LBN_RECO_VBSpol.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "c157cd32-2e52-4c7e-9b91-05a41aec33bb",
   "metadata": {},
   "outputs": [
    {
     "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>Rpt</th>\n",
       "      <th>Zepp_l1</th>\n",
       "      <th>Zepp_l2</th>\n",
       "      <th>Zepp_ll</th>\n",
       "      <th>detajj</th>\n",
       "      <th>eta_j1</th>\n",
       "      <th>eta_j2</th>\n",
       "      <th>eta_l1</th>\n",
       "      <th>eta_l2</th>\n",
       "      <th>mjj</th>\n",
       "      <th>mll</th>\n",
       "      <th>njet</th>\n",
       "      <th>nlep</th>\n",
       "      <th>pdgid_l1</th>\n",
       "      <th>pdgid_l2</th>\n",
       "      <th>phi_j1</th>\n",
       "      <th>phi_j2</th>\n",
       "      <th>phi_l1</th>\n",
       "      <th>phi_l2</th>\n",
       "      <th>pt_j1</th>\n",
       "      <th>pt_j2</th>\n",
       "      <th>pt_l1</th>\n",
       "      <th>pt_l2</th>\n",
       "      <th>ptll</th>\n",
       "      <th>ptmiss</th>\n",
       "      <th>ptphi</th>\n",
       "      <th>isLL</th>\n",
       "      <th>isMIX</th>\n",
       "      <th>isTT</th>\n",
       "      <th>isBKG</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3.121000</td>\n",
       "      <td>-0.493042</td>\n",
       "      <td>-2.089111</td>\n",
       "      <td>1.291077</td>\n",
       "      <td>3.776367</td>\n",
       "      <td>2.835449</td>\n",
       "      <td>-0.940918</td>\n",
       "      <td>0.454224</td>\n",
       "      <td>-1.141846</td>\n",
       "      <td>369.785309</td>\n",
       "      <td>251.217087</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>-13.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>-2.706543</td>\n",
       "      <td>0.575562</td>\n",
       "      <td>0.465759</td>\n",
       "      <td>3.018066</td>\n",
       "      <td>85.704681</td>\n",
       "      <td>34.702194</td>\n",
       "      <td>143.582367</td>\n",
       "      <td>64.647858</td>\n",
       "      <td>96.756378</td>\n",
       "      <td>20.823217</td>\n",
       "      <td>-1.292615</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.174566</td>\n",
       "      <td>2.262878</td>\n",
       "      <td>1.649446</td>\n",
       "      <td>1.956162</td>\n",
       "      <td>2.820435</td>\n",
       "      <td>-3.073242</td>\n",
       "      <td>-0.252808</td>\n",
       "      <td>0.599854</td>\n",
       "      <td>-0.013578</td>\n",
       "      <td>625.133850</td>\n",
       "      <td>72.537148</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>-11.0</td>\n",
       "      <td>-0.617432</td>\n",
       "      <td>0.904297</td>\n",
       "      <td>-2.550293</td>\n",
       "      <td>-1.547607</td>\n",
       "      <td>187.422958</td>\n",
       "      <td>122.559952</td>\n",
       "      <td>86.586983</td>\n",
       "      <td>46.310486</td>\n",
       "      <td>118.139442</td>\n",
       "      <td>160.278671</td>\n",
       "      <td>2.963379</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.101411</td>\n",
       "      <td>0.720459</td>\n",
       "      <td>-0.154663</td>\n",
       "      <td>0.282898</td>\n",
       "      <td>4.667969</td>\n",
       "      <td>-2.785645</td>\n",
       "      <td>1.882324</td>\n",
       "      <td>0.268799</td>\n",
       "      <td>-0.606323</td>\n",
       "      <td>964.531616</td>\n",
       "      <td>60.436550</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>-11.0</td>\n",
       "      <td>0.683594</td>\n",
       "      <td>-2.258301</td>\n",
       "      <td>-0.911011</td>\n",
       "      <td>3.036621</td>\n",
       "      <td>134.064819</td>\n",
       "      <td>63.955444</td>\n",
       "      <td>43.268951</td>\n",
       "      <td>20.095602</td>\n",
       "      <td>32.741360</td>\n",
       "      <td>52.529774</td>\n",
       "      <td>2.673340</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.319307</td>\n",
       "      <td>-0.882874</td>\n",
       "      <td>0.757812</td>\n",
       "      <td>0.062531</td>\n",
       "      <td>6.555664</td>\n",
       "      <td>3.874023</td>\n",
       "      <td>-2.681641</td>\n",
       "      <td>-0.286682</td>\n",
       "      <td>1.354004</td>\n",
       "      <td>2529.641357</td>\n",
       "      <td>104.620178</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>-11.0</td>\n",
       "      <td>-1.653076</td>\n",
       "      <td>2.024902</td>\n",
       "      <td>-0.781616</td>\n",
       "      <td>-0.118011</td>\n",
       "      <td>96.125755</td>\n",
       "      <td>94.413933</td>\n",
       "      <td>90.736679</td>\n",
       "      <td>31.937508</td>\n",
       "      <td>117.554039</td>\n",
       "      <td>36.320225</td>\n",
       "      <td>2.645425</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.143195</td>\n",
       "      <td>-1.701904</td>\n",
       "      <td>2.665771</td>\n",
       "      <td>0.481934</td>\n",
       "      <td>4.703613</td>\n",
       "      <td>1.912109</td>\n",
       "      <td>-2.791504</td>\n",
       "      <td>-2.141602</td>\n",
       "      <td>2.226074</td>\n",
       "      <td>2628.467041</td>\n",
       "      <td>848.424438</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>-13.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>1.287109</td>\n",
       "      <td>-0.681152</td>\n",
       "      <td>1.925781</td>\n",
       "      <td>-1.184326</td>\n",
       "      <td>274.317688</td>\n",
       "      <td>226.593491</td>\n",
       "      <td>144.917450</td>\n",
       "      <td>61.419807</td>\n",
       "      <td>83.550461</td>\n",
       "      <td>144.095169</td>\n",
       "      <td>-1.629096</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</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",
       "      <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>150907</th>\n",
       "      <td>0.119845</td>\n",
       "      <td>-1.498657</td>\n",
       "      <td>-1.836914</td>\n",
       "      <td>1.667786</td>\n",
       "      <td>5.092041</td>\n",
       "      <td>-1.640869</td>\n",
       "      <td>3.451172</td>\n",
       "      <td>-0.593506</td>\n",
       "      <td>-0.931763</td>\n",
       "      <td>1345.346436</td>\n",
       "      <td>73.458191</td>\n",
       "      <td>8.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>-13.0</td>\n",
       "      <td>-0.828613</td>\n",
       "      <td>1.972412</td>\n",
       "      <td>-0.823120</td>\n",
       "      <td>2.449219</td>\n",
       "      <td>147.390198</td>\n",
       "      <td>74.537453</td>\n",
       "      <td>55.850170</td>\n",
       "      <td>23.574209</td>\n",
       "      <td>32.622272</td>\n",
       "      <td>149.986572</td>\n",
       "      <td>1.382741</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>150908</th>\n",
       "      <td>0.471296</td>\n",
       "      <td>-2.624390</td>\n",
       "      <td>-1.764771</td>\n",
       "      <td>2.194580</td>\n",
       "      <td>4.749268</td>\n",
       "      <td>-1.862549</td>\n",
       "      <td>2.886719</td>\n",
       "      <td>-2.112305</td>\n",
       "      <td>-1.252686</td>\n",
       "      <td>1061.369263</td>\n",
       "      <td>131.066040</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>-13.0</td>\n",
       "      <td>-2.499512</td>\n",
       "      <td>0.810181</td>\n",
       "      <td>-0.052177</td>\n",
       "      <td>-2.156250</td>\n",
       "      <td>151.726135</td>\n",
       "      <td>63.179993</td>\n",
       "      <td>98.920914</td>\n",
       "      <td>45.671574</td>\n",
       "      <td>85.310318</td>\n",
       "      <td>38.308739</td>\n",
       "      <td>-2.488425</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>150909</th>\n",
       "      <td>0.856618</td>\n",
       "      <td>0.081573</td>\n",
       "      <td>-0.641937</td>\n",
       "      <td>0.280182</td>\n",
       "      <td>3.092834</td>\n",
       "      <td>-0.361877</td>\n",
       "      <td>2.730957</td>\n",
       "      <td>1.266113</td>\n",
       "      <td>0.542603</td>\n",
       "      <td>320.076843</td>\n",
       "      <td>97.441307</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>-11.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>-2.647949</td>\n",
       "      <td>2.900391</td>\n",
       "      <td>0.397766</td>\n",
       "      <td>-1.020020</td>\n",
       "      <td>121.496910</td>\n",
       "      <td>40.693249</td>\n",
       "      <td>115.536629</td>\n",
       "      <td>36.656826</td>\n",
       "      <td>126.425659</td>\n",
       "      <td>27.028612</td>\n",
       "      <td>-0.017420</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>150910</th>\n",
       "      <td>0.109525</td>\n",
       "      <td>0.611725</td>\n",
       "      <td>1.862183</td>\n",
       "      <td>1.236954</td>\n",
       "      <td>3.763916</td>\n",
       "      <td>1.460205</td>\n",
       "      <td>-2.303711</td>\n",
       "      <td>0.189972</td>\n",
       "      <td>1.440430</td>\n",
       "      <td>538.265137</td>\n",
       "      <td>55.855526</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>-11.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>3.122559</td>\n",
       "      <td>0.601562</td>\n",
       "      <td>1.177734</td>\n",
       "      <td>-0.714233</td>\n",
       "      <td>84.040810</td>\n",
       "      <td>76.863762</td>\n",
       "      <td>32.649853</td>\n",
       "      <td>21.669298</td>\n",
       "      <td>32.998287</td>\n",
       "      <td>45.887920</td>\n",
       "      <td>-0.360784</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>150911</th>\n",
       "      <td>0.326320</td>\n",
       "      <td>-2.287918</td>\n",
