoch: 8 | Count: 13500 \\ Start Time: 05:39:48 | Loss: 45\n",
      "Phase: train | Epoch: 8 | Count: 13600 \\ Start Time: 05:40:52 | Loss: 44\n",
      "Phase: train | Epoch: 9 | Count: 13700 \\ Start Time: 05:41:56 | Loss: 42\n",
      "Phase: train | Epoch: 9 | Count: 13800 \\ Start Time: 05:43:00 | Loss: 45\n",
      "Phase: train | Epoch: 9 | Count: 13900 \\ Start Time: 05:44:04 | Loss: 43\n",
      "Phase: train | Epoch: 9 | Count: 14000 \\ Start Time: 05:45:08 | Loss: 42\n",
      "Phase: train | Epoch: 9 | Count: 14100 \\ Start Time: 05:46:13 | Loss: 43\n",
      "Phase: train | Epoch: 9 | Count: 14200 \\ Start Time: 05:47:17 | Loss: 41\n",
      "Phase: train | Epoch: 9 | Count: 14300 \\ Start Time: 05:48:21 | Loss: 40\n",
      "Phase: train | Epoch: 9 | Count: 14400 \\ Start Time: 05:49:26 | Loss: 42\n",
      "Phase: train | Epoch: 9 | Count: 14500 \\ Start Time: 05:50:33 | Loss: 39\n",
      "Phase: train | Epoch: 9 | Count: 14600 \\ Start Time: 05:51:38 | Loss: 40\n",
      "Phase: train | Epoch: 9 | Count: 14700 \\ Start Time: 05:52:41 | Loss: 43\n",
      "Phase: train | Epoch: 9 | Count: 14800 \\ Start Time: 05:53:45 | Loss: 41\n",
      "Phase: train | Epoch: 9 | Count: 14900 \\ Start Time: 05:54:49 | Loss: 40\n",
      "Phase: train | Epoch: 9 | Count: 15000 \\ Start Time: 05:55:54 | Loss: 42\n",
      "Phase: train | Epoch: 9 | Count: 15100 \\ Start Time: 05:56:59 | Loss: 41\n",
      "Phase: train | Epoch: 9 | Count: 15200 \\ Start Time: 05:58:03 | Loss: 40\n"
     ]
    }
   ],
   "source": [
    "# TRAINING AND TESTING\n",
    "print(\"training...\")\n",
    "model.start(phase='train') "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Generating Images"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_model = Test(category)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataloader = test_model.dataloaders['train']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = dataloader.next_batch(test_model.batch_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(64, 784)"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x1dff2ce5508>"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(npy_data[int(index)].reshape(28,28))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "reconstructing images...\n",
      "C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\\src\\data\\output\\images\\moon\\moon_test_0.png\n",
      "C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\\src\\data\\output\\images\\moon\\moon_test_1.png\n",
      "C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\\src\\data\\output\\images\\moon\\moon_test_2.png\n",
      "C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\\src\\data\\output\\images\\moon\\moon_test_3.png\n",
      "C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\\src\\data\\output\\images\\moon\\moon_test_4.png\n",
      "C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\\src\\data\\output\\images\\moon\\moon_test_5.png\n",
      "C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\\src\\data\\output\\images\\moon\\moon_test_6.png\n",
      "C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\\src\\data\\output\\images\\moon\\moon_test_7.png\n",
      "C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\\src\\data\\output\\images\\moon\\moon_test_8.png\n",
      "C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\\src\\data\\output\\images\\moon\\moon_test_9.png\n",
