train_models.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "a77ac107-1d3c-498c-95f1-984858cb96fb",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "<h3>Tutorial notebook for:</h3>\n",
    "\n",
    " - Loading datasets\n",
    " - Training standard Logistic Regression (LR) models from *sklearn*\n",
    " - Training standard XGBoost (XGB) models from *xgboost*\n",
    " - Training PyTorch Deep Neural Network (DNN) models as constructed in *models.py*"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7237780c",
   "metadata": {
    "tags": []
   },
   "source": [
    "# Preamble"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "8d5ed870-ec23-46a9-8fc1-c06ddd94250c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'/home/dan-ley/icml-gces'"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import os\n",
    "os.getcwd()  # this notebook should be at the root level"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "68a87ac5",
   "metadata": {
    "tags": []
   },
   "source": [
    "# Load Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "bebaf3a7-9ef8-4a45-9e18-74e57455c7e3",
   "metadata": {},
   "outputs": [],
   "source": [
    "bold = lambda x: '\\033[1m' + x + '\\033[0m'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "bdcecc8c-83ca-42cb-898a-24776cfbf353",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1mDataset Names: \u001b[0mcompas, german_credit, adult_income, default_credit, heloc\n"
     ]
    }
   ],
   "source": [
    "from datasets import dataset_loader\n",
    "from models import performance\n",
    "print(bold('Dataset Names: ') + ', '.join(dataset_loader().datasets.keys()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "24cb34a4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1mHeloc Dataset:\u001b[0m\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ExternalRiskEstimate</th>\n",
       "      <th>MSinceOldestTradeOpen</th>\n",
       "      <th>MSinceMostRecentTradeOpen</th>\n",
       "      <th>AverageMInFile</th>\n",
       "      <th>NumSatisfactoryTrades</th>\n",
       "      <th>NumTrades60Ever2DerogPubRec</th>\n",
       "      <th>NumTrades90Ever2DerogPubRec</th>\n",
       "      <th>PercentTradesNeverDelq</th>\n",
       "      <th>MSinceMostRecentDelq</th>\n",
       "      <th>MaxDelq2PublicRecLast12M</th>\n",
       "      <th>...</th>\n",
       "      <th>MSinceMostRecentInqexcl7days</th>\n",
       "      <th>NumInqLast6M</th>\n",
       "      <th>NumInqLast6Mexcl7days</th>\n",
       "      <th>NetFractionRevolvingBurden</th>\n",
       "      <th>NetFractionInstallBurden</th>\n",
       "      <th>NumRevolvingTradesWBalance</th>\n",
       "      <th>NumInstallTradesWBalance</th>\n",
       "      <th>NumBank2NatlTradesWHighUtilization</th>\n",
       "      <th>PercentTradesWBalance</th>\n",
       "      <th>RiskPerformance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>55.0</td>\n",
       "      <td>144.0</td>\n",
       "      <td>4</td>\n",
       "      <td>84</td>\n",
       "      <td>20</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>83</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>33.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>69.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>61.0</td>\n",
       "      <td>58.0</td>\n",
       "      <td>15</td>\n",
       "      <td>41</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>100</td>\n",
       "      <td>15.0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>67.0</td>\n",
       "      <td>66.0</td>\n",
       "      <td>5</td>\n",
       "      <td>24</td>\n",
       "      <td>9</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>100</td>\n",
       "      <td>15.0</td>\n",
       "      <td>7</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>53.0</td>\n",
       "      <td>66.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>86.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>66.0</td>\n",
       "      <td>169.0</td>\n",
       "      <td>1</td>\n",
       "      <td>73</td>\n",
       "      <td>28</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>93</td>\n",
       "      <td>76.0</td>\n",
       "      <td>6</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5</td>\n",
       "      <td>4</td>\n",
       "      <td>72.0</td>\n",
       "      <td>83.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>91.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>81.0</td>\n",
       "      <td>333.0</td>\n",
       "      <td>27</td>\n",
       "      <td>132</td>\n",
       "      <td>12</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>100</td>\n",
       "      <td>15.0</td>\n",
       "      <td>7</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>51.0</td>\n",
       "      <td>89.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10454</th>\n",
       "      <td>73.0</td>\n",
       "      <td>131.0</td>\n",
       "      <td>5</td>\n",
       "      <td>57</td>\n",
       "      <td>21</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>95</td>\n",
       "      <td>80.0</td>\n",
       "      <td>6</td>\n",
       "      <td>...</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>26.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>100.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10455</th>\n",
       "      <td>65.0</td>\n",
       "      <td>147.0</td>\n",
       "      <td>39</td>\n",
       "      <td>68</td>\n",
       "      <td>11</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>92</td>\n",
       "      <td>28.0</td>\n",
       "      <td>6</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>86.0</td>\n",
       "      <td>53.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10456</th>\n",
       "      <td>74.0</td>\n",
       "      <td>129.0</td>\n",
       "      <td>6</td>\n",
       "      <td>64</td>\n",
       "      <td>18</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>100</td>\n",
       "      <td>15.0</td>\n",
       "      <td>6</td>\n",
       "      <td>...</td>\n",
       "      <td>3.0</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>6.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>56.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10457</th>\n",
       "      <td>72.0</td>\n",
       "      <td>234.0</td>\n",
       "      <td>12</td>\n",
       "      <td>113</td>\n",
       "      <td>42</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>96</td>\n",
       "      <td>35.0</td>\n",
       "      <td>6</td>\n",
       "      <td>...</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10458</th>\n",
       "      <td>66.0</td>\n",
       "      <td>28.0</td>\n",
       "      <td>1</td>\n",
       "      <td>17</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>100</td>\n",
       "      <td>15.0</td>\n",
       "      <td>7</td>\n",
       "      <td>...</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>67.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>100.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>9871 rows × 24 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       ExternalRiskEstimate  MSinceOldestTradeOpen  MSinceMostRecentTradeOpen  \\\n",
       "0                      55.0                  144.0                          4   \n",
       "1                      61.0                   58.0                         15   \n",
       "2                      67.0                   66.0                          5   \n",
       "3                      66.0                  169.0                          1   \n",
       "4                      81.0                  333.0                         27   \n",
       "...                     ...                    ...                        ...   \n",
       "10454                  73.0                  131.0                          5   \n",
       "10455                  65.0                  147.0                         39   \n",
       "10456                  74.0                  129.0                          6   \n",
       "10457                  72.0                  234.0                         12   \n",
       "10458                  66.0                   28.0                          1   \n",
       "\n",
       "       AverageMInFile  NumSatisfactoryTrades  NumTrades60Ever2DerogPubRec  \\\n",
       "0                  84                     20                            3   \n",
       "1                  41                      2                            4   \n",
       "2                  24                      9                            0   \n",
       "3                  73                     28                            1   \n",
       "4                 132                     12                            0   \n",
       "...               ...                    ...                          ...   \n",
       "10454              57                     21                            0   \n",
       "10455              68                     11                            0   \n",