       "      <td>-2.883133</td>\n",
       "      <td>2.585526</td>\n",
       "      <td>2.670746</td>\n",
       "      <td>0.089996</td>\n",
       "      <td>2.760742</td>\n",
       "      <td>-0.862549</td>\n",
       "      <td>-1.457764</td>\n",
       "      <td>363.036804</td>\n",
       "      <td>105.035767</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>-13.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>-2.616211</td>\n",
       "      <td>1.083008</td>\n",
       "      <td>0.357605</td>\n",
       "      <td>-2.375488</td>\n",
       "      <td>105.873451</td>\n",
       "      <td>76.026176</td>\n",
       "      <td>90.974586</td>\n",
       "      <td>28.871809</td>\n",
       "      <td>65.490448</td>\n",
       "      <td>23.610096</td>\n",
       "      <td>-2.485100</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>150912 rows × 30 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "             Rpt   Zepp_l1   Zepp_l2   Zepp_ll    detajj    eta_j1    eta_j2  \\\n",
       "0       3.121000 -0.493042 -2.089111  1.291077  3.776367  2.835449 -0.940918   \n",
       "1       0.174566  2.262878  1.649446  1.956162  2.820435 -3.073242 -0.252808   \n",
       "2       0.101411  0.720459 -0.154663  0.282898  4.667969 -2.785645  1.882324   \n",
       "3       0.319307 -0.882874  0.757812  0.062531  6.555664  3.874023 -2.681641   \n",
       "4       0.143195 -1.701904  2.665771  0.481934  4.703613  1.912109 -2.791504   \n",
       "...          ...       ...       ...       ...       ...       ...       ...   \n",
       "150907  0.119845 -1.498657 -1.836914  1.667786  5.092041 -1.640869  3.451172   \n",
       "150908  0.471296 -2.624390 -1.764771  2.194580  4.749268 -1.862549  2.886719   \n",
       "150909  0.856618  0.081573 -0.641937  0.280182  3.092834 -0.361877  2.730957   \n",
       "150910  0.109525  0.611725  1.862183  1.236954  3.763916  1.460205 -2.303711   \n",
       "150911  0.326320 -2.287918 -2.883133  2.585526  2.670746  0.089996  2.760742   \n",
       "\n",
       "          eta_l1    eta_l2          mjj         mll  njet  nlep  pdgid_l1  \\\n",
       "0       0.454224 -1.141846   369.785309  251.217087   4.0   2.0     -13.0   \n",
       "1       0.599854 -0.013578   625.133850   72.537148   6.0   2.0      11.0   \n",
       "2       0.268799 -0.606323   964.531616   60.436550   5.0   2.0      11.0   \n",
       "3      -0.286682  1.354004  2529.641357  104.620178   4.0   2.0      13.0   \n",
       "4      -2.141602  2.226074  2628.467041  848.424438   6.0   2.0     -13.0   \n",
       "...          ...       ...          ...         ...   ...   ...       ...   \n",
       "150907 -0.593506 -0.931763  1345.346436   73.458191   8.0   2.0      13.0   \n",
       "150908 -2.112305 -1.252686  1061.369263  131.066040   4.0   2.0      11.0   \n",
       "150909  1.266113  0.542603   320.076843   97.441307   5.0   2.0     -11.0   \n",
       "150910  0.189972  1.440430   538.265137   55.855526   5.0   2.0     -11.0   \n",
       "150911 -0.862549 -1.457764   363.036804  105.035767   6.0   2.0     -13.0   \n",
       "\n",
       "        pdgid_l2    phi_j1    phi_j2    phi_l1    phi_l2       pt_j1  \\\n",
       "0           11.0 -2.706543  0.575562  0.465759  3.018066   85.704681   \n",
       "1          -11.0 -0.617432  0.904297 -2.550293 -1.547607  187.422958   \n",
       "2          -11.0  0.683594 -2.258301 -0.911011  3.036621  134.064819   \n",
       "3          -11.0 -1.653076  2.024902 -0.781616 -0.118011   96.125755   \n",
       "4           11.0  1.287109 -0.681152  1.925781 -1.184326  274.317688   \n",
       "...          ...       ...       ...       ...       ...         ...   \n",
       "150907     -13.0 -0.828613  1.972412 -0.823120  2.449219  147.390198   \n",
       "150908     -13.0 -2.499512  0.810181 -0.052177 -2.156250  151.726135   \n",
       "150909      13.0 -2.647949  2.900391  0.397766 -1.020020  121.496910   \n",
       "150910      13.0  3.122559  0.601562  1.177734 -0.714233   84.040810   \n",
       "150911      13.0 -2.616211  1.083008  0.357605 -2.375488  105.873451   \n",
       "\n",
       "             pt_j2       pt_l1      pt_l2        ptll      ptmiss     ptphi  \\\n",
       "0        34.702194  143.582367  64.647858   96.756378   20.823217 -1.292615   \n",
       "1       122.559952   86.586983  46.310486  118.139442  160.278671  2.963379   \n",
       "2        63.955444   43.268951  20.095602   32.741360   52.529774  2.673340   \n",
       "3        94.413933   90.736679  31.937508  117.554039   36.320225  2.645425   \n",
       "4       226.593491  144.917450  61.419807   83.550461  144.095169 -1.629096   \n",
       "...            ...         ...        ...         ...         ...       ...   \n",
       "150907   74.537453   55.850170  23.574209   32.622272  149.986572  1.382741   \n",
       "150908   63.179993   98.920914  45.671574   85.310318   38.308739 -2.488425   \n",
       "150909   40.693249  115.536629  36.656826  126.425659   27.028612 -0.017420   \n",
       "150910   76.863762   32.649853  21.669298   32.998287   45.887920 -0.360784   \n",
       "150911   76.026176   90.974586  28.871809   65.490448   23.610096 -2.485100   \n",
       "\n",
       "        isLL  isMIX  isTT  isBKG  \n",
       "0        1.0    0.0   0.0    0.0  \n",
       "1        1.0    0.0   0.0    0.0  \n",
       "2        1.0    0.0   0.0    0.0  \n",
       "3        1.0    0.0   0.0    0.0  \n",
       "4        1.0    0.0   0.0    0.0  \n",
       "...      ...    ...   ...    ...  \n",
       "150907   0.0    0.0   0.0    1.0  \n",
       "150908   0.0    0.0   0.0    1.0  \n",
       "150909   0.0    0.0   0.0    1.0  \n",
       "150910   0.0    0.0   0.0    1.0  \n",
       "150911   0.0    0.0   0.0    1.0  \n",
       "\n",
       "[150912 rows x 30 columns]"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "\n",
    "#df = pd.read_pickle('https://pandora.infn.it/public/8d6872/dl/dataset_pol_vbs_full2018v9_VBS_WW_2j.pkl')\n",
    "df = pd.read_pickle('https://pandora.infn.it/public/63bb2f/dl/dataset_pol_vbs_full2018v9_VBS_WW_2j_new.pkl')\n",
    "pd.set_option('display.max_columns', None)\n",
    "df = df.drop(columns=['LHEcosthetastarcmww_lminus', 'LHEcosthetastarcmww_lplus', 'LHEPart_pt_0', 'LHEPart_pt_1', 'LHEPart_eta_0', 'LHEPart_eta_1', 'LHEPart_phi_0', 'LHEPart_phi_1', 'LHEPart_mass_0', 'LHEPart_mass_1', 'LHEPart_pdgId_0', 'LHEPart_pdgId_1',])\n",
    "components = ['pt', 'eta', 'phi', 'mass', 'pdgId']\n",
    "for particle in range(2,8):\n",
    "    for component in components:\n",
    "        df = df.drop(columns=['LHEPart_' + component + '_' + str(particle)])\n",
    "#df = df.drop(columns=['nlep', 'njet'])\n",
    "df.dropna(inplace=True)\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "650062c8-00b6-4f1e-abeb-63228d8126b9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['pt_l1', 'eta_l1', 'phi_l1', 'pt_l2', 'eta_l2', 'phi_l2', 'pt_j1', 'eta_j1', 'phi_j1', 'pt_j2', 'eta_j2', 'phi_j2', 'ptmiss', 'ptphi', 'isLL', 'isMIX', 'isTT', 'isBKG']\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>pt_l1</th>\n",
       "      <th>eta_l1</th>\n",
       "      <th>phi_l1</th>\n",
       "      <th>pt_l2</th>\n",
       "      <th>eta_l2</th>\n",
       "      <th>phi_l2</th>\n",
       "      <th>pt_j1</th>\n",
       "      <th>eta_j1</th>\n",
       "      <th>phi_j1</th>\n",
       "      <th>pt_j2</th>\n",
       "      <th>eta_j2</th>\n",
       "      <th>phi_j2</th>\n",
       "      <th>ptmiss</th>\n",
       "      <th>ptphi</th>\n",
       "      <th>isLL</th>\n",
       "      <th>isMIX</th>\n",
       "      <th>isTT</th>\n",
       "      <th>isBKG</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>143.582367</td>\n",
       "      <td>0.454224</td>\n",
       "      <td>0.465759</td>\n",
       "      <td>64.647858</td>\n",
       "      <td>-1.141846</td>\n",
       "      <td>3.018066</td>\n",
       "      <td>85.704681</td>\n",
       "      <td>2.835449</td>\n",
       "      <td>-2.706543</td>\n",
       "      <td>34.702194</td>\n",
       "      <td>-0.940918</td>\n",
       "      <td>0.575562</td>\n",
       "      <td>20.823217</td>\n",
       "      <td>-1.292615</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>86.586983</td>\n",
       "      <td>0.599854</td>\n",
       "      <td>-2.550293</td>\n",
       "      <td>46.310486</td>\n",
       "      <td>-0.013578</td>\n",
       "      <td>-1.547607</td>\n",
       "      <td>187.422958</td>\n",
       "      <td>-3.073242</td>\n",
       "      <td>-0.617432</td>\n",
       "      <td>122.559952</td>\n",
       "      <td>-0.252808</td>\n",
       "      <td>0.904297</td>\n",
       "      <td>160.278671</td>\n",
       "      <td>2.963379</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>43.268951</td>\n",
       "      <td>0.268799</td>\n",
       "      <td>-0.911011</td>\n",
       "      <td>20.095602</td>\n",
       "      <td>-0.606323</td>\n",
       "      <td>3.036621</td>\n",
       "      <td>134.064819</td>\n",
       "      <td>-2.785645</td>\n",
       "      <td>0.683594</td>\n",
       "      <td>63.955444</td>\n",
       "      <td>1.882324</td>\n",
       "      <td>-2.258301</td>\n",
       "      <td>52.529774</td>\n",
       "      <td>2.673340</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>90.736679</td>\n",
       "      <td>-0.286682</td>\n",
       "      <td>-0.781616</td>\n",
       "      <td>31.937508</td>\n",
       "      <td>1.354004</td>\n",
       "      <td>-0.118011</td>\n",
       "      <td>96.125755</td>\n",
       "      <td>3.874023</td>\n",
       "      <td>-1.653076</td>\n",
       "      <td>94.413933</td>\n",
       "      <td>-2.681641</td>\n",
       "      <td>2.024902</td>\n",