      "C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\\src\\data\\output\\images\\moon\\moon_test_10.png\n",
      "C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\\src\\data\\output\\images\\moon\\moon_test_11.png\n",
      "C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\\src\\data\\output\\images\\moon\\moon_test_12.png\n",
      "C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\\src\\data\\output\\images\\moon\\moon_test_13.png\n",
      "C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\\src\\data\\output\\images\\moon\\moon_test_14.png\n",
      "C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\\src\\data\\output\\images\\moon\\moon_test_15.png\n",
      "C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\\src\\data\\output\\images\\moon\\moon_test_16.png\n",
      "C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\\src\\data\\output\\images\\moon\\moon_test_17.png\n",
      "C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\\src\\data\\output\\images\\moon\\moon_test_18.png\n",
      "C:\\Users\\ozgur\\Desktop\\sandbox\\sliver-maestro\\src\\data\\output\\images\\moon\\moon_test_19.png\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([[5.00000000e-01, 5.00000000e-01, 5.00000000e-01, 5.00000000e-01,\n",
       "        5.00000000e-01, 5.00000000e-01, 5.00000000e-01, 5.00000000e-01,\n",
       "        5.00000000e-01, 5.00000000e-01, 5.00000000e-01, 5.00000000e-01,\n",
       "        5.00000000e-01, 5.00000000e-01, 5.00000000e-01, 5.00000000e-01,\n",
       "        5.00000000e-01, 5.00000000e-01, 5.00000000e-01, 5.00000000e-01,\n",
       "        5.00000000e-01, 5.00000000e-01, 5.00000000e-01, 5.00000000e-01,\n",
       "        5.00000000e-01, 5.00000000e-01, 5.00000000e-01, 5.00000000e-01,\n",
       "        5.00000000e-01, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 0.00000000e+00, 0.00000000e+00,\n",
       "        0.00000000e+00, 0.00000000e+00, 0.00000000e+00, 0.00000000e+00,\n",
       "        0.00000000e+00, 0.00000000e+00, 0.00000000e+00, 0.00000000e+00,\n",
       "        8.19102888e-36, 1.21299873e-31, 2.43768660e-28, 4.35705412e-26,\n",
       "        6.61047344e-25, 3.53194129e-24, 3.38512426e-23, 2.92931393e-22,\n",
       "        1.87593067e-22, 3.30519289e-24, 1.02133313e-26, 5.28658802e-29,\n",
       "        6.30634334e-31, 1.11414192e-32, 2.20495786e-34, 8.80728136e-38,\n",
       "        0.00000000e+00, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 0.00000000e+00, 0.00000000e+00,\n",
       "        0.00000000e+00, 0.00000000e+00, 0.00000000e+00, 0.00000000e+00,\n",
       "        8.46151993e-37, 3.37886955e-29, 3.74563527e-22, 2.80040924e-16,\n",
       "        8.26553583e-12, 8.89715146e-09, 3.46231531e-07, 7.69037797e-07,\n",
       "        4.28566750e-07, 4.84906366e-07, 2.74184936e-06, 1.30046847e-05,\n",
       "        3.04942091e-06, 2.16441833e-08, 5.83587900e-11, 8.26753325e-13,\n",
       "        6.15994614e-14, 4.31065570e-15, 5.07276021e-17, 2.46308274e-21,\n",
       "        2.31593208e-32, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 0.00000000e+00, 0.00000000e+00,\n",
       "        0.00000000e+00, 0.00000000e+00, 4.36654398e-31, 2.29484512e-23,\n",
       "        8.98907164e-17, 1.50583799e-11, 9.99355692e-08, 3.89217566e-05,\n",