       "10456              64                     18                            1   \n",
       "10457             113                     42                            2   \n",
       "10458              17                      4                            0   \n",
       "\n",
       "       NumTrades90Ever2DerogPubRec  PercentTradesNeverDelq  \\\n",
       "0                                0                      83   \n",
       "1                                4                     100   \n",
       "2                                0                     100   \n",
       "3                                1                      93   \n",
       "4                                0                     100   \n",
       "...                            ...                     ...   \n",
       "10454                            0                      95   \n",
       "10455                            0                      92   \n",
       "10456                            1                     100   \n",
       "10457                            2                      96   \n",
       "10458                            0                     100   \n",
       "\n",
       "       MSinceMostRecentDelq  MaxDelq2PublicRecLast12M  ...  \\\n",
       "0                       2.0                         3  ...   \n",
       "1                      15.0                         0  ...   \n",
       "2                      15.0                         7  ...   \n",
       "3                      76.0                         6  ...   \n",
       "4                      15.0                         7  ...   \n",
       "...                     ...                       ...  ...   \n",
       "10454                  80.0                         6  ...   \n",
       "10455                  28.0                         6  ...   \n",
       "10456                  15.0                         6  ...   \n",
       "10457                  35.0                         6  ...   \n",
       "10458                  15.0                         7  ...   \n",
       "\n",
       "       MSinceMostRecentInqexcl7days  NumInqLast6M  NumInqLast6Mexcl7days  \\\n",
       "0                               0.0             0                      0   \n",
       "1                               0.0             0                      0   \n",
       "2                               0.0             4                      4   \n",
       "3                               0.0             5                      4   \n",
       "4                               0.0             1                      1   \n",
       "...                             ...           ...                    ...   \n",
       "10454                           7.0             0                      0   \n",
       "10455                           1.0             1                      1   \n",
       "10456                           3.0             4                      4   \n",
       "10457                           6.0             0                      0   \n",
       "10458                           3.0             3                      2   \n",
       "\n",
       "       NetFractionRevolvingBurden  NetFractionInstallBurden  \\\n",
       "0                            33.0                      74.0   \n",
       "1                             0.0                      74.0   \n",
       "2                            53.0                      66.0   \n",
       "3                            72.0                      83.0   \n",
       "4                            51.0                      89.0   \n",
       "...                           ...                       ...   \n",
       "10454                        26.0                      74.0   \n",
       "10455                        86.0                      53.0   \n",
       "10456                         6.0                      74.0   \n",
       "10457                        19.0                      74.0   \n",
       "10458                        67.0                      74.0   \n",
       "\n",
       "       NumRevolvingTradesWBalance  NumInstallTradesWBalance  \\\n",
       "0                             8.0                       1.0   \n",
       "1                             0.0                       2.0   \n",
       "2                             4.0                       2.0   \n",
       "3                             6.0                       4.0   \n",
       "4                             3.0                       1.0   \n",
       "...                           ...                       ...   \n",
       "10454                         5.0                       2.0   \n",
       "10455                         2.0                       2.0   \n",
       "10456                         5.0                       2.0   \n",
       "10457                         4.0                       1.0   \n",
       "10458                         2.0                       1.0   \n",
       "\n",
       "       NumBank2NatlTradesWHighUtilization  PercentTradesWBalance  \\\n",
       "0                                     1.0                   69.0   \n",
       "1                                     1.0                    0.0   \n",
       "2                                     1.0                   86.0   \n",
       "3                                     3.0                   91.0   \n",
       "4                                     0.0                   80.0   \n",
       "...                                   ...                    ...   \n",
       "10454                                 0.0                  100.0   \n",
       "10455                                 1.0                   80.0   \n",
       "10456                                 0.0                   56.0   \n",
       "10457                                 0.0                   38.0   \n",
       "10458                                 0.0                  100.0   \n",
       "\n",
       "       RiskPerformance  \n",
       "0                    0  \n",
       "1                    0  \n",
       "2                    0  \n",
       "3                    0  \n",
       "4                    0  \n",
       "...                ...  \n",
       "10454                1  \n",
       "10455                0  \n",
       "10456                0  \n",
       "10457                0  \n",
       "10458                0  \n",
       "\n",
       "[9871 rows x 24 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "name = 'heloc'\n",
    "dropped_features = []\n",
    "# drop continuous features with:\n",
    "# dataset_loader().continuous_features[dataset]\n",
    "dataset = dataset_loader(name, dropped_features=dropped_features, n_bins=None)\n",
    "print(bold('{} Dataset:'.format(name.replace('_', ' ').title())))\n",
    "dataset.data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "e824d845-2d2f-46dd-8faa-87afc80f4bab",
   "metadata": {
    "scrolled": true,
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(0, 'ExternalRiskEstimate')\n",
      "(1, 'MSinceOldestTradeOpen')\n",
      "(2, 'MSinceMostRecentTradeOpen')\n",
      "(3, 'AverageMInFile')\n",
      "(4, 'NumSatisfactoryTrades')\n",
      "(5, 'NumTrades60Ever2DerogPubRec')\n",
      "(6, 'NumTrades90Ever2DerogPubRec')\n",
      "(7, 'PercentTradesNeverDelq')\n",
      "(8, 'MSinceMostRecentDelq')\n",
      "(9, 'MaxDelq2PublicRecLast12M')\n",
      "(10, 'MaxDelqEver')\n",
      "(11, 'NumTotalTrades')\n",
      "(12, 'NumTradesOpeninLast12M')\n",
      "(13, 'PercentInstallTrades')\n",
      "(14, 'MSinceMostRecentInqexcl7days')\n",
      "(15, 'NumInqLast6M')\n",
      "(16, 'NumInqLast6Mexcl7days')\n",
      "(17, 'NetFractionRevolvingBurden')\n",
      "(18, 'NetFractionInstallBurden')\n",
      "(19, 'NumRevolvingTradesWBalance')\n",
      "(20, 'NumInstallTradesWBalance')\n",
      "(21, 'NumBank2NatlTradesWHighUtilization')\n"
     ]
    }
   ],
   "source": [
    "# Inspect features\n",
    "features = dataset.features[:-1]\n",
    "print(*zip(range(len(features)-1), features[:-1]), sep='\\n')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "67af9cee-5511-416a-b737-068d614ed177",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "# Logistic Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "e8f9a131-70e0-4ebf-9134-177c2e35c624",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1mProportion of 1s in Training Data:\u001b[0m 70.12%\n",
      "\u001b[1mProportion of 1s in Test Data:\u001b[0m 69.5%\n"
     ]
    }
   ],
   "source": [
    "from models import lr_normalisers, lr_class_weights, lr_max_iters\n",
    "x_train, y_train, x_test, y_test, x_means, x_stds =\\\n",
    "dataset.get_split(normalise=lr_normalisers[name], shuffle=False,  # do not shuffle, important for repeatability\n",
    "                  return_mean_std=True, print_outputs=True)\n",
    "normalise = [x_means, x_stds] if lr_normalisers[name] else None"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "dd98f71f-65e9-49f3-9a98-631313c2b5e1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1mDataset:\u001b[0m German Credit\n",
      "\n",