       "      <td>36.320225</td>\n",
       "      <td>2.645425</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>144.917450</td>\n",
       "      <td>-2.141602</td>\n",
       "      <td>1.925781</td>\n",
       "      <td>61.419807</td>\n",
       "      <td>2.226074</td>\n",
       "      <td>-1.184326</td>\n",
       "      <td>274.317688</td>\n",
       "      <td>1.912109</td>\n",
       "      <td>1.287109</td>\n",
       "      <td>226.593491</td>\n",
       "      <td>-2.791504</td>\n",
       "      <td>-0.681152</td>\n",
       "      <td>144.095169</td>\n",
       "      <td>-1.629096</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>150907</th>\n",
       "      <td>55.850170</td>\n",
       "      <td>-0.593506</td>\n",
       "      <td>-0.823120</td>\n",
       "      <td>23.574209</td>\n",
       "      <td>-0.931763</td>\n",
       "      <td>2.449219</td>\n",
       "      <td>147.390198</td>\n",
       "      <td>-1.640869</td>\n",
       "      <td>-0.828613</td>\n",
       "      <td>74.537453</td>\n",
       "      <td>3.451172</td>\n",
       "      <td>1.972412</td>\n",
       "      <td>149.986572</td>\n",
       "      <td>1.382741</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>150908</th>\n",
       "      <td>98.920914</td>\n",
       "      <td>-2.112305</td>\n",
       "      <td>-0.052177</td>\n",
       "      <td>45.671574</td>\n",
       "      <td>-1.252686</td>\n",
       "      <td>-2.156250</td>\n",
       "      <td>151.726135</td>\n",
       "      <td>-1.862549</td>\n",
       "      <td>-2.499512</td>\n",
       "      <td>63.179993</td>\n",
       "      <td>2.886719</td>\n",
       "      <td>0.810181</td>\n",
       "      <td>38.308739</td>\n",
       "      <td>-2.488425</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>150909</th>\n",
       "      <td>115.536629</td>\n",
       "      <td>1.266113</td>\n",
       "      <td>0.397766</td>\n",
       "      <td>36.656826</td>\n",
       "      <td>0.542603</td>\n",
       "      <td>-1.020020</td>\n",
       "      <td>121.496910</td>\n",
       "      <td>-0.361877</td>\n",
       "      <td>-2.647949</td>\n",
       "      <td>40.693249</td>\n",
       "      <td>2.730957</td>\n",
       "      <td>2.900391</td>\n",
       "      <td>27.028612</td>\n",
       "      <td>-0.017420</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>150910</th>\n",
       "      <td>32.649853</td>\n",
       "      <td>0.189972</td>\n",
       "      <td>1.177734</td>\n",
       "      <td>21.669298</td>\n",
       "      <td>1.440430</td>\n",
       "      <td>-0.714233</td>\n",
       "      <td>84.040810</td>\n",
       "      <td>1.460205</td>\n",
       "      <td>3.122559</td>\n",
       "      <td>76.863762</td>\n",
       "      <td>-2.303711</td>\n",
       "      <td>0.601562</td>\n",
       "      <td>45.887920</td>\n",
       "      <td>-0.360784</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>150911</th>\n",
       "      <td>90.974586</td>\n",
       "      <td>-0.862549</td>\n",
       "      <td>0.357605</td>\n",
       "      <td>28.871809</td>\n",
       "      <td>-1.457764</td>\n",
       "      <td>-2.375488</td>\n",
       "      <td>105.873451</td>\n",
       "      <td>0.089996</td>\n",
       "      <td>-2.616211</td>\n",
       "      <td>76.026176</td>\n",
       "      <td>2.760742</td>\n",
       "      <td>1.083008</td>\n",
       "      <td>23.610096</td>\n",
       "      <td>-2.485100</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>150912 rows × 18 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "             pt_l1    eta_l1    phi_l1      pt_l2    eta_l2    phi_l2  \\\n",
       "0       143.582367  0.454224  0.465759  64.647858 -1.141846  3.018066   \n",
       "1        86.586983  0.599854 -2.550293  46.310486 -0.013578 -1.547607   \n",
       "2        43.268951  0.268799 -0.911011  20.095602 -0.606323  3.036621   \n",
       "3        90.736679 -0.286682 -0.781616  31.937508  1.354004 -0.118011   \n",
       "4       144.917450 -2.141602  1.925781  61.419807  2.226074 -1.184326   \n",
       "...            ...       ...       ...        ...       ...       ...   \n",
       "150907   55.850170 -0.593506 -0.823120  23.574209 -0.931763  2.449219   \n",
       "150908   98.920914 -2.112305 -0.052177  45.671574 -1.252686 -2.156250   \n",
       "150909  115.536629  1.266113  0.397766  36.656826  0.542603 -1.020020   \n",
       "150910   32.649853  0.189972  1.177734  21.669298  1.440430 -0.714233   \n",
       "150911   90.974586 -0.862549  0.357605  28.871809 -1.457764 -2.375488   \n",
       "\n",
       "             pt_j1    eta_j1    phi_j1       pt_j2    eta_j2    phi_j2  \\\n",
       "0        85.704681  2.835449 -2.706543   34.702194 -0.940918  0.575562   \n",
       "1       187.422958 -3.073242 -0.617432  122.559952 -0.252808  0.904297   \n",
       "2       134.064819 -2.785645  0.683594   63.955444  1.882324 -2.258301   \n",
       "3        96.125755  3.874023 -1.653076   94.413933 -2.681641  2.024902   \n",
       "4       274.317688  1.912109  1.287109  226.593491 -2.791504 -0.681152   \n",
       "...            ...       ...       ...         ...       ...       ...   \n",
       "150907  147.390198 -1.640869 -0.828613   74.537453  3.451172  1.972412   \n",
       "150908  151.726135 -1.862549 -2.499512   63.179993  2.886719  0.810181   \n",
       "150909  121.496910 -0.361877 -2.647949   40.693249  2.730957  2.900391   \n",
       "150910   84.040810  1.460205  3.122559   76.863762 -2.303711  0.601562   \n",
       "150911  105.873451  0.089996 -2.616211   76.026176  2.760742  1.083008   \n",
       "\n",
       "            ptmiss     ptphi  isLL  isMIX  isTT  isBKG  \n",
       "0        20.823217 -1.292615   1.0    0.0   0.0    0.0  \n",
       "1       160.278671  2.963379   1.0    0.0   0.0    0.0  \n",
       "2        52.529774  2.673340   1.0    0.0   0.0    0.0  \n",
       "3        36.320225  2.645425   1.0    0.0   0.0    0.0  \n",
       "4       144.095169 -1.629096   1.0    0.0   0.0    0.0  \n",
       "...            ...       ...   ...    ...   ...    ...  \n",
       "150907  149.986572  1.382741   0.0    0.0   0.0    1.0  \n",
       "150908   38.308739 -2.488425   0.0    0.0   0.0    1.0  \n",
       "150909   27.028612 -0.017420   0.0    0.0   0.0    1.0  \n",
       "150910   45.887920 -0.360784   0.0    0.0   0.0    1.0  \n",
       "150911   23.610096 -2.485100   0.0    0.0   0.0    1.0  \n",
       "\n",
       "[150912 rows x 18 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "targets = ['LL', 'MIX', 'TT', 'BKG']\n",
    "particles = ['l', 'j']\n",
    "components = ['pt', 'eta', 'phi']\n",
    "reco_columns = []\n",
    "for particle in particles:\n",
    "    for i in range(1,3):\n",
    "        for component in components:\n",
    "            reco_columns.append(component + '_' + particle + str(i))\n",
    "reco_columns.append('ptmiss')\n",
    "reco_columns.append('ptphi')\n",
    "for target in targets:\n",
    "    reco_columns.append('is' + target)\n",
    "print(reco_columns)\n",
    "df_reco = df[reco_columns]\n",
    "df_reco"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "8ac33c4a-b5ab-48f8-9354-c446a67b5ee5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             Rpt          mjj         mll  njet  nlep  pdgid_l1  pdgid_l2  \\\n",
      "0       3.121000   369.785309  251.217087   4.0   2.0     -13.0      11.0   \n",
      "1       0.174566   625.133850   72.537148   6.0   2.0      11.0     -11.0   \n",
      "2       0.101411   964.531616   60.436550   5.0   2.0      11.0     -11.0   \n",
      "3       0.319307  2529.641357  104.620178   4.0   2.0      13.0     -11.0   \n",
      "4       0.143195  2628.467041  848.424438   6.0   2.0     -13.0      11.0   \n",
      "...          ...          ...         ...   ...   ...       ...       ...   \n",
      "150907  0.119845  1345.346436   73.458191   8.0   2.0      13.0     -13.0   \n",
      "150908  0.471296  1061.369263  131.066040   4.0   2.0      11.0     -13.0   \n",
      "150909  0.856618   320.076843   97.441307   5.0   2.0     -11.0      13.0   \n",
      "150910  0.109525   538.265137   55.855526   5.0   2.0     -11.0      13.0   \n",
      "150911  0.326320   363.036804  105.035767   6.0   2.0     -13.0      13.0   \n",
      "\n",
      "              ptll   Zepp_l1   Zepp_l2   Zepp_ll    detajj  isLL  isMIX  isTT  \\\n",
      "0        96.756378 -0.493042 -2.089111  1.291077  3.776367   1.0    0.0   0.0   \n",
      "1       118.139442  2.262878  1.649446  1.956162  2.820435   1.0    0.0   0.0   \n",
      "2        32.741360  0.720459 -0.154663  0.282898  4.667969   1.0    0.0   0.0   \n",
      "3       117.554039 -0.882874  0.757812  0.062531  6.555664   1.0    0.0   0.0   \n",
      "4        83.550461 -1.701904  2.665771  0.481934  4.703613   1.0    0.0   0.0   \n",
      "...            ...       ...       ...       ...       ...   ...    ...   ...   \n",
      "150907   32.622272 -1.498657 -1.836914  1.667786  5.092041   0.0    0.0   0.0   \n",
      "150908   85.310318 -2.624390 -1.764771  2.194580  4.749268   0.0    0.0   0.0   \n",
      "150909  126.425659  0.081573 -0.641937  0.280182  3.092834   0.0    0.0   0.0   \n",
      "150910   32.998287  0.611725  1.862183  1.236954  3.763916   0.0    0.0   0.0   \n",
      "150911   65.490448 -2.287918 -2.883133  2.585526  2.670746   0.0    0.0   0.0   \n",
      "\n",
      "        isBKG  \n",
      "0         0.0  \n",
      "1         0.0  \n",
      "2         0.0  \n",
      "3         0.0  \n",
      "4         0.0  \n",
      "...       ...  \n",
      "150907    1.0  \n",
      "150908    1.0  \n",
      "150909    1.0  \n",
      "150910    1.0  \n",