       "        1.27537106e-03, 3.30846733e-03, 6.57348835e-04, 2.92773730e-05,\n",
       "        2.51801453e-06, 1.89709988e-06, 6.64956906e-06, 9.28034387e-06,\n",
       "        7.80659775e-07, 9.49366630e-09, 2.62985606e-10, 8.27959853e-11,\n",
       "        9.81452281e-11, 4.96350426e-11, 2.43614460e-12, 1.02897234e-15,\n",
       "        3.81185084e-24, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 0.00000000e+00, 0.00000000e+00,\n",
       "        5.16796071e-39, 4.82009459e-27, 6.97092565e-18, 6.47787865e-11,\n",
       "        3.95361576e-06, 2.29900749e-03, 4.75855917e-02, 1.60642833e-01,\n",
       "        1.87650144e-01, 4.13182601e-02, 6.35723933e-04, 2.72214970e-06,\n",
       "        5.70087337e-08, 2.39433096e-08, 4.32168612e-08, 2.03330615e-08,\n",
       "        8.55893023e-10, 2.07994993e-11, 4.41075822e-12, 1.54245384e-11,\n",
       "        1.37010153e-10, 3.34039713e-10, 7.47644238e-11, 3.34826351e-13,\n",
       "        2.25803066e-19, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 0.00000000e+00, 0.00000000e+00,\n",
       "        2.68635931e-28, 4.68915653e-17, 3.56016550e-09, 6.14715624e-04,\n",
       "        3.14463466e-01, 8.16724777e-01, 8.07593822e-01, 7.22294450e-01,\n",
       "        6.18181467e-01, 2.02044234e-01, 2.20363727e-03, 2.46859349e-06,\n",
       "        7.37266337e-09, 4.62892419e-10, 2.20823818e-10, 5.55538393e-11,\n",
       "        2.28158573e-12, 8.56245549e-14, 4.13186297e-14, 4.76039401e-13,\n",
       "        1.82967357e-11, 1.99994882e-10, 1.74003645e-10, 3.26989802e-12,\n",
       "        2.32124780e-17, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 0.00000000e+00, 7.06952450e-34,\n",
       "        2.09846286e-19, 8.11963052e-10, 8.33090511e-04, 6.88090980e-01,\n",
       "        9.82672215e-01, 9.84078884e-01, 9.54944968e-01, 9.26419258e-01,\n",
       "        9.36816156e-01, 8.50220263e-01, 7.27448687e-02, 3.79613666e-05,\n",
       "        1.01405560e-08, 3.26274147e-11, 2.20411233e-12, 3.85330548e-13,\n",
       "        2.45412550e-14, 1.13295400e-15, 5.04135693e-16, 8.65768384e-15,\n",
       "        1.11981821e-12, 5.38348775e-11, 1.35316341e-10, 3.53851041e-12,\n",
       "        1.51387965e-17, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 0.00000000e+00, 8.72184462e-26,\n",
       "        5.67075358e-13, 3.17739032e-05, 3.38508129e-01, 9.78991389e-01,\n",
       "        9.93631184e-01, 9.89076018e-01, 9.78934228e-01, 9.85133052e-01,\n",
       "        9.94713128e-01, 9.92518544e-01, 6.70922816e-01, 4.35145572e-04,\n",
       "        1.56856430e-08, 3.90143690e-12, 4.13098036e-14, 4.53140519e-15,\n",
       "        4.54230629e-16, 3.37443335e-17, 1.90504840e-17, 5.09883712e-16,\n",
       "        1.60116400e-13, 2.00803575e-11, 8.12529696e-11, 1.27292014e-12,\n",
       "        5.97468162e-19, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 4.73217931e-38, 7.80601620e-20,\n",
       "        1.22057662e-08, 1.17185730e-02, 8.47394645e-01, 9.77900088e-01,\n",
       "        9.81938243e-01, 9.77379620e-01, 9.84308898e-01, 9.95650232e-01,\n",
       "        9.98718619e-01, 9.96403575e-01, 5.89758098e-01, 1.15902316e-04,\n",
       "        1.84979232e-09, 1.91434040e-13, 8.40925575e-16, 5.94588012e-17,\n",
       "        8.79018615e-18, 1.98648251e-18, 4.00980600e-18, 2.60491695e-16,\n",
       "        1.21290117e-13, 1.69986074e-11, 5.98479946e-11, 4.23210078e-13,\n",