      "\u001b[1mTrain Accuracy:\u001b[0m 78.88%\n",
      "\u001b[1mTest Accuracy:\u001b[0m 75.5%\n",
      "\u001b[1mProportion of 1s Predicted (Train):\u001b[0m 77.5%\n",
      "\u001b[1mProportion of 1s Predicted (Test):\u001b[0m 71.0%\n",
      "\n",
      "\u001b[1mNormalise:\u001b[0m False\n",
      "\u001b[1mDNN Negative Predictions (Train):\u001b[0m 22.5% (180 Inputs out of 800)\n",
      "\u001b[1mData Negative Predictions (Train):\u001b[0m 29.88% (239 Inputs out of 800)\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegression  \n",
    "print(\"\\033[1mDataset:\\033[0m {}\\n\".format(name.replace('_', ' ').title()))\n",
    "B = LogisticRegression(max_iter=lr_max_iters[name], class_weight=lr_class_weights[name])\n",
    "B.fit(x_train, y_train)\n",
    "B_name = 'lr'\n",
    "performance(B, x_train, x_test, y_train, y_test, normalise=None)  # already normalised where necessary\n",
    "\n",
    "norm = False if normalise is None else True\n",
    "print(\"\\n\\033[1mNormalise:\\033[0m {}\".format(norm))\n",
    "x_aff_size = (1-B.predict(x_train)).sum()\n",
    "x_size = x_train.shape[0]\n",
    "y_aff_size = (1-y_train).astype(int).sum()\n",
    "print(\"\\033[1mDNN Negative Predictions (Train):\\033[0m {}% ({} Inputs out of {})\"\\\n",
    "      .format(round(x_aff_size/x_size*100, 2), x_aff_size, x_size))\n",
    "print(\"\\033[1mData Negative Predictions (Train):\\033[0m {}% ({} Inputs out of {})\"\\\n",
    "      .format(round(y_aff_size/x_size*100, 2), y_aff_size, x_size))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "d51f4e1a-1a69-41b9-a66a-f50bbbd0da18",
   "metadata": {},
   "outputs": [],
   "source": [
    "if False:  # replace False with True to save model in the desired location\n",
    "    with open('models/{}_{}.pkl'.format(name, B_name), 'wb') as f:\n",
    "        pickle.dump(B, f)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0933bc73-39b7-4b56-9159-7b0a5cb789ac",
   "metadata": {
    "tags": []
   },
   "source": [
    "# XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "12f34829-1323-4612-a331-667a413b036e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1mProportion of 1s in Training Data:\u001b[0m 47.19%\n",
      "\u001b[1mProportion of 1s in Test Data:\u001b[0m 51.09%\n"
     ]
    }
   ],
   "source": [
    "from xgboost import XGBClassifier\n",
    "from models import xgb_depths, xgb_estimators, xgb_regularizers, xgb_learning_rates, xgb_colsamples\n",
    "\n",
    "x_train, y_train, x_test, y_test, x_means, x_stds =\\\n",
    "dataset.get_split(normalise=False, shuffle=False,  # do not shuffle, important for repeatability\n",
    "                  return_mean_std=True, print_outputs=True)\n",
    "\n",
    "# alternatively compute best params per dataset using grid search (below)\n",
    "gamma, reg_alpha, reg_lambda = xgb_regularizers[name]\n",
    "best_params = {'max_depth': xgb_depths[name], 'n_estimators': xgb_estimators[name],\n",
    "               'gamma': gamma, 'reg_alpha': reg_alpha, 'reg_lambda': reg_lambda,\n",
    "               'learning_rate': xgb_learning_rates[name], 'colsample_bytree': xgb_colsamples[name]}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "b85bdc51-fc60-4c70-b50f-a1db0073432e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1mDataset:\u001b[0m Heloc\n",
      "\n",
      "\u001b[1mTrain Accuracy:\u001b[0m 77.22%\n",
      "\u001b[1mTest Accuracy:\u001b[0m 73.77%\n",
      "\u001b[1mProportion of 1s Predicted (Train):\u001b[0m 45.25%\n",
      "\u001b[1mProportion of 1s Predicted (Test):\u001b[0m 46.63%\n",
      "\n",
      "\u001b[1mNormalise:\u001b[0m False\n",
      "\u001b[1mDNN Negative Predictions (Train):\u001b[0m 54.75% (4323 Inputs out of 7896)\n",
      "\u001b[1mData Negative Predictions (Train):\u001b[0m 52.81% (4170 Inputs out of 7896)\n"
     ]
    }
   ],
   "source": [
    "print(bold(\"Dataset:\") + \" {}\\n\".format(name.replace('_', ' ').title()))\n",
    "\n",
    "B = XGBClassifier(**best_params)\n",
    "B.fit(x_train, y_train)\n",
    "B_name = 'xgb'\n",
    "performance(B, x_train, x_test, y_train, y_test, normalise=None)  # no normalisation in xgb\n",
    "normalise = None\n",
    "\n",
    "norm = False if normalise is None else True\n",
    "print(\"\\n\\033[1mNormalise:\\033[0m {}\".format(norm))\n",
    "x_aff_size = (1-B.predict(x_train)).sum()\n",
    "x_size = x_train.shape[0]\n",
    "y_aff_size = (1-y_train).astype(int).sum()\n",
    "print(\"\\033[1mDNN Negative Predictions (Train):\\033[0m {}% ({} Inputs out of {})\"\\\n",
    "      .format(round(x_aff_size/x_size*100, 2), x_aff_size, x_size))\n",
    "print(\"\\033[1mData Negative Predictions (Train):\\033[0m {}% ({} Inputs out of {})\"\\\n",
    "      .format(round(y_aff_size/x_size*100, 2), y_aff_size, x_size))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "2aa4f0cc-7b1f-427e-9687-0a9f9e7a56bb",
   "metadata": {},
   "outputs": [],
   "source": [
    "if False:  # replace False with True to save model in the desired location\n",
    "    with open('models/{}_{}.pkl'.format(name, B_name), 'wb') as f:\n",
    "        pickle.dump(B, f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "cf1e07cf-5ba5-4602-a0e7-2945e31c5c8b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', callbacks=None,\n",
       "              colsample_bylevel=1, colsample_bynode=1, colsample_bytree=1,\n",
       "              early_stopping_rounds=None, enable_categorical=False,\n",
       "              eval_metric=None, feature_types=None, gamma=4, gpu_id=-1,\n",
       "              grow_policy='depthwise', importance_type=None,\n",
       "              interaction_constraints='', learning_rate=0.3, max_bin=256,\n",
       "              max_cat_threshold=64, max_cat_to_onehot=4, max_delta_step=0,\n",
       "              max_depth=6, max_leaves=0, min_child_weight=1, missing=nan,\n",
       "              monotone_constraints='()', n_estimators=100, n_jobs=0,\n",
       "              num_parallel_tree=1, predictor='auto', random_state=0, ...)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "B"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "29018ce4-aa89-4446-b873-f97635ad9428",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "### Grid Search (Used to find the hyperparameters stored in models.py)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "93c70ef9-7a7d-4de0-9fad-91254ddc4db3",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import GridSearchCV\n",
    "params = {'max_depth': [3,6,10],\n",
    "          'learning_rate': [0.01, 0.1, 0.5],\n",
    "          'n_estimators': [100, 500, 1000],\n",
    "          'colsample_bytree': [0.3, 0.7]}\n",
    "xgbr = XGBClassifier(seed = 20)\n",
    "clf = GridSearchCV(estimator=xgbr, param_grid=params,\n",
    "                   scoring='neg_mean_squared_error', verbose=1)\n",
    "clf.fit(x_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8f1b9ccc-938e-42d6-960b-de1aea439d45",
   "metadata": {},
   "outputs": [],
   "source": [
    "B = XGBClassifier(**clf.best_params_)\n",
    "B.fit(x_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5494982d-414c-4d1d-97f0-4974f6239a70",
   "metadata": {},
   "outputs": [],
   "source": [
    "performance(B, x_train, x_test, y_train, y_test, normalise=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "acf943dc-475c-49b9-a2e2-e5e7b1533758",
   "metadata": {},
   "source": [
    "# DNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "993e2e1f-cddb-4b8c-9738-494fb18c476c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1mDataset:\u001b[0m German Credit\n",
      "\u001b[1mNormalise:\u001b[0m True\n",
      "\u001b[1mBatch Size:\u001b[0m 128\n"
     ]
    }
   ],
   "source": [
    "import torch\n",
    "from models import dnn_normalisers, dnn_batch_sizes, dnn_input_output_dims,\\\n",
    "dnn_widths_depths, dnn_dropouts, dnn_criteria, dnn_learning_rates,\\\n",
    "dnn_weight_decays, dnn_num_epochs\n",
    "normalise = True#dnn_normalisers[name]  # set to True if necessary for model performance\n",
    "batch_size = dnn_batch_sizes[name]\n",
    "print(\"\\033[1mDataset:\\033[0m {}\".format(name.replace('_', ' ').title()))\n",
    "print(\"\\033[1mNormalise:\\033[0m {}\".format(normalise))\n",
    "print(\"\\033[1mBatch Size:\\033[0m {}\".format(batch_size))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "88fb13de-20ca-4632-bfa7-21a2fffb2d05",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1mProportion of 1s in Training Data:\u001b[0m 70.12%\n",
      "\u001b[1mProportion of 1s in Test Data:\u001b[0m 69.5%\n"
     ]
    }
   ],
   "source": [
    "normalise = dnn_normalisers[name]\n",
    "x_train, y_train, x_test, y_test, x_means, x_stds =\\\n",
    "dataset.get_split(normalise=normalise, shuffle=False,  # do not shuffle, important for repeatability\n",
    "                  return_mean_std=True, print_outputs=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cf1fe301-2dcd-415d-b361-f494d9ee321e",
   "metadata": {
    "tags": []
   },
   "source": [
    "### Construct DNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "ed8b30b9",