      "150911    1.0  \n",
      "\n",
      "[150912 rows x 16 columns]\n"
     ]
    }
   ],
   "source": [
    "df_aux = df[['Rpt', 'mjj', 'mll', 'njet', 'nlep', 'pdgid_l1', 'pdgid_l2', 'ptll', 'Zepp_l1', 'Zepp_l2', 'Zepp_ll', 'detajj', 'isLL', 'isMIX', 'isTT', 'isBKG']]\n",
    "print(df_aux)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "e2dcd06f-898a-4fc1-991b-a723448bcaf6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting lbn\n",
      "  Downloading lbn-1.2.2.tar.gz (13 kB)\n",
      "  Preparing metadata (setup.py) ... \u001b[?25ldone\n",
      "\u001b[?25hRequirement already satisfied: numpy in /usr/local/lib/python3.8/dist-packages (from lbn) (1.21.3)\n",
      "Requirement already satisfied: tensorflow in /usr/local/lib/python3.8/dist-packages (from lbn) (2.7.0rc1)\n",
      "Requirement already satisfied: h5py>=2.9.0 in /usr/local/lib/python3.8/dist-packages (from tensorflow->lbn) (2.10.0)\n",
      "Requirement already satisfied: opt-einsum>=2.3.2 in /usr/local/lib/python3.8/dist-packages (from tensorflow->lbn) (3.3.0)\n",
      "Requirement already satisfied: tensorflow-io-gcs-filesystem>=0.21.0 in /usr/local/lib/python3.8/dist-packages (from tensorflow->lbn) (0.21.0)\n",
      "Requirement already satisfied: flatbuffers<3.0,>=1.12 in /usr/local/lib/python3.8/dist-packages (from tensorflow->lbn) (2.0)\n",
      "Requirement already satisfied: libclang>=9.0.1 in /usr/local/lib/python3.8/dist-packages (from tensorflow->lbn) (12.0.0)\n",
      "Requirement already satisfied: gast<0.5.0,>=0.2.1 in /usr/local/lib/python3.8/dist-packages (from tensorflow->lbn) (0.4.0)\n",
      "Requirement already satisfied: tensorflow-estimator<2.8,~=2.7.0rc0 in /usr/local/lib/python3.8/dist-packages (from tensorflow->lbn) (2.7.0rc0)\n",
      "Requirement already satisfied: tensorboard~=2.6 in /usr/local/lib/python3.8/dist-packages (from tensorflow->lbn) (2.7.0)\n",
      "Requirement already satisfied: keras<2.8,>=2.7.0rc0 in /usr/local/lib/python3.8/dist-packages (from tensorflow->lbn) (2.7.0)\n",
      "Requirement already satisfied: keras-preprocessing>=1.1.1 in /usr/local/lib/python3.8/dist-packages (from tensorflow->lbn) (1.1.2)\n",
      "Requirement already satisfied: astunparse>=1.6.0 in /usr/local/lib/python3.8/dist-packages (from tensorflow->lbn) (1.6.3)\n",
      "Requirement already satisfied: typing-extensions>=3.6.6 in /usr/local/lib/python3.8/dist-packages (from tensorflow->lbn) (3.10.0.2)\n",
      "Requirement already satisfied: six>=1.12.0 in /usr/lib/python3/dist-packages (from tensorflow->lbn) (1.14.0)\n",
      "Requirement already satisfied: protobuf>=3.9.2 in /usr/local/lib/python3.8/dist-packages (from tensorflow->lbn) (3.19.0)\n",
      "Requirement already satisfied: termcolor>=1.1.0 in /usr/local/lib/python3.8/dist-packages (from tensorflow->lbn) (1.1.0)\n",
      "Requirement already satisfied: google-pasta>=0.1.1 in /usr/local/lib/python3.8/dist-packages (from tensorflow->lbn) (0.2.0)\n",
      "Requirement already satisfied: wrapt>=1.11.0 in /usr/local/lib/python3.8/dist-packages (from tensorflow->lbn) (1.13.2)\n",
      "Requirement already satisfied: grpcio<2.0,>=1.24.3 in /usr/local/lib/python3.8/dist-packages (from tensorflow->lbn) (1.41.0)\n",
      "Requirement already satisfied: wheel<1.0,>=0.32.0 in /usr/local/lib/python3.8/dist-packages (from tensorflow->lbn) (0.37.0)\n",
      "Requirement already satisfied: absl-py>=0.4.0 in /usr/local/lib/python3.8/dist-packages (from tensorflow->lbn) (0.15.0)\n",
      "Requirement already satisfied: google-auth<3,>=1.6.3 in /usr/local/lib/python3.8/dist-packages (from tensorboard~=2.6->tensorflow->lbn) (2.3.0)\n",
      "Requirement already satisfied: google-auth-oauthlib<0.5,>=0.4.1 in /usr/local/lib/python3.8/dist-packages (from tensorboard~=2.6->tensorflow->lbn) (0.4.6)\n",
      "Requirement already satisfied: setuptools>=41.0.0 in /usr/local/lib/python3.8/dist-packages (from tensorboard~=2.6->tensorflow->lbn) (59.6.0)\n",
      "Requirement already satisfied: tensorboard-data-server<0.7.0,>=0.6.0 in /usr/local/lib/python3.8/dist-packages (from tensorboard~=2.6->tensorflow->lbn) (0.6.1)\n",
      "Requirement already satisfied: requests<3,>=2.21.0 in /usr/lib/python3/dist-packages (from tensorboard~=2.6->tensorflow->lbn) (2.22.0)\n",
      "Requirement already satisfied: tensorboard-plugin-wit>=1.6.0 in /usr/local/lib/python3.8/dist-packages (from tensorboard~=2.6->tensorflow->lbn) (1.8.0)\n",
      "Requirement already satisfied: werkzeug>=0.11.15 in /usr/local/lib/python3.8/dist-packages (from tensorboard~=2.6->tensorflow->lbn) (2.0.2)\n",
      "Requirement already satisfied: markdown>=2.6.8 in /usr/local/lib/python3.8/dist-packages (from tensorboard~=2.6->tensorflow->lbn) (3.3.4)\n",
      "Requirement already satisfied: pyasn1-modules>=0.2.1 in /usr/local/lib/python3.8/dist-packages (from google-auth<3,>=1.6.3->tensorboard~=2.6->tensorflow->lbn) (0.2.8)\n",
      "Requirement already satisfied: cachetools<5.0,>=2.0.0 in /usr/local/lib/python3.8/dist-packages (from google-auth<3,>=1.6.3->tensorboard~=2.6->tensorflow->lbn) (4.2.4)\n",
      "Requirement already satisfied: rsa<5,>=3.1.4 in /usr/local/lib/python3.8/dist-packages (from google-auth<3,>=1.6.3->tensorboard~=2.6->tensorflow->lbn) (4.7.2)\n",
      "Requirement already satisfied: requests-oauthlib>=0.7.0 in /usr/local/lib/python3.8/dist-packages (from google-auth-oauthlib<0.5,>=0.4.1->tensorboard~=2.6->tensorflow->lbn) (1.3.0)\n",
      "Requirement already satisfied: pyasn1<0.5.0,>=0.4.6 in /usr/local/lib/python3.8/dist-packages (from pyasn1-modules>=0.2.1->google-auth<3,>=1.6.3->tensorboard~=2.6->tensorflow->lbn) (0.4.8)\n",
      "Requirement already satisfied: oauthlib>=3.0.0 in /usr/local/lib/python3.8/dist-packages (from requests-oauthlib>=0.7.0->google-auth-oauthlib<0.5,>=0.4.1->tensorboard~=2.6->tensorflow->lbn) (3.1.1)\n",
      "Building wheels for collected packages: lbn\n",
      "  Building wheel for lbn (setup.py) ... \u001b[?25ldone\n",
      "\u001b[?25h  Created wheel for lbn: filename=lbn-1.2.2-py3-none-any.whl size=15414 sha256=1e772e8465e60517cb22b979b89e969fe70a42d23640a9c1c0683bde8fe3979a\n",
      "  Stored in directory: /root/.cache/pip/wheels/34/5d/1d/7f33a54a32eb4e9bd2b3ac22eb576f64840734a604094b9ef2\n",
      "Successfully built lbn\n",
      "Installing collected packages: lbn\n",
      "Successfully installed lbn-1.2.2\n",
      "\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\n",
      "\u001b[33mWARNING: You are using pip version 21.3.1; however, version 23.1 is available.\n",
      "You should consider upgrading via the '/usr/bin/python3 -m pip install --upgrade pip' command.\u001b[0m\n"
     ]
    }
   ],
   "source": [
    "!pip install lbn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "b1c30b9d-7706-4fda-8bce-99ffc4a93c08",
   "metadata": {},
   "outputs": [],
   "source": [
    "import lbn\n",
    "import math\n",
    "import numpy as np\n",
    "import tensorflow as tf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "56ed54fe-471c-4381-a8ee-d0266f7f30dd",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             Rpt          mjj         mll  njet  nlep  pdgid_l1  pdgid_l2  \\\n",
      "0       3.121000   369.785309  251.217087   4.0   2.0     -13.0      11.0   \n",
      "1       0.174566   625.133850   72.537148   6.0   2.0      11.0     -11.0   \n",
      "2       0.101411   964.531616   60.436550   5.0   2.0      11.0     -11.0   \n",
      "3       0.319307  2529.641357  104.620178   4.0   2.0      13.0     -11.0   \n",
      "4       0.143195  2628.467041  848.424438   6.0   2.0     -13.0      11.0   \n",
      "...          ...          ...         ...   ...   ...       ...       ...   \n",
      "150907  0.119845  1345.346436   73.458191   8.0   2.0      13.0     -13.0   \n",
      "150908  0.471296  1061.369263  131.066040   4.0   2.0      11.0     -13.0   \n",
      "150909  0.856618   320.076843   97.441307   5.0   2.0     -11.0      13.0   \n",
      "150910  0.109525   538.265137   55.855526   5.0   2.0     -11.0      13.0   \n",
      "150911  0.326320   363.036804  105.035767   6.0   2.0     -13.0      13.0   \n",
      "\n",
      "              ptll   Zepp_l1   Zepp_l2   Zepp_ll    detajj  isLL  isMIX  isTT  \\\n",
      "0        96.756378 -0.493042 -2.089111  1.291077  3.776367   1.0    0.0   0.0   \n",
      "1       118.139442  2.262878  1.649446  1.956162  2.820435   1.0    0.0   0.0   \n",
      "2        32.741360  0.720459 -0.154663  0.282898  4.667969   1.0    0.0   0.0   \n",
      "3       117.554039 -0.882874  0.757812  0.062531  6.555664   1.0    0.0   0.0   \n",
      "4        83.550461 -1.701904  2.665771  0.481934  4.703613   1.0    0.0   0.0   \n",
      "...            ...       ...       ...       ...       ...   ...    ...   ...   \n",
      "150907   32.622272 -1.498657 -1.836914  1.667786  5.092041   0.0    0.0   0.0   \n",
      "150908   85.310318 -2.624390 -1.764771  2.194580  4.749268   0.0    0.0   0.0   \n",
      "150909  126.425659  0.081573 -0.641937  0.280182  3.092834   0.0    0.0   0.0   \n",
      "150910   32.998287  0.611725  1.862183  1.236954  3.763916   0.0    0.0   0.0   \n",
      "150911   65.490448 -2.287918 -2.883133  2.585526  2.670746   0.0    0.0   0.0   \n",
      "\n",
      "        isBKG  \n",
      "0         0.0  \n",
      "1         0.0  \n",
      "2         0.0  \n",
      "3         0.0  \n",
      "4         0.0  \n",
      "...       ...  \n",
      "150907    1.0  \n",
      "150908    1.0  \n",
      "150909    1.0  \n",
      "150910    1.0  \n",
      "150911    1.0  \n",
      "\n",
      "[150912 rows x 16 columns]\n",
      "            Rpt          mjj         mll  njet  nlep  pdgid_l1  pdgid_l2  \\\n",