       "        2.02694341e-20, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 1.55749078e-32, 3.76882439e-15,\n",
       "        1.11965919e-05, 2.12012097e-01, 8.86004269e-01, 9.22953486e-01,\n",
       "        9.08979893e-01, 9.38531756e-01, 9.83213186e-01, 9.96431947e-01,\n",
       "        9.96722162e-01, 9.16094720e-01, 6.11220812e-03, 2.03131250e-07,\n",
       "        5.90858786e-12, 1.84650133e-15, 1.50783958e-17, 1.25531782e-18,\n",
       "        3.05138217e-19, 2.53477556e-19, 2.54124422e-18, 4.39988459e-16,\n",
       "        2.28286438e-13, 2.43386075e-11, 6.42047110e-11, 2.91890345e-13,\n",
       "        5.27202917e-21, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 3.51059165e-27, 1.94518984e-11,\n",
       "        9.46756743e-04, 5.51989913e-01, 8.54055047e-01, 7.84963310e-01,\n",
       "        7.41813898e-01, 8.86265278e-01, 9.79611576e-01, 9.91536975e-01,\n",
       "        9.22993004e-01, 1.49977552e-02, 6.60917465e-07, 1.55839890e-11,\n",
       "        2.78805680e-15, 1.14895189e-17, 6.82423747e-19, 1.64954833e-19,\n",
       "        7.46001236e-20, 1.16275150e-19, 2.24617607e-18, 6.14343097e-16,\n",
       "        3.82601801e-13, 4.25633417e-11, 1.17818075e-10, 6.75788798e-13,\n",
       "        2.91012673e-20, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 2.14683863e-22, 6.70214062e-09,\n",
       "        9.16789006e-03, 6.54999614e-01, 7.92538881e-01, 6.39105380e-01,\n",
       "        6.12465918e-01, 8.55419397e-01, 9.70441759e-01, 9.53459918e-01,\n",
       "        1.09371506e-01, 6.21742947e-06, 4.67395532e-11, 2.38408731e-15,\n",
       "        6.36188662e-18, 6.08455499e-19, 4.26045914e-19, 4.47281579e-19,\n",
       "        3.46834323e-19, 4.35029328e-19, 4.99946804e-18, 1.09265207e-15,\n",
       "        8.51204686e-13, 1.33750566e-10, 4.95060992e-10, 5.60727810e-12,\n",
       "        2.29127313e-18, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 8.76814551e-19, 2.32409647e-07,\n",
       "        2.27650777e-02, 5.78127205e-01, 6.09458208e-01, 3.88474315e-01,\n",
       "        3.97071093e-01, 7.23573029e-01, 8.91636312e-01, 4.97298837e-01,\n",
       "        4.75099456e-04, 2.94976998e-09, 2.06832774e-14, 1.01995952e-17,\n",
       "        7.68036889e-19, 1.35625588e-18, 6.14879373e-18, 1.66232267e-17,\n",
       "        1.68172519e-17, 1.58063032e-17, 1.02951369e-16, 1.49621323e-14,\n",
       "        1.04597823e-11, 1.58610602e-09, 5.77906079e-09, 1.16545676e-10,\n",
       "        6.02389397e-16, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 1.76082096e-16, 2.84431508e-06,\n",
       "        4.58796807e-02, 4.17651802e-01, 2.24895373e-01, 6.65444061e-02,\n",
       "        6.41480237e-02, 1.81140453e-01, 2.30589315e-01, 7.54436618e-03,\n",
       "        6.26292149e-07, 1.95253405e-12, 6.64792750e-17, 1.06009606e-18,\n",
       "        2.98275766e-18, 5.35747313e-17, 5.44365046e-16, 1.41728009e-15,\n",
       "        1.37112548e-15, 1.81250079e-15, 1.81377079e-14, 2.61485820e-12,\n",
       "        1.05274633e-09, 7.36520391e-08, 1.47539623e-07, 3.20579496e-09,\n",
       "        9.27593370e-14, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 2.07965202e-15, 1.76361464e-05,\n",
       "        1.05869472e-01, 3.08244050e-01, 4.16741334e-02, 4.05259384e-03,\n",
       "        2.54086452e-03, 5.51798614e-03, 3.56650585e-03, 3.50307018e-05,\n",