   "metadata": {},
   "outputs": [],
   "source": [
    "from models import DNN\n",
    "\n",
    "# Hyperparameters \n",
    "input_dim, output_dim = dnn_input_output_dims[name]\n",
    "width, depth = dnn_widths_depths[name]\n",
    "dropout = dnn_dropouts[name]\n",
    "criterion = dnn_criteria[name]\n",
    "\n",
    "# Device configuration\n",
    "cuda = False  # torch.cuda.is_available()\n",
    "device = 'cuda' if cuda else 'cpu'\n",
    "\n",
    "torch.manual_seed(0)\n",
    "B = DNN(input_dim, width, depth, output_dim, dropout=dropout).to(device)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "06dfcef2-823f-4112-be73-9c354fbe97ca",
   "metadata": {},
   "source": [
    "### Train DNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "37278081",
   "metadata": {
    "scrolled": true,
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of Epochs = 276\n",
      "Number of Steps per Epoch = ceil(Length of Trainset / Batch Size) = 7\n",
      "Epoch [1/276], Step [7/7], Loss: 4.8243, Time: 0.10s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [2/276], Step [7/7], Loss: 4.7892, Time: 0.14s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [3/276], Step [7/7], Loss: 4.7939, Time: 0.17s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [4/276], Step [7/7], Loss: 4.7660, Time: 0.21s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [5/276], Step [7/7], Loss: 4.7771, Time: 0.25s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [6/276], Step [7/7], Loss: 4.7737, Time: 0.29s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [7/276], Step [7/7], Loss: 4.7467, Time: 0.32s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [8/276], Step [7/7], Loss: 4.7236, Time: 0.35s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [9/276], Step [7/7], Loss: 4.7153, Time: 0.39s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [10/276], Step [7/7], Loss: 4.7176, Time: 0.43s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [11/276], Step [7/7], Loss: 4.7016, Time: 0.47s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [12/276], Step [7/7], Loss: 4.6895, Time: 0.50s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [13/276], Step [7/7], Loss: 4.6744, Time: 0.53s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [14/276], Step [7/7], Loss: 4.6761, Time: 0.56s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [15/276], Step [7/7], Loss: 4.6585, Time: 0.60s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [16/276], Step [7/7], Loss: 4.6575, Time: 0.64s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [17/276], Step [7/7], Loss: 4.6437, Time: 0.67s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [18/276], Step [7/7], Loss: 4.6139, Time: 0.71s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [19/276], Step [7/7], Loss: 4.6238, Time: 0.74s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [20/276], Step [7/7], Loss: 4.5886, Time: 0.77s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [21/276], Step [7/7], Loss: 4.5962, Time: 0.82s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [22/276], Step [7/7], Loss: 4.5777, Time: 0.86s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [23/276], Step [7/7], Loss: 4.5618, Time: 0.90s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [24/276], Step [7/7], Loss: 4.5452, Time: 0.93s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [25/276], Step [7/7], Loss: 4.5655, Time: 0.96s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [26/276], Step [7/7], Loss: 4.5338, Time: 0.99s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [27/276], Step [7/7], Loss: 4.5318, Time: 1.03s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [28/276], Step [7/7], Loss: 4.5542, Time: 1.08s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [29/276], Step [7/7], Loss: 4.5110, Time: 1.11s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [30/276], Step [7/7], Loss: 4.4568, Time: 1.14s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [31/276], Step [7/7], Loss: 4.4776, Time: 1.17s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [32/276], Step [7/7], Loss: 4.4878, Time: 1.21s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [33/276], Step [7/7], Loss: 4.4846, Time: 1.24s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [34/276], Step [7/7], Loss: 4.4499, Time: 1.28s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [35/276], Step [7/7], Loss: 4.4336, Time: 1.32s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [36/276], Step [7/7], Loss: 4.3913, Time: 1.35s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [37/276], Step [7/7], Loss: 4.4516, Time: 1.39s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [38/276], Step [7/7], Loss: 4.3859, Time: 1.42s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [39/276], Step [7/7], Loss: 4.3763, Time: 1.46s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [40/276], Step [7/7], Loss: 4.4224, Time: 1.50s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [41/276], Step [7/7], Loss: 4.4111, Time: 1.53s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [42/276], Step [7/7], Loss: 4.3676, Time: 1.56s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [43/276], Step [7/7], Loss: 4.4243, Time: 1.60s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [44/276], Step [7/7], Loss: 4.3488, Time: 1.63s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [45/276], Step [7/7], Loss: 4.3892, Time: 1.69s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [46/276], Step [7/7], Loss: 4.3535, Time: 1.75s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [47/276], Step [7/7], Loss: 4.3760, Time: 1.78s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [48/276], Step [7/7], Loss: 4.3172, Time: 1.81s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [49/276], Step [7/7], Loss: 4.2981, Time: 1.85s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [50/276], Step [7/7], Loss: 4.3520, Time: 1.89s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [51/276], Step [7/7], Loss: 4.3154, Time: 1.92s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [52/276], Step [7/7], Loss: 4.3880, Time: 1.95s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [53/276], Step [7/7], Loss: 4.3064, Time: 1.99s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [54/276], Step [7/7], Loss: 4.3406, Time: 2.02s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [55/276], Step [7/7], Loss: 4.2910, Time: 2.06s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [56/276], Step [7/7], Loss: 4.3399, Time: 2.10s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [57/276], Step [7/7], Loss: 4.2882, Time: 2.14s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [58/276], Step [7/7], Loss: 4.3243, Time: 2.17s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [59/276], Step [7/7], Loss: 4.2667, Time: 2.20s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [60/276], Step [7/7], Loss: 4.3178, Time: 2.23s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [61/276], Step [7/7], Loss: 4.3907, Time: 2.28s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [62/276], Step [7/7], Loss: 4.3444, Time: 2.32s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [63/276], Step [7/7], Loss: 4.2867, Time: 2.35s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [64/276], Step [7/7], Loss: 4.3356, Time: 2.38s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [65/276], Step [7/7], Loss: 4.3647, Time: 2.41s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [66/276], Step [7/7], Loss: 4.3413, Time: 2.45s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [67/276], Step [7/7], Loss: 4.3478, Time: 2.50s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [68/276], Step [7/7], Loss: 4.3435, Time: 2.53s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [69/276], Step [7/7], Loss: 4.3530, Time: 2.57s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [70/276], Step [7/7], Loss: 4.2991, Time: 2.60s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [71/276], Step [7/7], Loss: 4.2369, Time: 2.63s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [72/276], Step [7/7], Loss: 4.3138, Time: 2.68s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [73/276], Step [7/7], Loss: 4.3335, Time: 2.73s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [74/276], Step [7/7], Loss: 4.3421, Time: 2.76s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [75/276], Step [7/7], Loss: 4.3090, Time: 2.80s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [76/276], Step [7/7], Loss: 4.3620, Time: 2.83s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [77/276], Step [7/7], Loss: 4.3275, Time: 2.86s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [78/276], Step [7/7], Loss: 4.3220, Time: 2.91s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [79/276], Step [7/7], Loss: 4.3968, Time: 2.95s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [80/276], Step [7/7], Loss: 4.3583, Time: 2.98s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [81/276], Step [7/7], Loss: 4.3068, Time: 3.01s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [82/276], Step [7/7], Loss: 4.3320, Time: 3.05s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [83/276], Step [7/7], Loss: 4.3236, Time: 3.10s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [84/276], Step [7/7], Loss: 4.3212, Time: 3.15s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [85/276], Step [7/7], Loss: 4.3247, Time: 3.20s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [86/276], Step [7/7], Loss: 4.3136, Time: 3.24s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [87/276], Step [7/7], Loss: 4.3067, Time: 3.27s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [88/276], Step [7/7], Loss: 4.2927, Time: 3.31s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [89/276], Step [7/7], Loss: 4.2777, Time: 3.35s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [90/276], Step [7/7], Loss: 4.2958, Time: 3.38s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [91/276], Step [7/7], Loss: 4.2981, Time: 3.42s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [92/276], Step [7/7], Loss: 4.2834, Time: 3.46s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [93/276], Step [7/7], Loss: 4.3035, Time: 3.50s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [94/276], Step [7/7], Loss: 4.2411, Time: 3.55s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [95/276], Step [7/7], Loss: 4.2731, Time: 3.58s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [96/276], Step [7/7], Loss: 4.2482, Time: 3.62s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [97/276], Step [7/7], Loss: 4.2851, Time: 3.66s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [98/276], Step [7/7], Loss: 4.2706, Time: 3.69s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [99/276], Step [7/7], Loss: 4.2786, Time: 3.74s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [100/276], Step [7/7], Loss: 4.3037, Time: 3.78s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [101/276], Step [7/7], Loss: 4.3077, Time: 3.81s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [102/276], Step [7/7], Loss: 4.3069, Time: 3.84s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [103/276], Step [7/7], Loss: 4.2753, Time: 3.88s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [104/276], Step [7/7], Loss: 4.3433, Time: 3.92s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [105/276], Step [7/7], Loss: 4.2735, Time: 3.97s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [106/276], Step [7/7], Loss: 4.3108, Time: 4.00s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [107/276], Step [7/7], Loss: 4.2872, Time: 4.03s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [108/276], Step [7/7], Loss: 4.3180, Time: 4.06s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [109/276], Step [7/7], Loss: 4.2796, Time: 4.10s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [110/276], Step [7/7], Loss: 4.2658, Time: 4.16s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [111/276], Step [7/7], Loss: 4.2282, Time: 4.19s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [112/276], Step [7/7], Loss: 4.2638, Time: 4.22s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [113/276], Step [7/7], Loss: 4.2637, Time: 4.26s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [114/276], Step [7/7], Loss: 4.3187, Time: 4.29s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [115/276], Step [7/7], Loss: 4.3330, Time: 4.34s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [116/276], Step [7/7], Loss: 4.2994, Time: 4.38s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [117/276], Step [7/7], Loss: 4.3117, Time: 4.41s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [118/276], Step [7/7], Loss: 4.2947, Time: 4.45s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [119/276], Step [7/7], Loss: 4.3413, Time: 4.50s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [120/276], Step [7/7], Loss: 4.3138, Time: 4.56s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [121/276], Step [7/7], Loss: 4.2790, Time: 4.62s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [122/276], Step [7/7], Loss: 4.2473, Time: 4.68s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [123/276], Step [7/7], Loss: 4.2106, Time: 4.74s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [124/276], Step [7/7], Loss: 4.2856, Time: 4.81s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [125/276], Step [7/7], Loss: 4.3184, Time: 4.84s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [126/276], Step [7/7], Loss: 4.2888, Time: 4.88s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [127/276], Step [7/7], Loss: 4.2522, Time: 4.91s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [128/276], Step [7/7], Loss: 4.2640, Time: 4.96s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [129/276], Step [7/7], Loss: 4.2397, Time: 5.00s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [130/276], Step [7/7], Loss: 4.2717, Time: 5.04s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [131/276], Step [7/7], Loss: 4.2402, Time: 5.08s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [132/276], Step [7/7], Loss: 4.2938, Time: 5.12s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [133/276], Step [7/7], Loss: 4.2419, Time: 5.16s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [134/276], Step [7/7], Loss: 4.2803, Time: 5.20s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [135/276], Step [7/7], Loss: 4.2349, Time: 5.23s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [136/276], Step [7/7], Loss: 4.3279, Time: 5.27s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [137/276], Step [7/7], Loss: 4.3205, Time: 5.32s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [138/276], Step [7/7], Loss: 4.2413, Time: 5.37s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [139/276], Step [7/7], Loss: 4.2241, Time: 5.42s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [140/276], Step [7/7], Loss: 4.2791, Time: 5.47s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [141/276], Step [7/7], Loss: 4.3179, Time: 5.51s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [142/276], Step [7/7], Loss: 4.2898, Time: 5.55s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [143/276], Step [7/7], Loss: 4.2486, Time: 5.59s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [144/276], Step [7/7], Loss: 4.2605, Time: 5.63s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [145/276], Step [7/7], Loss: 4.2074, Time: 5.68s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [146/276], Step [7/7], Loss: 4.1626, Time: 5.72s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [147/276], Step [7/7], Loss: 4.2560, Time: 5.75s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [148/276], Step [7/7], Loss: 4.2186, Time: 5.80s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [149/276], Step [7/7], Loss: 4.2433, Time: 5.83s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [150/276], Step [7/7], Loss: 4.2167, Time: 5.87s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [151/276], Step [7/7], Loss: 4.2168, Time: 5.90s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [152/276], Step [7/7], Loss: 4.1233, Time: 5.93s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [153/276], Step [7/7], Loss: 4.1418, Time: 5.97s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [154/276], Step [7/7], Loss: 4.1285, Time: 6.04s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [155/276], Step [7/7], Loss: 4.1441, Time: 6.09s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [156/276], Step [7/7], Loss: 4.1386, Time: 6.12s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [157/276], Step [7/7], Loss: 4.1813, Time: 6.16s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [158/276], Step [7/7], Loss: 4.0204, Time: 6.21s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [159/276], Step [7/7], Loss: 4.1217, Time: 6.24s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [160/276], Step [7/7], Loss: 4.0198, Time: 6.28s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [161/276], Step [7/7], Loss: 4.0601, Time: 6.32s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [162/276], Step [7/7], Loss: 4.0141, Time: 6.36s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [163/276], Step [7/7], Loss: 4.1999, Time: 6.42s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [164/276], Step [7/7], Loss: 3.8138, Time: 6.47s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [165/276], Step [7/7], Loss: 3.8924, Time: 6.50s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [166/276], Step [7/7], Loss: 3.8639, Time: 6.54s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [167/276], Step [7/7], Loss: 3.8649, Time: 6.58s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [168/276], Step [7/7], Loss: 3.8078, Time: 6.62s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [169/276], Step [7/7], Loss: 3.9145, Time: 6.67s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [170/276], Step [7/7], Loss: 3.8384, Time: 6.72s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [171/276], Step [7/7], Loss: 3.8692, Time: 6.76s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [172/276], Step [7/7], Loss: 