      "37731  1.062498   806.283325  153.874802   4.0   2.0      11.0     -11.0   \n",
      "37735  0.189368  1354.570557  414.468201   2.0   2.0      11.0     -13.0   \n",
      "37738  0.253936  1427.392212  109.180862   5.0   2.0      11.0     -13.0   \n",
      "37739  0.369804   669.901550   92.755089   7.0   2.0     -13.0      11.0   \n",
      "37746  0.033068  1197.946289  108.849846   3.0   2.0      13.0     -13.0   \n",
      "...         ...          ...         ...   ...   ...       ...       ...   \n",
      "75427  3.884041  1191.636841  323.387695   3.0   2.0      13.0     -13.0   \n",
      "75435  0.080335  1912.755981   63.266106   3.0   2.0     -11.0      11.0   \n",
      "75450  0.116048  1149.677002   55.297207   5.0   2.0      13.0     -11.0   \n",
      "75454  0.083136   896.681519  146.205841   3.0   2.0     -11.0      11.0   \n",
      "75455  0.096620  1323.256958  205.244522   2.0   2.0     -11.0      13.0   \n",
      "\n",
      "             ptll   Zepp_l1   Zepp_l2   Zepp_ll    detajj  isLL  isMIX  isTT  \\\n",
      "37731  209.241730 -0.103638 -0.082886  0.093262  3.842041   0.0    1.0   0.0   \n",
      "37735  130.982742  0.462891 -3.040283  1.288696  4.252441   0.0    1.0   0.0   \n",
      "37738   79.243851 -1.666626  0.043457  0.811584  5.731934   0.0    1.0   0.0   \n",
      "37739   32.381691 -0.338013 -1.833664  1.085838  4.739990   0.0    1.0   0.0   \n",
      "37746   88.814606 -0.094513  2.257050  1.081268  3.531311   0.0    1.0   0.0   \n",
      "...           ...       ...       ...       ...       ...   ...    ...   ...   \n",
      "75427  121.304588  0.888916  1.149414  1.019165  5.296875   0.0    1.0   0.0   \n",
      "75435  101.415688  1.032471  0.617920  0.825195  4.682617   0.0    1.0   0.0   \n",
      "75450   73.434639  0.287354  1.379639  0.833496  4.558105   0.0    1.0   0.0   \n",
      "75454   73.797089 -0.792107  1.124390  0.166142  3.048584   0.0    1.0   0.0   \n",
      "75455   83.150101 -0.765198  0.433014  0.166092  3.023071   0.0    1.0   0.0   \n",
      "\n",
      "       isBKG  \n",
      "37731    0.0  \n",
      "37735    0.0  \n",
      "37738    0.0  \n",
      "37739    0.0  \n",
      "37746    0.0  \n",
      "...      ...  \n",
      "75427    0.0  \n",
      "75435    0.0  \n",
      "75450    0.0  \n",
      "75454    0.0  \n",
      "75455    0.0  \n",
      "\n",
      "[7637 rows x 16 columns]\n",
      "isLL\n",
      "            Rpt          mjj         mll  njet  nlep  pdgid_l1  pdgid_l2  \\\n",
      "0      0.174566   625.133850   72.537148   6.0   2.0      11.0     -11.0   \n",
      "1      0.319307  2529.641357  104.620178   4.0   2.0      13.0     -11.0   \n",
      "2      0.143195  2628.467041  848.424438   6.0   2.0     -13.0      11.0   \n",
      "3      0.077832   826.541565   70.460709   6.0   2.0      13.0     -11.0   \n",
      "4      0.159875   735.234680  115.907455   4.0   2.0      13.0     -13.0   \n",
      "...         ...          ...         ...   ...   ...       ...       ...   \n",
      "30081  0.105021   740.435059   77.821083   5.0   2.0      13.0     -11.0   \n",
      "30082  1.568811   930.527039  123.436935   3.0   2.0      11.0     -13.0   \n",
      "30083  0.059037   861.158936  111.474136   6.0   2.0      11.0     -11.0   \n",
      "30084  0.267582  1836.835449   64.767357   3.0   2.0     -11.0      11.0   \n",
      "30085  0.691047   358.356628  151.662308   3.0   2.0     -13.0      13.0   \n",
      "\n",
      "             ptll   Zepp_l1   Zepp_l2   Zepp_ll    detajj  \n",
      "0      118.139442  2.262878  1.649446  1.956162  2.820435  \n",
      "1      117.554039 -0.882874  0.757812  0.062531  6.555664  \n",
      "2       83.550461 -1.701904  2.665771  0.481934  4.703613  \n",
      "3       98.435509  1.275055 -0.054352  0.610352  2.995911  \n",
      "4       80.886627 -0.893555  1.151855  0.129150  3.887207  \n",
      "...           ...       ...       ...       ...       ...  \n",
      "30081   82.318771 -0.952820  0.392761  0.280029  3.228882  \n",
      "30082   40.378323  0.717529 -0.355255  0.181137  6.047363  \n",
      "30083   48.853912 -0.709244  1.158325  0.224541  3.243896  \n",
      "30084   55.961037 -0.556885 -0.350098  0.453491  6.210449  \n",
      "30085   84.009300  0.100220 -2.625000  1.262390  3.921875  \n",
      "\n",
      "[30086 rows x 12 columns]\n",
      "           Rpt          mjj         mll  njet  nlep  pdgid_l1  pdgid_l2  \\\n",
      "0     3.121000   369.785309  251.217087   4.0   2.0     -13.0      11.0   \n",
      "1     0.101411   964.531616   60.436550   5.0   2.0      11.0     -11.0   \n",
      "2     0.217704   515.762207   80.798332   3.0   2.0      11.0     -13.0   \n",
      "3     0.043198  2178.861328   97.647255   5.0   2.0     -13.0      11.0   \n",
      "4     0.122515  1851.078735  125.590958   6.0   2.0      13.0     -11.0   \n",
      "...        ...          ...         ...   ...   ...       ...       ...   \n",
      "7637  0.105792   898.653320  144.515640   5.0   2.0      13.0     -13.0   \n",
      "7638  0.046662   811.794922   59.632343   6.0   2.0      13.0     -11.0   \n",
      "7639  0.198054  1044.429688   64.873459   7.0   2.0     -13.0      13.0   \n",
      "7640  0.050715  1305.179199  163.786316   4.0   2.0      11.0     -13.0   \n",
      "7641  0.675221  1366.365723  187.907181   3.0   2.0      13.0     -13.0   \n",
      "\n",
      "            ptll   Zepp_l1   Zepp_l2   Zepp_ll    detajj  \n",
      "0      96.756378 -0.493042 -2.089111  1.291077  3.776367  \n",
      "1      32.741360  0.720459 -0.154663  0.282898  4.667969  \n",
      "2     100.884567 -1.220703 -0.069397  0.645050  2.982300  \n",
      "3      46.198227 -0.650208 -2.467773  1.558990  5.109375  \n",
      "4      99.481163  2.071289  0.764282  1.417786  4.746094  \n",
      "...          ...       ...       ...       ...       ...  \n",
      "7637   63.539852  1.916626 -1.231567  0.342529  4.467041  \n",
      "7638   61.190830  1.317261  0.020126  0.668694  3.525146  \n",
      "7639   44.887016  0.384949  1.688843  1.036896  5.429688  \n",
      "7640   72.139420  1.652222 -0.519913  0.566154  3.358154  \n",
      "7641  187.205643 -1.050476 -3.060913  2.055695  5.346436  \n",
      "\n",
      "[7642 rows x 12 columns]\n",
      "isMIX\n",
      "            Rpt          mjj         mll  njet  nlep  pdgid_l1  pdgid_l2  \\\n",
      "0      0.144685   972.579529  287.280884   3.0   2.0      13.0     -11.0   \n",
      "1      0.159244   995.480347  120.022789   4.0   2.0     -11.0      11.0   \n",
      "2      0.174890  3644.313721  135.231796   7.0   2.0      13.0     -13.0   \n",
      "3      1.296082  1185.575928  223.417755   6.0   2.0     -13.0      13.0   \n",
      "4      0.813441   511.817902  182.637680   7.0   2.0     -11.0      13.0   \n",
      "...         ...          ...         ...   ...   ...       ...       ...   \n",
      "30086  0.366750  1429.780151   95.759857   4.0   2.0      13.0     -11.0   \n",
      "30087  0.055606  3249.864258  149.160645   5.0   2.0      13.0     -13.0   \n",
      "30088  0.073084   921.309448   79.331589   4.0   2.0     -13.0      11.0   \n",
      "30089  0.215730  1042.503784  218.219437   5.0   2.0     -11.0      13.0   \n",
      "30090  0.235714  1310.632080   90.200844   4.0   2.0      11.0     -11.0   \n",
      "\n",
      "             ptll   Zepp_l1   Zepp_l2   Zepp_ll    detajj  \n",
      "0       79.778275 -0.672699  2.002350  0.664825  3.185730  \n",
      "1      120.419258 -0.818359  0.789856  0.014252  3.652344  \n",
      "2       90.189827 -0.286255  1.295776  0.504761  6.626953  \n",
      "3      109.096802 -1.808258 -2.789185  2.298721  5.168213  \n",
      "4      251.359467 -0.678879  0.677567  0.000656  2.558746  \n",
      "...           ...       ...       ...       ...       ...  \n",
      "30086  174.372803 -1.137939 -1.458496  1.298218  5.673584  \n",
      "30087   72.698792  1.453613 -1.702881  0.124634  6.369629  \n",
      "30088   72.049942 -0.261688 -1.121857  0.691772  3.687073  \n",
      "30089   62.427948  1.695190 -1.410767  0.142212  4.693115  \n",
      "30090  111.024574 -0.619934 -2.064026  1.341980  4.867065  \n",
      "\n",
      "[30091 rows x 12 columns]\n",
      "           Rpt          mjj         mll  njet  nlep  pdgid_l1  pdgid_l2  \\\n",
      "0     1.062498   806.283325  153.874802   4.0   2.0      11.0     -11.0   \n",
      "1     0.189368  1354.570557  414.468201   2.0   2.0      11.0     -13.0   \n",
      "2     0.253936  1427.392212  109.180862   5.0   2.0      11.0     -13.0   \n",
      "3     0.369804   669.901550   92.755089   7.0   2.0     -13.0      11.0   \n",
      "4     0.033068  1197.946289  108.849846   3.0   2.0      13.0     -13.0   \n",
      "...        ...          ...         ...   ...   ...       ...       ...   \n",
      "7632  3.884041  1191.636841  323.387695   3.0   2.0      13.0     -13.0   \n",
      "7633  0.080335  1912.755981   63.266106   3.0   2.0     -11.0      11.0   \n",
      "7634  0.116048  1149.677002   55.297207   5.0   2.0      13.0     -11.0   \n",
      "7635  0.083136   896.681519  146.205841   3.0   2.0     -11.0      11.0   \n",
      "7636  0.096620  1323.256958  205.244522   2.0   2.0     -11.0      13.0   \n",
      "\n",
      "            ptll   Zepp_l1   Zepp_l2   Zepp_ll    detajj  \n",
      "0     209.241730 -0.103638 -0.082886  0.093262  3.842041  \n",
      "1     130.982742  0.462891 -3.040283  1.288696  4.252441  \n",
      "2      79.243851 -1.666626  0.043457  0.811584  5.731934  \n",
      "3      32.381691 -0.338013 -1.833664  1.085838  4.739990  \n",
      "4      88.814606 -0.094513  2.257050  1.081268  3.531311  \n",
      "...          ...       ...       ...       ...       ...  \n",