       "        1.54099677e-09, 8.45190396e-15, 2.91291735e-18, 1.35898009e-18,\n",
       "        7.13131464e-17, 5.59075923e-15, 5.94903155e-14, 8.15876636e-14,\n",
       "        5.70328154e-14, 1.31481753e-13, 3.38344543e-12, 6.82915113e-10,\n",
       "        1.58876276e-07, 4.50527159e-06, 4.21387449e-06, 6.08642239e-08,\n",
       "        2.31096500e-12, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 3.40099356e-15, 3.35865079e-05,\n",
       "        1.74207419e-01, 2.91472435e-01, 1.40065700e-02, 5.05361182e-04,\n",
       "        1.84263947e-04, 3.30874929e-04, 2.08267971e-04, 2.73925912e-06,\n",
       "        3.13438775e-10, 9.53101208e-15, 2.57969923e-17, 6.70180607e-17,\n",
       "        9.15075782e-15, 9.17588163e-13, 7.62145555e-12, 6.09556502e-12,\n",
       "        2.42495368e-12, 4.01648506e-12, 9.46924275e-11, 1.84883735e-08,\n",
       "        3.97112672e-06, 9.86543891e-05, 6.95889612e-05, 5.36547930e-07,\n",
       "        6.11668051e-12, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 3.91268362e-15, 2.04158787e-05,\n",
       "        1.37621447e-01, 2.66701967e-01, 9.59412567e-03, 2.03742035e-04,\n",
       "        5.70107914e-05, 1.51454267e-04, 2.73465790e-04, 2.09547197e-05,\n",
       "        3.15052588e-08, 1.47180289e-11, 1.84735216e-13, 3.70386474e-13,\n",
       "        1.48083611e-11, 5.35858580e-10, 2.97415870e-09, 1.92343563e-09,\n",
       "        3.62610497e-10, 1.23228081e-10, 4.95693153e-10, 4.01349887e-08,\n",
       "        1.13943406e-05, 5.74498612e-04, 5.39144210e-04, 2.14083661e-06,\n",
       "        2.76355848e-12, 5.00000000e-01],\n",
       "       [5.00000000e-01, 9.04393405e-39, 7.59510473e-15, 7.19744139e-06,\n",
       "        3.83758284e-02, 1.04579315e-01, 4.25559515e-03, 1.02290011e-04,\n",
       "        3.55513694e-05, 2.17334178e-04, 2.04104488e-03, 1.96171855e-03,\n",
       "        8.68610223e-05, 1.03264676e-06, 5.12146059e-08, 3.68595927e-08,\n",
       "        1.73693195e-07, 1.17729678e-06, 3.26046506e-06, 1.63575828e-06,\n",
       "        1.39788852e-07, 7.38905026e-09, 3.16523296e-09, 7.02752416e-08,\n",
       "        2.11997631e-05, 1.99889368e-03, 2.44623236e-03, 4.95814129e-06,\n",
       "        6.34481671e-13, 5.00000000e-01],\n",
       "       [5.00000000e-01, 1.87048548e-37, 9.27831759e-15, 1.46279001e-06,\n",
       "        2.61109113e-03, 6.48667105e-03, 5.23679599e-04, 3.55336088e-05,\n",
       "        2.77191684e-05, 3.79230711e-04, 1.23176100e-02, 8.41598585e-02,\n",
       "        6.76805228e-02, 1.56981368e-02, 3.64742987e-03, 1.71340944e-03,\n",
       "        1.64698786e-03, 2.26426730e-03, 2.19235220e-03, 5.92609693e-04,\n",
       "        2.82090459e-05, 6.45850605e-07, 9.98363063e-08, 8.93098672e-07,\n",
       "        1.45027399e-04, 9.27874632e-03, 8.15274287e-03, 8.96562597e-06,\n",
       "        2.49844280e-13, 5.00000000e-01],\n",
       "       [5.00000000e-01, 8.32807596e-38, 1.08926023e-15, 8.36979694e-08,\n",
       "        5.06263168e-05, 9.67968081e-05, 2.69979937e-05, 1.22521687e-05,\n",
       "        3.60844897e-05, 8.10031430e-04, 3.31003815e-02, 3.39934766e-01,\n",
       "        6.50156200e-01, 7.46225476e-01, 7.63556719e-01, 7.16242969e-01,\n",
       "        5.89579105e-01, 4.01058555e-01, 1.73024103e-01, 2.39267871e-02,\n",