3.8223, Time: 6.81s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [173/276], Step [7/7], Loss: 3.8228, Time: 6.85s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [174/276], Step [7/7], Loss: 3.7891, Time: 6.89s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [175/276], Step [7/7], Loss: 3.7078, Time: 6.93s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [176/276], Step [7/7], Loss: 3.6313, Time: 6.96s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [177/276], Step [7/7], Loss: 3.8239, Time: 7.02s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [178/276], Step [7/7], Loss: 3.6349, Time: 7.06s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [179/276], Step [7/7], Loss: 3.7594, Time: 7.10s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [180/276], Step [7/7], Loss: 3.5810, Time: 7.13s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [181/276], Step [7/7], Loss: 3.6007, Time: 7.16s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [182/276], Step [7/7], Loss: 3.7685, Time: 7.20s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [183/276], Step [7/7], Loss: 3.5636, Time: 7.25s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [184/276], Step [7/7], Loss: 3.5816, Time: 7.28s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [185/276], Step [7/7], Loss: 3.5560, Time: 7.32s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [186/276], Step [7/7], Loss: 3.5940, Time: 7.35s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [187/276], Step [7/7], Loss: 3.5321, Time: 7.39s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [188/276], Step [7/7], Loss: 3.5693, Time: 7.43s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [189/276], Step [7/7], Loss: 3.5068, Time: 7.47s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [190/276], Step [7/7], Loss: 3.5900, Time: 7.50s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [191/276], Step [7/7], Loss: 3.4797, Time: 7.54s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [192/276], Step [7/7], Loss: 3.5478, Time: 7.57s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [193/276], Step [7/7], Loss: 3.6184, Time: 7.62s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [194/276], Step [7/7], Loss: 3.5239, Time: 7.67s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [195/276], Step [7/7], Loss: 3.5168, Time: 7.70s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [196/276], Step [7/7], Loss: 3.4954, Time: 7.73s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [197/276], Step [7/7], Loss: 3.4651, Time: 7.77s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [198/276], Step [7/7], Loss: 3.4410, Time: 7.82s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [199/276], Step [7/7], Loss: 3.3593, Time: 7.86s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [200/276], Step [7/7], Loss: 3.4059, Time: 7.90s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [201/276], Step [7/7], Loss: 3.4895, Time: 7.93s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [202/276], Step [7/7], Loss: 3.3937, Time: 7.96s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [203/276], Step [7/7], Loss: 3.3761, Time: 8.00s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [204/276], Step [7/7], Loss: 3.3689, Time: 8.05s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [205/276], Step [7/7], Loss: 3.4096, Time: 8.10s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [206/276], Step [7/7], Loss: 3.4062, Time: 8.15s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [207/276], Step [7/7], Loss: 3.4655, Time: 8.20s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [208/276], Step [7/7], Loss: 3.5639, Time: 8.24s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [209/276], Step [7/7], Loss: 3.2636, Time: 8.28s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [210/276], Step [7/7], Loss: 3.1948, Time: 8.33s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [211/276], Step [7/7], Loss: 3.3477, Time: 8.36s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [212/276], Step [7/7], Loss: 3.1940, Time: 8.40s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [213/276], Step [7/7], Loss: 3.4347, Time: 8.44s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [214/276], Step [7/7], Loss: 3.3264, Time: 8.49s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [215/276], Step [7/7], Loss: 3.1746, Time: 8.53s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [216/276], Step [7/7], Loss: 3.1395, Time: 8.57s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [217/276], Step [7/7], Loss: 3.1151, Time: 8.61s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [218/276], Step [7/7], Loss: 3.2196, Time: 8.66s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [219/276], Step [7/7], Loss: 3.1091, Time: 8.71s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [220/276], Step [7/7], Loss: 3.3134, Time: 8.75s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [221/276], Step [7/7], Loss: 3.3718, Time: 8.78s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [222/276], Step [7/7], Loss: 3.2473, Time: 8.82s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [223/276], Step [7/7], Loss: 3.3677, Time: 8.86s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [224/276], Step [7/7], Loss: 3.1909, Time: 8.90s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [225/276], Step [7/7], Loss: 3.1697, Time: 8.94s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [226/276], Step [7/7], Loss: 3.1388, Time: 8.98s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [227/276], Step [7/7], Loss: 3.2591, Time: 9.02s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [228/276], Step [7/7], Loss: 3.2886, Time: 9.06s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [229/276], Step [7/7], Loss: 3.2504, Time: 9.10s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [230/276], Step [7/7], Loss: 3.0706, Time: 9.14s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [231/276], Step [7/7], Loss: 3.1707, Time: 9.18s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [232/276], Step [7/7], Loss: 3.1379, Time: 9.21s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [233/276], Step [7/7], Loss: 3.0315, Time: 9.25s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [234/276], Step [7/7], Loss: 3.1747, Time: 9.29s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [235/276], Step [7/7], Loss: 3.1635, Time: 9.33s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [236/276], Step [7/7], Loss: 2.9867, Time: 9.37s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [237/276], Step [7/7], Loss: 2.9715, Time: 9.41s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [238/276], Step [7/7], Loss: 3.1597, Time: 9.45s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [239/276], Step [7/7], Loss: 3.1159, Time: 9.49s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [240/276], Step [7/7], Loss: 3.0088, Time: 9.53s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [241/276], Step [7/7], Loss: 3.0395, Time: 9.57s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [242/276], Step [7/7], Loss: 3.1250, Time: 9.61s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [243/276], Step [7/7], Loss: 2.9811, Time: 9.66s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [244/276], Step [7/7], Loss: 3.0708, Time: 9.70s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [245/276], Step [7/7], Loss: 2.8566, Time: 9.75s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [246/276], Step [7/7], Loss: 2.9032, Time: 9.79s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [247/276], Step [7/7], Loss: 2.9088, Time: 9.83s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [248/276], Step [7/7], Loss: 2.9626, Time: 9.88s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [249/276], Step [7/7], Loss: 2.7747, Time: 9.92s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [250/276], Step [7/7], Loss: 2.9639, Time: 9.95s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [251/276], Step [7/7], Loss: 2.8411, Time: 10.00s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [252/276], Step [7/7], Loss: 3.1944, Time: 10.06s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [253/276], Step [7/7], Loss: 2.8111, Time: 10.10s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [254/276], Step [7/7], Loss: 2.9208, Time: 10.14s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [255/276], Step [7/7], Loss: 2.9425, Time: 10.18s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [256/276], Step [7/7], Loss: 2.9214, Time: 10.23s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [257/276], Step [7/7], Loss: 2.9192, Time: 10.26s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [258/276], Step [7/7], Loss: 2.8336, Time: 10.30s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [259/276], Step [7/7], Loss: 2.9250, Time: 10.35s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [260/276], Step [7/7], Loss: 2.9265, Time: 10.39s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [261/276], Step [7/7], Loss: 2.7632, Time: 10.44s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [262/276], Step [7/7], Loss: 2.8709, Time: 10.49s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [263/276], Step [7/7], Loss: 2.8634, Time: 10.53s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [264/276], Step [7/7], Loss: 2.7522, Time: 10.57s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [265/276], Step [7/7], Loss: 2.8063, Time: 10.60s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [266/276], Step [7/7], Loss: 2.8021, Time: 10.64s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [267/276], Step [7/7], Loss: 2.7798, Time: 10.67s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [268/276], Step [7/7], Loss: 2.6424, Time: 10.72s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [269/276], Step [7/7], Loss: 2.7153, Time: 10.75s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [270/276], Step [7/7], Loss: 2.7511, Time: 10.79s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [271/276], Step [7/7], Loss: 2.6538, Time: 10.83s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [272/276], Step [7/7], Loss: 2.6277, Time: 10.86s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [273/276], Step [7/7], Loss: 2.5832, Time: 10.90s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [274/276], Step [7/7], Loss: 2.7556, Time: 10.95s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [275/276], Step [7/7], Loss: 2.7669, Time: 11.01s\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the training set: 70.125%\n",
      "Proportion of 1's predicted: 100.000%\n",
      "Accuracy of the network on the test set: 69.5%\n",
      "Epoch [276/276], Step [7/7], Loss: 2.7094, Time: 11.04s\n",
      "Proportion of 1's predicted: 62.000%\n",
      "Accuracy of the network on the training set: 86.125%\n",
      "Proportion of 1's predicted: 67.000%\n",
      "Accuracy of the network on the test set: 75.5%\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1100x600 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of Epochs = 45\n",
      "Number of Steps per Epoch = ceil(Length of Trainset / Batch Size) = 7\n",
      "Epoch [1/45], Step [7/7], Loss: 2.6398, Time: 0.06s\n",
      "Proportion of 1's predicted: 61.250%\n",
      "Accuracy of the network on the training set: 86.625%\n",
      "Proportion of 1's predicted: 67.000%\n",
      "Accuracy of the network on the test set: 75.5%\n",
      "Epoch [2/45], Step [7/7], Loss: 2.5820, Time: 0.12s\n",
      "Proportion of 1's predicted: 60.875%\n",
      "Accuracy of the network on the training set: 86.5%\n",
      "Proportion of 1's predicted: 66.500%\n",
      "Accuracy of the network on the test set: 75.0%\n",
      "Epoch [3/45], Step [7/7], Loss: 2.6867, Time: 0.16s\n",
      "Proportion of 1's predicted: 60.500%\n",
      "Accuracy of the network on the training set: 86.625%\n",
      "Proportion of 1's predicted: 65.000%\n",
      "Accuracy of the network on the test set: 75.5%\n",
      "Epoch [4/45], Step [7/7], Loss: 2.6111, Time: 0.21s\n",
      "Proportion of 1's predicted: 59.875%\n",
      "Accuracy of the network on the training set: 86.0%\n",
      "Proportion of 1's predicted: 65.000%\n",
      "Accuracy of the network on the test set: 75.5%\n",
      "Epoch [5/45], Step [7/7], Loss: 2.7310, Time: 0.27s\n",
      "Proportion of 1's predicted: 59.500%\n",
      "Accuracy of the network on the training set: 85.625%\n",
      "Proportion of 1's predicted: 64.000%\n",
      "Accuracy of the network on the test set: 74.5%\n",
      "Epoch [6/45], Step [7/7], Loss: 2.6747, Time: 0.32s\n",
      "Proportion of 1's predicted: 59.500%\n",
      "Accuracy of the network on the training set: 85.625%\n",
      "Proportion of 1's predicted: 64.000%\n",
      "Accuracy of the network on the test set: 74.5%\n",
      "Epoch [7/45], Step [7/7], Loss: 2.6917, Time: 0.36s\n",
      "Proportion of 1's predicted: 59.500%\n",
      "Accuracy of the network on the training set: 85.625%\n",
      "Proportion of 1's predicted: 64.000%\n",
      "Accuracy of the network on the test set: 74.5%\n",
      "Epoch [8/45], Step [7/7], Loss: 2.7470, Time: 0.40s\n",
      "Proportion of 1's predicted: 59.500%\n",
      "Accuracy of the network on the training set: 85.625%\n",
      "Proportion of 1's predicted: 64.000%\n",
      "Accuracy of the network on the test set: 74.5%\n",
      "Epoch [9/45], Step [7/7], Loss: 2.7169, Time: 0.44s\n",
      "Proportion of 1's predicted: 59.750%\n",
      "Accuracy of the network on the training set: 85.625%\n",
      "Proportion of 1's predicted: 64.000%\n",
      "Accuracy of the network on the test set: 74.5%\n",
      "Epoch [10/45], Step [7/7], Loss: 2.5935, Time: 0.48s\n",
      "Proportion of 1's predicted: 59.750%\n",
      "Accuracy of the network on the training set: 85.625%\n",
      "Proportion of 1's predicted: 65.000%\n",
      "Accuracy of the network on the test set: 75.5%\n",
      "Epoch [11/45], Step [7/7], Loss: 2.7779, Time: 0.52s\n",
      "Proportion of 1's predicted: 59.625%\n",
      "Accuracy of the network on the training set: 85.75%\n",
      "Proportion of 1's predicted: 64.000%\n",
      "Accuracy of the network on the test set: 74.5%\n",
      "Epoch [12/45], Step [7/7], Loss: 2.7410, Time: 0.56s\n",
      "Proportion of 1's predicted: 59.750%\n",
      "Accuracy of the network on the training set: 85.625%\n",
      "Proportion of 1's predicted: 64.000%\n",
      "Accuracy of the network on the test set: 74.5%\n",
      "Epoch [13/45], Step [7/7], Loss: 2.7761, Time: 0.59s\n",
      "Proportion of 1's predicted: 59.875%\n",
      "Accuracy of the network on the training set: 85.75%\n",
      "Proportion of 1's predicted: 64.500%\n",
      "Accuracy of the network on the test set: 75.0%\n",
      "Epoch [14/45], Step [7/7], Loss: 2.7723, Time: 0.63s\n",
      "Proportion of 1's predicted: 60.000%\n",
      "Accuracy of the network on the training set: 85.875%\n",
      "Proportion of 1's predicted: 65.000%\n",
      "Accuracy of the network on the test set: 75.5%\n",
      "Epoch [15/45], Step [7/7], Loss: 2.7514, Time: 0.68s\n",
      "Proportion of 1's predicted: 60.375%\n",
      "Accuracy of the network on the training set: 86.25%\n",
      "Proportion of 1's predicted: 65.500%\n",
      "Accuracy of the network on the test set: 76.0%\n",
      "Epoch [16/45], Step [7/7], Loss: 2.7109, Time: 0.72s\n",
      "Proportion of 1's predicted: 60.625%\n",
      "Accuracy of the network on the training set: 86.25%\n",
      "Proportion of 1's predicted: 65.500%\n",
      "Accuracy of the network on the test set: 76.0%\n",
      "Epoch [17/45], Step [7/7], Loss: 2.8100, Time: 0.75s\n",
      "Proportion of 1's predicted: 60.875%\n",
      "Accuracy of the network on the training set: 86.5%\n",
      "Proportion of 1's predicted: 66.000%\n",
      "Accuracy of the network on the test set: 76.5%\n",
      "Epoch [18/45], Step [7/7], Loss: 2.7848, Time: 0.78s\n",
      "Proportion of 1's predicted: 61.000%\n",
      "Accuracy of the network on the training set: 86.625%\n",
      "Proportion of 1's predicted: 66.000%\n",
      "Accuracy of the network on the test set: 76.5%\n",
      "Epoch [19/45], Step [7/7], Loss: 2.8240, Time: 0.82s\n",
      "Proportion of 1's predicted: 61.375%\n",
      "Accuracy of the network on the training set: 86.25%\n",
      "Proportion of 1's predicted: 66.000%\n",
      "Accuracy of the network on the test set: 76.5%\n",
      "Epoch [20/45], Step [7/7], Loss: 2.7686, Time: 0.87s\n",
      "Proportion of 1's predicted: 61.375%\n",
      "Accuracy of the network on the training set: 86.25%\n",
      "Proportion of 1's predicted: 66.000%\n",
      "Accuracy of the network on the test set: 76.5%\n",
      "Epoch [21/45], Step [7/7], Loss: 2.8268, Time: 0.91s\n",
      "Proportion of 1's predicted: 61.750%\n",
      "Accuracy of the network on the training set: 86.125%\n",
      "Proportion of 1's predicted: 66.000%\n",
      "Accuracy of the network on the test set: 76.5%\n",
      "Epoch [22/45], Step [7/7], Loss: 2.8071, Time: 0.95s\n",
      "Proportion of 1's predicted: 62.125%\n",
      "Accuracy of the network on the training set: 86.0%\n",
      "Proportion of 1's predicted: 66.500%\n",
      "Accuracy of the network on the test set: 76.0%\n",
      "Epoch [23/45], Step [7/7], Loss: 2.7745, Time: 0.98s\n",
      "Proportion of 1's predicted: 63.125%\n",
      "Accuracy of the network on the training set: 85.5%\n",
      "Proportion of 1's predicted: 67.500%\n",
      "Accuracy of the network on the test set: 76.0%\n",
      "Epoch [24/45], Step [7/7], Loss: 2.8349, Time: 1.02s\n",
      "Proportion of 1's predicted: 63.375%\n",
      "Accuracy of the network on the training set: 85.25%\n",
      "Proportion of 1's predicted: 67.500%\n",
      "Accuracy of the network on the test set: 76.0%\n",
      "Epoch [25/45], Step [7/7], Loss: 2.8209, Time: 1.07s\n",
      "Proportion of 1's predicted: 63.625%\n",
      "Accuracy of the network on the training set: 85.25%\n",
      "Proportion of 1's predicted: 67.500%\n",
      "Accuracy of the network on the test set: 76.0%\n",
      "Epoch [26/45], Step [7/7], Loss: 2.8822, Time: 1.11s\n",