      "7632  121.304588  0.888916  1.149414  1.019165  5.296875  \n",
      "7633  101.415688  1.032471  0.617920  0.825195  4.682617  \n",
      "7634   73.434639  0.287354  1.379639  0.833496  4.558105  \n",
      "7635   73.797089 -0.792107  1.124390  0.166142  3.048584  \n",
      "7636   83.150101 -0.765198  0.433014  0.166092  3.023071  \n",
      "\n",
      "[7637 rows x 12 columns]\n",
      "isTT\n",
      "            Rpt          mjj         mll  njet  nlep  pdgid_l1  pdgid_l2  \\\n",
      "0      0.284201  1638.271240  159.267258   3.0   2.0      13.0     -11.0   \n",
      "1      0.265005  1223.334473  258.683960   5.0   2.0     -11.0      13.0   \n",
      "2      0.924994   675.782715   72.993866   4.0   2.0      13.0     -13.0   \n",
      "3      0.095162  1844.334839   51.016777   4.0   2.0      11.0     -13.0   \n",
      "4      0.773607   457.260864  142.724167   2.0   2.0      11.0     -11.0   \n",
      "...         ...          ...         ...   ...   ...       ...       ...   \n",
      "30165  0.500550   753.336182  227.237274   4.0   2.0      13.0     -13.0   \n",
      "30166  0.064646  1407.892090  102.510735   4.0   2.0     -13.0      13.0   \n",
      "30167  0.435524   997.746826  597.755554   3.0   2.0      11.0     -13.0   \n",
      "30168  0.493099   995.869202   62.829956   4.0   2.0      11.0     -13.0   \n",
      "30169  1.092131   337.764465  179.845459   5.0   2.0      13.0     -11.0   \n",
      "\n",
      "             ptll   Zepp_l1   Zepp_l2   Zepp_ll    detajj  \n",
      "0      100.149376 -0.099640  0.118835  0.009598  4.788086  \n",
      "1      133.424225  1.393677 -1.589844  0.098083  4.741455  \n",
      "2      122.668091 -1.283630 -0.154358  0.718994  4.720581  \n",
      "3       56.149624  0.458984 -0.585510  0.063263  5.704590  \n",
      "4       63.458973  1.013916 -0.544800  0.234558  3.848145  \n",
      "...           ...       ...       ...       ...       ...  \n",
      "30165  134.228195  0.064636 -3.026672  1.481018  4.715698  \n",
      "30166   37.505882  0.427002 -1.093018  0.333008  4.366699  \n",
      "30167  108.736771 -2.967896  0.404663  1.281616  3.598389  \n",
      "30168  275.332062 -0.438660 -0.099907  0.269283  3.550781  \n",
      "30169  159.380676 -0.715363 -0.398224  0.556793  2.681824  \n",
      "\n",
      "[30170 rows x 12 columns]\n",
      "           Rpt          mjj         mll  njet  nlep  pdgid_l1  pdgid_l2  \\\n",
      "0     0.588571  1290.506470  198.609283   5.0   2.0     -13.0      13.0   \n",
      "1     0.336086   559.518250  177.509857   6.0   2.0     -11.0      13.0   \n",
      "2     0.420149   619.843750   59.186520   4.0   2.0      11.0     -13.0   \n",
      "3     0.734643   665.590149  155.680145   4.0   2.0      11.0     -11.0   \n",
      "4     0.433596  1475.237305  294.201843   4.0   2.0      13.0     -11.0   \n",
      "...        ...          ...         ...   ...   ...       ...       ...   \n",
      "7553  0.682076   433.814911  161.762146   6.0   2.0      11.0     -13.0   \n",
      "7554  0.782204   719.900513  204.333481   2.0   2.0     -13.0      11.0   \n",
      "7555  0.215456   986.111145  143.475830   2.0   2.0      11.0     -13.0   \n",
      "7556  0.253746   572.394043   81.091911   2.0   2.0      13.0     -13.0   \n",
      "7557  0.094356  1185.074341   70.881134   5.0   2.0     -13.0      13.0   \n",
      "\n",
      "            ptll   Zepp_l1   Zepp_l2   Zepp_ll    detajj  \n",
      "0     136.366882  1.116089  1.834106  1.475098  4.328369  \n",
      "1     124.331139  0.123322  1.349731  0.736526  2.599365  \n",
      "2     173.305862 -2.892212 -3.249512  3.070862  2.873535  \n",
      "3     141.330353 -1.258591 -1.186577  1.222584  3.782104  \n",
      "4     235.658615 -0.392090 -2.121338  1.256714  4.005859  \n",
      "...          ...       ...       ...       ...       ...  \n",
      "7553  108.902786 -0.017944 -0.932770  0.475357  3.033203  \n",
      "7554  131.995422  0.576569 -1.340088  0.381760  4.156250  \n",
      "7555   87.482880 -0.008087 -0.188360  0.098224  3.673340  \n",
      "7556   79.197441 -0.053833  0.473877  0.210022  3.595459  \n",
      "7557   99.522560 -0.281635 -0.484497  0.383066  3.601074  \n",
      "\n",
      "[7558 rows x 12 columns]\n",
      "isBKG\n",
      "            Rpt          mjj         mll  njet  nlep  pdgid_l1  pdgid_l2  \\\n",
      "0      0.258718   797.644653  158.335999   3.0   2.0     -11.0      13.0   \n",
      "1      0.106093   681.477905  110.376381   5.0   2.0      13.0     -11.0   \n",
      "2      0.345514  1063.075195  182.773041   4.0   2.0      13.0     -11.0   \n",
      "3      0.829721   710.275391   92.980476   5.0   2.0     -13.0      11.0   \n",
      "4      0.594790   334.772186  116.750854   5.0   2.0      13.0     -13.0   \n",
      "...         ...          ...         ...   ...   ...       ...       ...   \n",
      "30115  1.037049   472.241180  227.299652   3.0   2.0     -13.0      13.0   \n",
      "30116  0.471296  1061.369263  131.066040   4.0   2.0      11.0     -13.0   \n",
      "30117  0.856618   320.076843   97.441307   5.0   2.0     -11.0      13.0   \n",
      "30118  0.109525   538.265137   55.855526   5.0   2.0     -11.0      13.0   \n",
      "30119  0.326320   363.036804  105.035767   6.0   2.0     -13.0      13.0   \n",
      "\n",
      "             ptll   Zepp_l1   Zepp_l2   Zepp_ll    detajj  \n",
      "0       95.949997 -1.334961 -2.400391  1.867676  3.438965  \n",
      "1      109.168106 -0.699844  1.109177  0.204666  2.967072  \n",
      "2      371.016937 -0.363770 -0.384766  0.374268  2.669189  \n",
      "3       40.632839  1.170984  1.813721  1.492352  5.181152  \n",
      "4       88.207520 -1.078369 -0.835327  0.956848  3.021851  \n",
      "...           ...       ...       ...       ...       ...  \n",
      "30115   32.587646  2.830383 -0.114929  1.357727  4.521851  \n",
      "30116   85.310318 -2.624390 -1.764771  2.194580  4.749268  \n",
      "30117  126.425659  0.081573 -0.641937  0.280182  3.092834  \n",
      "30118   32.998287  0.611725  1.862183  1.236954  3.763916  \n",
      "30119   65.490448 -2.287918 -2.883133  2.585526  2.670746  \n",
      "\n",
      "[30120 rows x 12 columns]\n",
      "            Rpt          mjj         mll  njet  nlep  pdgid_l1  pdgid_l2  \\\n",
      "0      0.058838   553.380859   63.815094   4.0   2.0     -13.0      13.0   \n",
      "1      0.083872   451.707306   51.988937   5.0   2.0      13.0     -11.0   \n",
      "2      0.623096   485.333588  121.201881   4.0   2.0      13.0     -11.0   \n",
      "3      0.637398   373.100800  236.437119   4.0   2.0      13.0     -13.0   \n",
      "4      0.265464   650.201416  139.071747   8.0   2.0     -13.0      11.0   \n",
      "...         ...          ...         ...   ...   ...       ...       ...   \n",
      "7603   0.152315  1284.729614  107.785294   5.0   2.0      11.0     -13.0   \n",
      "7604   0.099884  1178.487915   58.049061   5.0   2.0      13.0     -13.0   \n",
      "7605  10.076541  1110.979614  246.452667   6.0   2.0      11.0     -13.0   \n",
      "7606   0.770655   648.448486  289.530182   5.0   2.0     -13.0      13.0   \n",
      "7607   0.119845  1345.346436   73.458191   8.0   2.0      13.0     -13.0   \n",
      "\n",
      "            ptll   Zepp_l1   Zepp_l2   Zepp_ll    detajj  \n",
      "0      61.002121 -1.055420 -2.676453  1.865936  2.826294  \n",
      "1      35.268883  2.289246  1.172546  1.730896  3.063354  \n",
      "2      39.086514  0.526001  1.592407  1.059204  3.814453  \n",
      "3      98.766220 -1.013306  0.955414  0.028946  2.733643  \n",
      "4     137.190659  2.925720  1.753113  2.339417  2.724731  \n",
      "...          ...       ...       ...       ...       ...  \n",
      "7603   88.638725  2.583984  0.506653  1.545319  4.955078  \n",
      "7604   96.551170 -0.772095 -1.504028  1.138062  4.424561  \n",
      "7605  118.143929  1.085205  1.030029  1.057617  6.661621  \n",
      "7606  138.834961  2.467346  1.359741  1.913544  2.922485  \n",
      "7607   32.622272 -1.498657 -1.836914  1.667786  5.092041  \n",
      "\n",
      "[7608 rows x 12 columns]\n",
      "(120467, 5, 16) (120467, 4) (30445, 5, 16) (30445, 4)\n",
      "[[[ 1.02637782e+02 -7.18859750e+01 -4.82670921e+01 ...  1.64944649e+00\n",
      "    1.95616245e+00  2.82043457e+00]\n",
      "  [ 4.63147551e+01  1.07379321e+00 -4.62980352e+01 ...  1.64944649e+00\n",
      "    1.95616245e+00  2.82043457e+00]\n",
      "  [ 2.02961538e+03  1.52818697e+02 -1.08507194e+02 ...  1.64944649e+00\n",
      "    1.95616245e+00  2.82043457e+00]\n",
      "  [ 1.26497362e+02  7.57712663e+01  9.63309762e+01 ...  1.64944649e+00\n",
      "    1.95616245e+00  2.82043457e+00]\n",
      "  [ 1.60278671e+02 -1.57740164e+02  2.84129038e+01 ...  1.64944649e+00\n",
      "    1.95616245e+00  2.82043457e+00]]\n",
      "\n",
      " [[ 9.44909579e+01  6.44027137e+01 -6.39174108e+01 ...  7.57812500e-01\n",
      "    6.25305176e-02  6.55566406e+00]\n",
      "  [ 6.59686087e+01  3.17153734e+01 -3.76025018e+00 ...  7.57812500e-01\n",
      "    6.25305176e-02  6.55566406e+00]\n",
      "  [ 2.31453720e+03 -7.90029108e+00 -9.58005545e+01 ...  7.57812500e-01\n",
      "    6.25305176e-02  6.55566406e+00]\n",
      "  [ 6.92879889e+02 -4.14155316e+01  8.48454150e+01 ...  7.57812500e-01\n",
      "    6.25305176e-02  6.55566406e+00]\n",
      "  [ 3.63202248e+01 -3.19404951e+01  1.72905610e+01 ...  7.57812500e-01\n",
      "    6.25305176e-02  6.55566406e+00]]\n",
      "\n",
      " [[ 6.25357017e+02 -5.03698607e+01  1.35882098e+02 ...  2.66577148e+00\n",
      "    4.81933594e-01  4.70361328e+00]\n",
      "  [ 2.87794171e+02  2.31504317e+01 -5.68898080e+01 ...  2.66577148e+00\n",