       "        8.83535715e-04, 2.84634934e-05, 6.78586912e-06, 4.57423521e-05,\n",
       "        2.44038319e-03, 4.34059463e-02, 1.77704580e-02, 1.46180364e-05,\n",
       "        3.68397505e-13, 5.00000000e-01],\n",
       "       [5.00000000e-01, 3.58419637e-39, 1.10558562e-17, 1.70656100e-09,\n",
       "        1.15653586e-06, 2.73035766e-06, 2.60425281e-06, 7.37307073e-06,\n",
       "        7.45171346e-05, 1.76749076e-03, 3.74340154e-02, 2.62485355e-01,\n",
       "        6.14353359e-01, 8.63999903e-01, 9.58888054e-01, 9.78998780e-01,\n",
       "        9.72764015e-01, 9.22426164e-01, 6.61604941e-01, 1.18188746e-01,\n",
       "        4.60323598e-03, 2.66247545e-04, 1.24456928e-04, 7.58278649e-04,\n",
       "        1.46647161e-02, 7.97192156e-02, 1.87108982e-02, 1.99088463e-05,\n",
       "        1.87941047e-12, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 1.06533328e-19, 6.19137033e-11,\n",
       "        1.97987362e-07, 1.16443448e-06, 2.18543119e-06, 1.30244316e-05,\n",
       "        1.89024955e-04, 2.59160902e-03, 1.67836025e-02, 4.93480675e-02,\n",
       "        1.24135315e-01, 3.93498689e-01, 8.07887971e-01, 9.46546435e-01,\n",
       "        9.63390827e-01, 9.30941939e-01, 7.41053700e-01, 1.96059272e-01,\n",
       "        1.04729990e-02, 7.18359312e-04, 3.12937598e-04, 1.24072155e-03,\n",
       "        1.24595715e-02, 4.07599881e-02, 8.34136922e-03, 1.69387058e-05,\n",
       "        1.14942448e-11, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 4.85507966e-21, 7.59991289e-12,\n",
       "        2.23743527e-07, 6.21691379e-06, 1.82719396e-05, 9.31848408e-05,\n",
       "        7.59485294e-04, 3.19751888e-03, 4.63787327e-03, 4.25429409e-03,\n",
       "        6.16379315e-03, 2.16136985e-02, 1.23084292e-01, 4.37667608e-01,\n",
       "        6.83767796e-01, 7.25153208e-01, 5.92924416e-01, 2.33226657e-01,\n",
       "        2.28508394e-02, 1.51232909e-03, 3.50361748e-04, 6.20196806e-04,\n",
       "        3.40748322e-03, 8.31013359e-03, 1.64218561e-03, 5.25595215e-06,\n",
       "        1.35165889e-11, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 2.49312358e-22, 7.30962545e-13,\n",
       "        1.72954316e-07, 3.63432919e-05, 2.92889919e-04, 1.45205029e-03,\n",
       "        4.90446063e-03, 4.54737432e-03, 1.43086701e-03, 5.39035304e-04,\n",
       "        5.94023732e-04, 1.79040723e-03, 8.58180970e-03, 4.08486687e-02,\n",
       "        1.43396199e-01, 3.11615735e-01, 4.04398650e-01, 2.72983909e-01,\n",
       "        5.50782792e-02, 4.29482199e-03, 5.88977768e-04, 4.16483555e-04,\n",
       "        9.03420965e-04, 1.06565934e-03, 1.25195380e-04, 2.59654485e-07,\n",
       "        3.56836387e-13, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 1.51714646e-24, 1.03650532e-14,\n",
       "        1.19955121e-08, 1.94194781e-05, 8.04641226e-04, 7.57344114e-03,\n",
       "        1.57872569e-02, 4.46833763e-03, 5.05438016e-04, 1.61861259e-04,\n",
       "        2.86266149e-04, 1.22672308e-03, 5.22755180e-03, 1.87326465e-02,\n",
       "        6.61242679e-02, 2.02250481e-01, 3.68667215e-01, 3.34708095e-01,\n",
       "        1.06868684e-01, 1.12294983e-02, 1.22433237e-03, 3.60637816e-04,\n",
       "        2.12354164e-04, 6.46673943e-05, 2.03440004e-06, 7.12567005e-10,\n",
       "        3.15927904e-17, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 7.03382308e-28, 1.72209210e-17,\n",