      "Proportion of 1's predicted: 64.000%\n",
      "Accuracy of the network on the training set: 85.375%\n",
      "Proportion of 1's predicted: 67.500%\n",
      "Accuracy of the network on the test set: 76.0%\n",
      "Epoch [27/45], Step [7/7], Loss: 2.8461, Time: 1.15s\n",
      "Proportion of 1's predicted: 64.375%\n",
      "Accuracy of the network on the training set: 85.25%\n",
      "Proportion of 1's predicted: 68.000%\n",
      "Accuracy of the network on the test set: 75.5%\n",
      "Epoch [28/45], Step [7/7], Loss: 2.7535, Time: 1.19s\n",
      "Proportion of 1's predicted: 64.375%\n",
      "Accuracy of the network on the training set: 85.25%\n",
      "Proportion of 1's predicted: 68.000%\n",
      "Accuracy of the network on the test set: 75.5%\n",
      "Epoch [29/45], Step [7/7], Loss: 2.8996, Time: 1.22s\n",
      "Proportion of 1's predicted: 64.375%\n",
      "Accuracy of the network on the training set: 85.25%\n",
      "Proportion of 1's predicted: 68.000%\n",
      "Accuracy of the network on the test set: 75.5%\n",
      "Epoch [30/45], Step [7/7], Loss: 2.8385, Time: 1.27s\n",
      "Proportion of 1's predicted: 64.625%\n",
      "Accuracy of the network on the training set: 85.25%\n",
      "Proportion of 1's predicted: 68.000%\n",
      "Accuracy of the network on the test set: 75.5%\n",
      "Epoch [31/45], Step [7/7], Loss: 2.8255, Time: 1.31s\n",
      "Proportion of 1's predicted: 64.625%\n",
      "Accuracy of the network on the training set: 85.25%\n",
      "Proportion of 1's predicted: 68.000%\n",
      "Accuracy of the network on the test set: 75.5%\n",
      "Epoch [32/45], Step [7/7], Loss: 2.8658, Time: 1.34s\n",
      "Proportion of 1's predicted: 65.125%\n",
      "Accuracy of the network on the training set: 85.25%\n",
      "Proportion of 1's predicted: 68.500%\n",
      "Accuracy of the network on the test set: 76.0%\n",
      "Epoch [33/45], Step [7/7], Loss: 2.9090, Time: 1.37s\n",
      "Proportion of 1's predicted: 65.500%\n",
      "Accuracy of the network on the training set: 85.375%\n",
      "Proportion of 1's predicted: 68.500%\n",
      "Accuracy of the network on the test set: 76.0%\n",
      "Epoch [34/45], Step [7/7], Loss: 2.8166, Time: 1.41s\n",
      "Proportion of 1's predicted: 65.750%\n",
      "Accuracy of the network on the training set: 85.125%\n",
      "Proportion of 1's predicted: 69.000%\n",
      "Accuracy of the network on the test set: 75.5%\n",
      "Epoch [35/45], Step [7/7], Loss: 2.9641, Time: 1.44s\n",
      "Proportion of 1's predicted: 66.125%\n",
      "Accuracy of the network on the training set: 85.0%\n",
      "Proportion of 1's predicted: 69.000%\n",
      "Accuracy of the network on the test set: 75.5%\n",
      "Epoch [36/45], Step [7/7], Loss: 2.9098, Time: 1.49s\n",
      "Proportion of 1's predicted: 66.375%\n",
      "Accuracy of the network on the training set: 84.75%\n",
      "Proportion of 1's predicted: 69.000%\n",
      "Accuracy of the network on the test set: 75.5%\n",
      "Epoch [37/45], Step [7/7], Loss: 2.8448, Time: 1.52s\n",
      "Proportion of 1's predicted: 66.750%\n",
      "Accuracy of the network on the training set: 85.125%\n",
      "Proportion of 1's predicted: 70.000%\n",
      "Accuracy of the network on the test set: 76.5%\n",
      "Epoch [38/45], Step [7/7], Loss: 2.9051, Time: 1.56s\n",
      "Proportion of 1's predicted: 67.875%\n",
      "Accuracy of the network on the training set: 84.75%\n",
      "Proportion of 1's predicted: 70.000%\n",
      "Accuracy of the network on the test set: 76.5%\n",
      "Epoch [39/45], Step [7/7], Loss: 2.8943, Time: 1.60s\n",
      "Proportion of 1's predicted: 68.000%\n",
      "Accuracy of the network on the training set: 84.625%\n",
      "Proportion of 1's predicted: 70.000%\n",
      "Accuracy of the network on the test set: 76.5%\n",
      "Epoch [40/45], Step [7/7], Loss: 2.9439, Time: 1.64s\n",
      "Proportion of 1's predicted: 68.250%\n",
      "Accuracy of the network on the training set: 84.375%\n",
      "Proportion of 1's predicted: 70.000%\n",
      "Accuracy of the network on the test set: 76.5%\n",
      "Epoch [41/45], Step [7/7], Loss: 2.9881, Time: 1.69s\n",
      "Proportion of 1's predicted: 68.250%\n",
      "Accuracy of the network on the training set: 84.375%\n",
      "Proportion of 1's predicted: 70.000%\n",
      "Accuracy of the network on the test set: 76.5%\n",
      "Epoch [42/45], Step [7/7], Loss: 2.9393, Time: 1.72s\n",
      "Proportion of 1's predicted: 68.500%\n",
      "Accuracy of the network on the training set: 84.125%\n",
      "Proportion of 1's predicted: 70.500%\n",
      "Accuracy of the network on the test set: 77.0%\n",
      "Epoch [43/45], Step [7/7], Loss: 2.9232, Time: 1.76s\n",
      "Proportion of 1's predicted: 69.250%\n",
      "Accuracy of the network on the training set: 83.625%\n",
      "Proportion of 1's predicted: 71.000%\n",
      "Accuracy of the network on the test set: 77.5%\n",
      "Epoch [44/45], Step [7/7], Loss: 2.8737, Time: 1.79s\n",
      "Proportion of 1's predicted: 69.375%\n",
      "Accuracy of the network on the training set: 83.75%\n",
      "Proportion of 1's predicted: 71.000%\n",
      "Accuracy of the network on the test set: 77.5%\n",
      "Epoch [45/45], Step [7/7], Loss: 2.9665, Time: 1.83s\n",
      "Proportion of 1's predicted: 69.625%\n",
      "Accuracy of the network on the training set: 83.75%\n",
      "Proportion of 1's predicted: 71.500%\n",
      "Accuracy of the network on the test set: 78.0%\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1100x600 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from models import DNN_trainer\n",
    "\n",
    "inps = torch.from_numpy(x_train).float()\n",
    "tgts = torch.from_numpy(y_train).float()\n",
    "train_dataset = torch.utils.data.TensorDataset(inps, tgts)\n",
    "\n",
    "inps = torch.from_numpy(x_test).float()\n",
    "tgts = torch.from_numpy(y_test).float()\n",
    "test_dataset = torch.utils.data.TensorDataset(inps, tgts)\n",
    "\n",
    "\n",
    "for i, learning_rate in enumerate(dnn_learning_rates[name]):\n",
    "    # Initialise optimizer\n",
    "    optimizer = torch.optim.Adam(B.parameters(), lr=learning_rate,\n",
    "                                 weight_decay=dnn_weight_decays[name][i])\n",
    "\n",
    "    # Train model\n",
    "    n_epochs = dnn_num_epochs[name][i]\n",
    "    trainer = DNN_trainer(B, train_dataset, n_epochs, criterion, optimizer,\n",
    "                          device, batch_size=batch_size, val_ratio=None,\n",
    "                          test_dataset=test_dataset)\n",
    "    B = trainer.train(n_epochs, patience=None, print_outputs=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "369a42b5-f72c-4c3e-8e3d-0262372b942d",
   "metadata": {},
   "source": [
    "### Assess Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "7b12f25c-2bc7-4995-921b-f8b53f864922",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1mDataset:\u001b[0m German Credit\n",
      "\n",
      "\u001b[1mTrain Accuracy:\u001b[0m 83.75%\n",
      "\u001b[1mTest Accuracy:\u001b[0m 78.0%\n",
      "\u001b[1mProportion of 1s Predicted (Train):\u001b[0m 69.62%\n",
      "\u001b[1mProportion of 1s Predicted (Test):\u001b[0m 71.5%\n",
      "\n",
      "\u001b[1mWidth x Depth:\u001b[0m 50 x 10\n",
      "\u001b[1mNormalise:\u001b[0m True\n",
      "\u001b[1mDNN Negative Predictions (Train):\u001b[0m 30.38% (243 Inputs out of 800)\n",
      "\u001b[1mData Negative Predictions (Train):\u001b[0m 29.88% (239 Inputs out of 800)\n"
     ]
    }
   ],
   "source": [
    "print(\"\\033[1mDataset:\\033[0m {}\\n\".format(name.replace('_', ' ').title()))\n",
    "\n",
    "B_name = 'dnn'\n",
    "performance(B, x_train, x_test, y_train, y_test, normalise=None)  # already normalised where necessary\n",
    "\n",
    "print(\"\\n\\033[1mWidth x Depth:\\033[0m {} x {}\".format(width, depth))\n",
    "print(\"\\033[1mNormalise:\\033[0m {}\".format(normalise))\n",
    "x_aff_size = (1-B.predict(x_train)).sum()\n",
    "x_size = x_train.shape[0]\n",
    "y_aff_size = (1-y_train).astype(int).sum()\n",
    "print(\"\\033[1mDNN Negative Predictions (Train):\\033[0m {}% ({} Inputs out of {})\"\\\n",
    "      .format(round(x_aff_size/x_size*100, 2), x_aff_size, x_size))\n",
    "print(\"\\033[1mData Negative Predictions (Train):\\033[0m {}% ({} Inputs out of {})\"\\\n",
    "      .format(round(y_aff_size/x_size*100, 2), y_aff_size, x_size))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "618381c3-b6bb-4573-acfa-c4c108201a67",
   "metadata": {
    "tags": []
   },
   "source": [
    "### Save Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "da58d4b8-5747-4a04-a653-f089bbe0b642",
   "metadata": {},
   "outputs": [],
   "source": [
    "if False:  # replace False with True to save model in the desired location\n",
    "    with open('models/{}_{}.pkl'.format(name, B_name), 'wb') as f:\n",
    "        pickle.dump(B, f)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}