      "    4.81933594e-01  4.70361328e+00]\n",
      "  [ 9.48469452e+02  7.67807328e+01  2.63353210e+02 ...  2.66577148e+00\n",
      "    4.81933594e-01  4.70361328e+00]\n",
      "  [ 1.85431109e+03  1.76028613e+02 -1.42683346e+02 ...  2.66577148e+00\n",
      "    4.81933594e-01  4.70361328e+00]\n",
      "  [ 1.44095169e+02 -8.39599922e+00 -1.43850356e+02 ...  2.66577148e+00\n",
      "    4.81933594e-01  4.70361328e+00]]\n",
      "\n",
      " ...\n",
      "\n",
      " [[ 2.21192834e+02  1.06516524e+02  4.47542485e+01 ... -6.41937256e-01\n",
      "    2.80181885e-01  3.09283447e+00]\n",
      "  [ 4.21867328e+01  1.91843245e+01 -3.12359502e+01 ... -6.41937256e-01\n",
      "    2.80181885e-01  3.09283447e+00]\n",
      "  [ 1.29539437e+02 -1.06991675e+02 -5.75697897e+01 ... -6.41937256e-01\n",
      "    2.80181885e-01  3.09283447e+00]\n",
      "  [ 3.13596926e+02 -3.95152422e+01  9.72039757e+00 ... -6.41937256e-01\n",
      "    2.80181885e-01  3.09283447e+00]\n",
      "  [ 2.70286121e+01  2.70245115e+01 -4.70801257e-01 ... -6.41937256e-01\n",
      "    2.80181885e-01  3.09283447e+00]]\n",
      "\n",
      " [[ 3.32407824e+01  1.25055027e+01  3.01599948e+01 ...  1.86218262e+00\n",
      "    1.23695374e+00  3.76391602e+00]\n",
      "  [ 4.83153381e+01  1.63732267e+01 -1.41942217e+01 ...  1.86218262e+00\n",
      "    1.23695374e+00  3.76391602e+00]\n",
      "  [ 1.90731934e+02 -8.40255863e+01  1.59954121e+00 ...  1.86218262e+00\n",
      "    1.23695374e+00  3.76391602e+00]\n",
      "  [ 3.88590600e+02  6.33705094e+01  4.34996141e+01 ...  1.86218262e+00\n",
      "    1.23695374e+00  3.76391602e+00]\n",
      "  [ 4.58879204e+01  4.29336748e+01 -1.61987903e+01 ...  1.86218262e+00\n",
      "    1.23695374e+00  3.76391602e+00]]\n",
      "\n",
      " [[ 1.26967615e+02  8.52193379e+01  3.18439922e+01 ... -2.88313293e+00\n",
      "    2.58552551e+00  2.67074585e+00]\n",
      "  [ 6.53816112e+01 -2.08055192e+01 -2.00177852e+01 ... -2.88313293e+00\n",
      "    2.58552551e+00  2.67074585e+00]\n",
      "  [ 1.06302493e+02 -9.15945696e+01 -5.31001176e+01 ... -2.88313293e+00\n",
      "    2.58552551e+00  2.67074585e+00]\n",
      "  [ 6.03450878e+02  3.56314521e+01  6.71593562e+01 ... -2.88313293e+00\n",
      "    2.58552551e+00  2.67074585e+00]\n",
      "  [ 2.36100960e+01 -1.87024426e+01 -1.44102489e+01 ... -2.88313293e+00\n",
      "    2.58552551e+00  2.67074585e+00]]] [[[ 1.58650683e+02  1.28288092e+02  6.44830320e+01 ... -2.08911133e+00\n",
      "    1.29107666e+00  3.77636719e+00]\n",
      "  [ 1.11574899e+02 -6.41552623e+01  7.96541411e+00 ... -2.08911133e+00\n",
      "    1.29107666e+00  3.77636719e+00]\n",
      "  [ 7.32635522e+02 -7.77212135e+01 -3.61207058e+01 ... -2.08911133e+00\n",
      "    1.29107666e+00  3.77636719e+00]\n",
      "  [ 5.12308822e+01  2.91112125e+01  1.88886101e+01 ... -2.08911133e+00\n",
      "    1.29107666e+00  3.77636719e+00]\n",
      "  [ 2.08232174e+01  5.71821452e+00 -2.00226973e+01 ... -2.08911133e+00\n",
      "    1.29107666e+00  3.77636719e+00]]\n",
      "\n",
      " [[ 4.48415376e+01  2.65215935e+01 -3.41878229e+01 ... -1.54663086e-01\n",
      "    2.82897949e-01  4.66796875e+00]\n",
      "  [ 2.39040132e+01 -1.99849867e+01  2.10559478e+00 ... -1.54663086e-01\n",
      "    2.82897949e-01  4.66796875e+00]\n",
      "  [ 1.09074814e+03  1.03941524e+02  8.46731090e+01 ... -1.54663086e-01\n",
      "    2.82897949e-01  4.66796875e+00]\n",
      "  [ 2.14921958e+02 -4.05867980e+01 -4.94268216e+01 ... -1.54663086e-01\n",
      "    2.82897949e-01  4.66796875e+00]\n",
      "  [ 5.25297737e+01 -4.68753746e+01  2.37081501e+01 ... -1.54663086e-01\n",
      "    2.82897949e-01  4.66796875e+00]]\n",
      "\n",
      " [[ 8.03423830e+01 -5.37026210e+01 -4.85309443e+01 ... -6.93969727e-02\n",
      "    6.45050049e-01  2.98229980e+00]\n",
      "  [ 4.90332260e+01 -3.85995701e+01  7.81238982e+00 ... -6.93969727e-02\n",
      "    6.45050049e-01  2.98229980e+00]\n",
      "  [ 1.64434663e+02 -2.19132017e+00  1.28177635e+02 ... -6.93969727e-02\n",
      "    6.45050049e-01  2.98229980e+00]\n",
      "  [ 4.88516079e+02  1.00176649e+02 -1.99264323e+01 ... -6.93969727e-02\n",
      "    6.45050049e-01  2.98229980e+00]\n",
      "  [ 8.70388641e+01 -6.38348200e+00 -8.68044643e+01 ... -6.93969727e-02\n",
      "    6.45050049e-01  2.98229980e+00]]\n",
      "\n",
      " ...\n",
      "\n",
      " [[ 8.29258995e+02 -1.68404175e+02 -8.57942493e+01 ...  1.03002930e+00\n",
      "    1.05761719e+00  6.66162109e+00]\n",
      "  [ 3.50225639e+02  8.41677827e+01  2.95574270e+00 ...  1.03002930e+00\n",
      "    1.05761719e+00  6.66162109e+00]\n",
      "  [ 2.02780954e+02  3.29279643e+01  2.60308376e+01 ...  1.03002930e+00\n",
      "    1.05761719e+00  6.66162109e+00]\n",
      "  [ 1.53947676e+03  3.63898643e+01 -9.59530113e+00 ...  1.03002930e+00\n",
      "    1.05761719e+00  6.66162109e+00]\n",
      "  [ 8.88250046e+01  4.01798012e+01  7.92178327e+01 ...  1.03002930e+00\n",
      "    1.05761719e+00  6.66162109e+00]]\n",
      "\n",
      " [[ 2.82976228e+02  1.93135085e+02 -6.73316905e+01 ...  1.35974121e+00\n",
      "    1.91354370e+00  2.92248535e+00]\n",
      "  [ 8.25851642e+01 -7.90452937e+01 -1.17800516e+01 ...  1.35974121e+00\n",
      "    1.91354370e+00  2.92248535e+00]\n",
      "  [ 2.08135365e+02 -1.70880346e+02  1.14373031e+02 ...  1.35974121e+00\n",
      "    1.91354370e+00  2.92248535e+00]\n",
      "  [ 1.12303625e+03 -1.28110004e+01 -1.02353960e+02 ...  1.35974121e+00\n",
      "    1.91354370e+00  2.92248535e+00]\n",
      "  [ 1.87099884e+02  1.75851927e+02 -6.38941807e+01 ...  1.35974121e+00\n",
      "    1.91354370e+00  2.92248535e+00]]\n",
      "\n",
      " [[ 6.59789151e+01  3.79745764e+01 -4.09533033e+01 ... -1.83691406e+00\n",
      "    1.66778564e+00  5.09204102e+00]\n",
      "  [ 3.45696863e+01 -1.81458412e+01  1.50489795e+01 ... -1.83691406e+00\n",
      "    1.66778564e+00  5.09204102e+00]\n",
      "  [ 3.94524025e+02  9.96208009e+01 -1.08625809e+02 ... -1.83691406e+00\n",
      "    1.66778564e+00  5.09204102e+00]\n",
      "  [ 1.17653780e+03 -2.91371425e+01  6.86065506e+01 ... -1.83691406e+00\n",
      "    1.66778564e+00  5.09204102e+00]\n",
      "  [ 1.49986572e+02  2.80397876e+01  1.47342262e+02 ... -1.83691406e+00\n",
      "    1.66778564e+00  5.09204102e+00]]]\n",
      "[[1. 0. 0. 0.]\n",
      " [1. 0. 0. 0.]\n",
      " [1. 0. 0. 0.]\n",
      " ...\n",
      " [0. 0. 0. 1.]\n",
      " [0. 0. 0. 1.]\n",
      " [0. 0. 0. 1.]] [[1. 0. 0. 0.]\n",
      " [1. 0. 0. 0.]\n",
      " [1. 0. 0. 0.]\n",
      " ...\n",
      " [0. 0. 0. 1.]\n",
      " [0. 0. 0. 1.]\n",
      " [0. 0. 0. 1.]]\n"
     ]
    }
   ],
   "source": [
    "nclasses = 4\n",
    "features = df.columns[:len(df.columns)-nclasses]\n",
    "NDIM = len(features)\n",
    "#df = df[df['isMIX']==0]\n",
    "#df = df.drop(columns=['isMIX'])\n",
    "msk = np.random.rand(len(df_reco)) < 0.8\n",
    "df_train = df_reco[msk]\n",
    "df_val = df_reco[~msk]\n",
    "df_aux_train = df_aux[msk]\n",
    "df_aux_val = df_aux[~msk]\n",
    "print(df_aux)\n",
    "print(df_aux_val[df_aux_val['isMIX'] == 1])\n",
    "particles = ['l1', 'l2', 'j1', 'j2', 'miss']\n",
    "#X = df_train.values[:,0:NDIM]\n",
    "#Y = df_train.values[:,NDIM:NDIM+nclasses] # isLL, isMIX, isTT, isBKG\n",
    "\n",
    "#X_val = df_val.values[:,0:NDIM]\n",
    "#Y_val = df_val.values[:,NDIM:NDIM+nclasses] # isLL, isMIX, isTT, isBKG\n",
    "\n",
    "#inputs = np.array([np.float32,(len(df),6,4)])\n",
    "targets = ['isLL', 'isMIX', 'isTT', 'isBKG']\n",
    "inputs = np.ones((len(df_train),5,16))\n",
    "inputs_val = np.ones((len(df_val),5,16))\n",
    "labels = np.ones((len(df_train),4))\n",
    "labels_val = np.ones((len(df_val),4))\n",
    "baseline = 0\n",
    "baseline_val = 0\n",
    "baseline_targets = [0]\n",
    "for itarget,target in enumerate(targets):\n",
    "    print(target)\n",
    "    if target == 'isMIX':\n",
    "        baseline += len(df_train[df_train['isLL']==1])\n",
    "        baseline_val += len(df_val[df_val['isLL']==1])\n",
    "    if target == 'isTT':\n",
    "        baseline += len(df_train[df_train['isMIX']==1])\n",
    "        baseline_val += len(df_val[df_val['isMIX']==1])\n",
    "    if target == 'isBKG':\n",
    "        baseline += len(df_train[df_train['isTT']==1])\n",
    "        baseline_val += len(df_val[df_val['isTT']==1])\n",
    "    baseline_targets.append(baseline_val + len(df_val[df_val[targets[itarget]]==1]))\n",
    "    tmp_df = df_train[df_train[target]==1]\n",
    "    tmp_df_val = df_val[df_val[target]==1]\n",
    "    tmp_df_aux = df_aux_train[df_aux_train[target]==1]\n",
    "    tmp_df_aux_val = df_aux_val[df_aux_val[target]==1]\n",
    "    df_main = pd.DataFrame(tmp_df.iloc[:,0:].to_numpy(), columns=tmp_df.columns.values.tolist()[0:])\n",
    "    _df_aux = pd.DataFrame(tmp_df_aux.iloc[:,0:len(tmp_df_aux.columns)-nclasses].to_numpy(), columns=tmp_df_aux.columns.values.tolist()[0:len(tmp_df_aux.columns)-nclasses])\n",
    "    #df_main = pd.DataFrame(tmp_df.iloc[:,0:].to_numpy(), columns=tmp_df.columns.values.tolist()[0:])\n",
    "    #df_aux = pd.DataFrame(tmp_df.iloc[:,0:].to_numpy(), columns=tmp_df.columns.values.tolist()[0:len(tmp_df_aux.columns)-nclasses])\n",
    "    print(_df_aux)\n",
    "    for event,index in zip(range(len(tmp_df)),tmp_df.index.values.tolist()):\n",
    "        labels[event+baseline,...] = np.array([df_main[target][event] for target in targets])\n",
    "        for i,particle in enumerate(particles):\n",
    "            if particle != 'miss':\n",
    "                inputs[event+baseline,i,...] = np.concatenate((df_main['pt_' + particle][event]*np.array([math.cosh(df_main['eta_' + particle][event]),\n",