       "        5.12726875e-11, 3.43035907e-07, 6.69390574e-05, 1.89308962e-03,\n",
       "        5.94936591e-03, 1.64261484e-03, 2.14123633e-04, 1.42445846e-04,\n",
       "        7.73011358e-04, 7.96456821e-03, 4.10623588e-02, 9.75158811e-02,\n",
       "        1.82444781e-01, 3.19064677e-01, 4.13466394e-01, 3.10771585e-01,\n",
       "        8.83730054e-02, 9.20398999e-03, 7.80980510e-04, 9.55375581e-05,\n",
       "        1.27440726e-05, 6.47834440e-07, 2.48685672e-09, 3.44580118e-14,\n",
       "        2.84128765e-24, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 5.28276977e-32, 7.67248388e-21,\n",
       "        3.39872580e-14, 3.63605701e-10, 1.51194797e-07, 1.11644877e-05,\n",
       "        1.21719459e-04, 1.54233756e-04, 9.41598992e-05, 2.35801941e-04,\n",
       "        3.87549214e-03, 8.63950774e-02, 4.04167354e-01, 5.08187056e-01,\n",
       "        4.02310610e-01, 2.69780874e-01, 1.50896505e-01, 4.81761321e-02,\n",
       "        6.46890886e-03, 4.35166730e-04, 2.22744802e-05, 1.03842899e-06,\n",
       "        2.91324902e-08, 1.77975856e-10, 3.87152943e-14, 5.82913006e-21,\n",
       "        1.19750766e-34, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 7.39593895e-37, 6.48942343e-24,\n",
       "        5.07417304e-17, 3.01361068e-13, 4.81349162e-11, 1.65255465e-09,\n",
       "        2.28548416e-08, 1.24306368e-07, 4.39020397e-07, 3.03796696e-06,\n",
       "        6.46211629e-05, 1.60135620e-03, 9.70028341e-03, 7.18300045e-03,\n",
       "        1.21189677e-03, 1.47915023e-04, 2.33764895e-05, 3.55826410e-06,\n",
       "        3.06888978e-07, 1.32374129e-08, 3.52227220e-10, 5.63645484e-12,\n",
       "        2.79446825e-14, 1.23119333e-17, 6.06581719e-23, 3.52741639e-32,\n",
       "        0.00000000e+00, 5.00000000e-01],\n",
       "       [5.00000000e-01, 0.00000000e+00, 0.00000000e+00, 7.73530340e-24,\n",
       "        1.42827606e-15, 6.65311610e-13, 2.64872461e-13, 3.69032180e-15,\n",
       "        3.38824793e-17, 1.07279680e-18, 1.36398076e-19, 5.53656153e-20,\n",
       "        7.99893798e-20, 3.05463169e-19, 8.88220131e-19, 6.02289694e-19,\n",
       "        1.11794701e-19, 2.02121437e-20, 8.34147024e-21, 4.76898986e-21,\n",
       "        1.22098068e-21, 8.18661609e-23, 1.71001477e-24, 1.02228030e-26,\n",
       "        6.14123162e-30, 6.81408935e-35, 0.00000000e+00, 0.00000000e+00,\n",
       "        0.00000000e+00, 5.00000000e-01],\n",
       "       [5.00000000e-01, 5.00000000e-01, 5.00000000e-01, 5.00000000e-01,\n",
       "        5.00000000e-01, 5.00000000e-01, 5.00000000e-01, 5.00000000e-01,\n",
       "        5.00000000e-01, 5.00000000e-01, 5.00000000e-01, 5.00000000e-01,\n",
       "        5.00000000e-01, 5.00000000e-01, 5.00000000e-01, 5.00000000e-01,\n",
       "        5.00000000e-01, 5.00000000e-01, 5.00000000e-01, 5.00000000e-01,\n",
       "        5.00000000e-01, 5.00000000e-01, 5.00000000e-01, 5.00000000e-01,\n",
       "        5.00000000e-01, 5.00000000e-01, 5.00000000e-01, 5.00000000e-01,\n",
       "        5.00000000e-01, 5.00000000e-01]])"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAO0AAADtCAYAAABTTfKPAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy86wFpkAAAACXBIWXMAAAsTAAALEwEAmpwYAAAJyElEQVR4nO3dy0vUbRjG8WfyUGmZmZORmSVZtKiW7dq06LDoBEJ/QEHRpl2bamGrtq2jQGrXrkJIhLCgKEICE9JSA48gpqkzHuvdvby8dN+PzagzV34/yy5u+2Vd/cCb55nEr1+/AgAd63L9AAD+DKUFxFBaQAylBcRQWkAMpQXEFHrhnTt32AcBOXLr1q3E736dNy0ghtICYigtIIbSAmIoLSDG/emx5/bt28v5HFgB69bZ/ycXFBS4s4uLi2YWO2TCIZSla2xs/OMZ3rSAGEoLiKG0gBhKC4ihtIAYSguIobSAmIz3tFgdZWVlbn7lyhUzu3DhgpnF9rQdHR1m9uLFC3e2vb3dzAYHB93ZVCplZt7uOIS1sx/mTQuIobSAGEoLiKG0gBhKC4ihtIAYVj5LlEj89o6tf9XV1ZnZ1atX3dnjx4+bWW1trTu7adMmM/PWOrH1yZ49e8ysurrana2srDSzN2/euLMDAwNmNjU15c7Ozc2Z2cLCgjsb+37kE960gBhKC4ihtIAYSguIobSAGEoLiKG0gBj2tP/h7WIbGhrc2bt375pZVVWVO1tUVGRmsSN0sf2xJbaXTKfTZjYyMuLOjo+Pm9nPnz/d2eLiYjPbuHGjO+t9r2ZnZ91ZL48982rjTQuIobSAGEoLiKG0gBhKC4ihtIAYVj7/UVFRYWanT592Z71PqPOOjMVkc8Ogt7bp7u52Z588eWJmL1++dGeHh4fNbGZmxp31jtDF1lT5tppZKbxpATGUFhBDaQExlBYQQ2kBMZQWEENpATFrak9bWOj/cWtqaswsttf0joXt3LnTnU0mk2bmHdsLIYT+/n4za2pqMrPm5mb3605MTJhZbB/qHRf09tmxPLaz9vLYMyt94h5vWkAMpQXEUFpADKUFxFBaQAylBcT8dSsfb0XiHb0LIYRUKmVmT58+dWdbWloy/n3Ly8vNbGhoyJ1tb283s8nJSTPL1xWHty6K3TyZzZ8pX78fv8ObFhBDaQExlBYQQ2kBMZQWEENpATGUFhAjt6ctLS118/r6ejMbGxtzZ72rP+fn5/0Hy4J3behK/r75yNuXKu1SVxJvWkAMpQXEUFpADKUFxFBaQAylBcTIrXzOnTvn5tPT02bW19eX8Ww2n8jGqgLLiTctIIbSAmIoLSCG0gJiKC0ghtICYigtICYv97Rbt241s2PHjrmz3ifFeXvYEPxdLLtW5AvetIAYSguIobSAGEoLiKG0gBhKC4jJy5XPxYsXzaygoMCd7enpMTPv1sMQWOtAA29aQAylBcRQWkAMpQXEUFpADKUFxFBaQExO9rTr1vn/V5w6dcrMXr9+7c56n4yXzR42kUi4OTterBbetIAYSguIobSAGEoLiKG0gBhKC4jJycqnsrLSzUtLS83s7du37uzc3FxGzxQTW1N5K59sPnEP+D/etIAYSguIobSAGEoLiKG0gBhKC4ihtICYnOxpk8mkm6dSKTPr6+tzZ1fqiFxxcbGbFxba38rJycnlfhysYbxpATGUFhBDaQExlBYQQ2kBMZQWEJOTlc/mzZvd3DsGNzMzs9yPsySxT+urra01s46OjuV+HKxhvGkBMZQWEENpATGUFhBDaQExlBYQk5OVz7Zt29y8vLzczEpKSpb5aZZmdnbWzauqqsyss7PTneW2RvwJ3rSAGEoLiKG0gBhKC4ihtIAYSguIobSAmJzsaWM3G3pH97x9aAjx2xozNT8/7+a9vb1mtn37dnd2eHg4o2fC0iUSiYxnV+qGz0zxpgXEUFpADKUFxFBaQAylBcRQWkBMTlY+3d3dbj43N2dmu3btcmffvXtnZiv5o/v+/n4z279/vzubTqfNbGJiIuNnUuStZmJrG+8Wz9htmt7xyNi6b7XxpgXEUFpADKUFxFBaQAylBcRQWkAMpQXE5GRP29XV5ebeMbe9e/e6s94+bmFhwX+wLHi7vKGhIXe2vr7ezGKfEuh97ampKTPL5nvh7UND8I9eFhUVubPZ7GkLC+1/zl4WQgjT09Nm9uPHD3d2tY/u8aYFxFBaQAylBcRQWkAMpQXEUFpATE5WPt7RuxBCaGtrM7PDhw+7s8lk0sxiq5dseD/2Hx8fd2e9GyQPHDjgzh45csTMvONmsWfy/o42bNjgzpaVlZlZbPXirbi8I4wh+Gus2GxPT4+ZxVY+q403LSCG0gJiKC0ghtICYigtIIbSAmIoLSAmJ3vamNbWVjM7efKkO3vmzBkze/z4sTvrHc/K5vjV4uKim4+NjZnZx48f3dnv37+bWV1dnZlVV1e7X7eystLMYp8CuGXLFjOLXWWaSqXMLHad7OjoqJl5e9gQQvj69aub5xPetIAYSguIobSAGEoLiKG0gBhKC4jJy5XPly9fzGxyctKdPX/+vJl5txOGEEJzc7OZxY6yecfgYrxZbw0Vgv8JhN5RxNjK59ChQ2bmrYNCCKG0tNTMYrcxejc9euugWD44OOjOen+/q33bYgxvWkAMpQXEUFpADKUFxFBaQAylBcRQWkBMXu5pves7Hz586M7evHnTzC5fvuzO1tTUmFlTU5M7OzIyYmaxo3me2I7QuzbU22kPDAy4X7e8vNzM9u3b585WVFSYWex74T3X58+f3dlPnz6ZmXdNbQjxfXg+4U0LiKG0gBhKC4ihtIAYSguIobSAmLxc+XhaWlrc3Lsp8NKlS+7s2bNnzWz9+vXu7IMHD8xseHjYnc1mJZRIJMzMu/mwpKTE/bo7duwwM2+lE0IIs7OzZtbV1eXOvn//3sy8Y4gh+DdTzs/Pu7P5dvzOw5sWEENpATGUFhBDaQExlBYQQ2kBMZQWECO3p41dVfro0SMz+/Dhgzt77do1Mzt69Kg76+01Y7+vdxwttl/0riT1dtbeFakhhHDw4EEzKyz0/9m0t7eb2atXr9zZzs5OM0un0+6s0q41G7xpATGUFhBDaQExlBYQQ2kBMZQWECO38onxfuzvrRNCCOH69etmduLECXf2xo0bZtbQ0ODOep8y5x2vC8FfgXnrIu/GyxBCGB0dNbPW1lZ39v79+2bW29vrzmZzTHGt4E0LiKG0gBhKC4ihtIAYSguIobSAGEoLiPnr9rTZ8Paaz549c2efP39uZrt373ZnvatbvU/yCyGEnp4eM2tra8toLoQQZmZmzGytHIHLV7xpATGUFhBDaQExlBYQQ2kBMZQWEMPKZ5l4a5Bv3765s/fu3Vvux8FfjDctIIbSAmIoLSCG0gJiKC0ghtICYigtICbjPW1jY+NyPgeAJeJNC4ihtIAYSguIobSAGEoLiKG0gJgEN+sBWnjTAmIoLSCG0gJiKC0ghtICYigtIOYfnIOTDW7a4xsAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "print(\"reconstructing images...\")\n",
    "test_model.xrecons_grid(int(index), data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}