    "                                                                                                math.cos(df_main['phi_' + particle][event]),\n",
    "                                                                                                math.sin(df_main['phi_' + particle][event]),\n",
    "                                                                                                math.sinh(df_main['eta_' + particle][event])]),\n",
    "                                                                    _df_aux.loc[event].to_numpy()))\n",
    "            else:\n",
    "                inputs[event+baseline,i,...] = np.concatenate((df_main['ptmiss'][event]*np.array([1.,\n",
    "                                                                                                math.cos(df_main['ptphi'][event]),\n",
    "                                                                                                math.sin(df_main['ptphi'][event]),\n",
    "                                                                                                0.01*np.random.randn()]),\n",
    "                                                                    _df_aux.loc[event].to_numpy()))\n",
    "                \n",
    "    df_main_val = pd.DataFrame(tmp_df_val.iloc[:,0:].to_numpy(), columns=tmp_df_val.columns.values.tolist()[0:])\n",
    "    _df_aux_val = pd.DataFrame(tmp_df_aux_val.iloc[:,0:len(tmp_df_aux_val.columns)-nclasses].to_numpy(), columns=tmp_df_aux_val.columns.values.tolist()[0:len(tmp_df_aux_val.columns)-nclasses])    \n",
    "    #df_main_val = pd.DataFrame(tmp_df_val.iloc[:,0:].to_numpy(), columns=tmp_df_val.columns.values.tolist()[0:])\n",
    "    #df_aux_val = pd.DataFrame(tmp_df_val.iloc[:,0:].to_numpy(), columns=tmp_df_val.columns.values.tolist()[0:])\n",
    "    print(_df_aux_val) \n",
    "    for event,index in zip(range(len(tmp_df_val)),tmp_df_val.index.values.tolist()):\n",
    "        #print(event, index)\n",
    "        labels_val[event+baseline_val,...] = np.array([df_main_val['isLL'][event], df_main_val['isMIX'][event], df_main_val['isTT'][event], df_main_val['isBKG'][event]])\n",
    "        for i,particle in enumerate(particles):\n",
    "            if particle != 'miss':\n",
    "                inputs_val[event+baseline_val,i,...] = np.concatenate((df_main_val['pt_' + particle][event]*np.array([math.cosh(df_main_val['eta_' + particle][event]),\n",
    "                                                                                                math.cos(df_main_val['phi_' + particle][event]),\n",
    "                                                                                                math.sin(df_main_val['phi_' + particle][event]),\n",
    "                                                                                                math.sinh(df_main_val['eta_' + particle][event])]),\n",
    "                                                                    _df_aux_val.loc[event].to_numpy()))\n",
    "            else:\n",
    "                inputs_val[event+baseline_val,i,...] = np.concatenate((df_main_val['ptmiss'][event]*np.array([1.,\n",
    "                                                                                                math.cos(df_main_val['ptphi'][event]),\n",
    "                                                                                                math.sin(df_main_val['ptphi'][event]),\n",
    "                                                                                                0.01*np.random.randn()]),\n",
    "                                                                    _df_aux_val.loc[event].to_numpy()))\n",
    "    \n",
    "print(inputs.shape, labels.shape, inputs_val.shape, labels_val.shape)\n",
    "print(inputs, inputs_val)\n",
    "print(labels, labels_val)\n",
    "# print(inputs)\n",
    "# print(labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "03494827-30c3-459d-bc3e-86eb8f55e7b7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(5, 16) (120467, 5, 16)\n",
      "(120467, 5, 16) (120467, 4) (30445, 5, 16) (30445, 4)\n"
     ]
    }
   ],
   "source": [
    "lorentz_shape = (inputs.shape[1], inputs.shape[2])\n",
    "print(lorentz_shape, inputs.shape)\n",
    "print(inputs.shape, labels.shape, inputs_val.shape, labels_val.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "ac56b7b3-c0d4-463d-9aec-7541044c929d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(120467, 5, 16)\n",
      "['E', 'beta', 'eta', 'gamma', 'm', 'p', 'pair_cos', 'pair_dr', 'pair_ds', 'pair_dy', 'phi', 'pt', 'px', 'py', 'pz']\n",
      "30\n",
      "n features = 1815.0\n",
      "feature shape =  (120467, 1815)\n",
      "tf.Tensor(\n",
      "[[ 2.7227255e+01  4.0443047e+01  3.1164484e+01 ... -2.5487390e-01\n",
      "  -9.0808943e-02  4.0907353e-01]\n",
      " [ 4.9066437e+01  1.1382572e+02  3.8103065e+01 ... -5.9241492e-01\n",
      "  -2.1107131e-01  9.5082819e-01]\n",
      " [ 6.2194408e+01  1.5943414e+02  5.3895813e+01 ... -4.2505088e-01\n",
      "  -1.5144125e-01  6.8220824e-01]\n",
      " ...\n",
      " [ 9.2316055e+00  2.0716318e+01  1.3468748e+01 ... -2.7948982e-01\n",
      "  -9.9579334e-02  4.4858220e-01]\n",
      " [ 1.3761437e+01  3.1781992e+01  1.1414896e+01 ... -3.4013337e-01\n",
      "  -1.2118601e-01  5.4591531e-01]\n",
      " [ 1.5984755e+01  4.5235638e+01  2.9660007e+01 ... -2.4134700e-01\n",
      "  -8.5989445e-02  3.8736281e-01]], shape=(120467, 1815), dtype=float32) 120467\n"
     ]
    }
   ],
   "source": [
    "from lbn import LBN, LBNLayer\n",
    "#inputs = create_four_vectors((1200,6))\n",
    "print(inputs.shape)\n",
    "#print(inputs)\n",
    "inputs = inputs.astype('float32')\n",
    "lbn = LBN(30, boost_mode=LBN.PAIRS)\n",
    "#lbn = LBN(5, boost_mode=LBN.COMBINATIONS)\n",
    "\n",
    "# run build, which creates trainable variables and an eager callable in eager mode\n",
    "lbn.build(inputs.shape, features=[\"E\", \"pt\", \"eta\", \"phi\", \"m\", \"pair_cos\", \"pair_dr\", \"pair_dy\"])\n",
    "\n",
    "# show available features\n",
    "print(lbn.available_features)\n",
    "print(lbn.n_particles)\n",
    "# build certain features for all 10 boosted particles\n",
    "features = lbn(inputs)\n",
    "print('n features =', lbn.n_features)\n",
    "print('feature shape = ', features.shape)\n",
    "# print features (10 x E, 10 x pt, 10 x eta, 10 x phi, 10 x m, 45 x pair_cos)\n",
    "print(features, len(features))\n",
    "\n",
    "# lbn_layer = LBNLayer((5, 16), n_particles=5, boost_mode=LBN.COMBINATIONS, features=[\"E\", \"pt\", \"eta\", \"phi\", \"m\", \"pair_cos\", \"pair_dr\", \"pair_ds\", \"pair_dy\"])\n",
    "# # define the keras model and add the lbn\n",
    "# model = tf.keras.models.Sequential()\n",
    "# model.add(lbn_layer)\n",
    "# N_LAYERS=8\n",
    "# for layer in range(N_LAYERS):\n",
    "#     model.add(tf.keras.layers.Dense(1024, activation = 'relu'))\n",
    "# model.add(tf.keras.layers.Dense(4, activation = 'softmax'))\n",
    "\n",
    "\n",
    "# # compile the model (we have to pass a loss or it won't compile)\n",
    "# model.compile(loss=\"categorical_crossentropy\")\n",
    "\n",
    "# # compute features\n",
    "# features = model.predict(inputs)\n",
    "\n",
    "# print(features)\n",
    "# #model.fit(inputs, features, validation_split=0.2)\n",
    "# model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "bc26e027-8929-4c05-8995-bc477450a479",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.7.0-rc1\n"
     ]
    }
   ],
   "source": [
    "import tqdm\n",
    "from tensorflow.keras.models import Sequential, Model\n",
    "from tensorflow.keras.layers import Input, Activation, Dense, Dropout, InputLayer\n",
    "from tensorflow.keras.constraints import max_norm\n",
    "from tensorflow.keras import initializers\n",
    "from tensorflow.keras.models import load_model\n",
    "from tensorflow.keras.optimizers import Adam\n",
    "from sklearn.utils import shuffle\n",
    "import copy\n",
    "print(tf.__version__)\n",
    "from tensorflow.python.client import device_lib\n",
    "tf.config.run_functions_eagerly(True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "id": "52abbb0d-f65b-408f-9465-36991164434b",
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "from matplotlib.pyplot import cm\n",
    "import numpy as np\n",
    "import matplotlib as mpl\n",
    "%matplotlib inline\n",
    "mpl.rcParams['figure.dpi'] = 150"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "id": "2bd8aa6e-8ec9-462c-937d-0430fcdec8da",
   "metadata": {},
   "outputs": [],
   "source": [
    "class SimpleLBNNeuralNetwork ( tf.Module ):\n",
    "    #def __init__ (self, inputs_shape, lorentz_shape, nEpochs, learning_rate, N_NODES, n_layers, n_features, n_outputsC=nclasses):\n",
    "    def __init__ (self, lorentz_shape, nEpochs, nBatch, learning_rate, N_NODES, n_layers, n_features, n_particles, boost_mode, n_outputsC=nclasses):\n",
    "        #self.inputs_shape = inputs_shape\n",
    "        self.lorentz_shape = lorentz_shape        \n",
    "        self.learning_rate = learning_rate\n",
    "        self.optimizer  = tf.optimizers.Adam (self.learning_rate)\n",
    "        self.nEpochs = nEpochs\n",
    "        self.n_batch = nBatch\n",
    "        self.N_NODES = N_NODES\n",
    "        self.n_layers = n_layers\n",
    "        self.n_features = n_features\n",
    "        self.n_particles = n_particles\n",
    "        self.boost_mode = boost_mode\n",
    "        self.n_outputsC = n_outputsC\n",
    "        self.weights = self.build (self.n_features, self.N_NODES)\n",
    "        \n",
    "                                \n",
    "    # Define the structure of the model\n",
    "    def build (self, n_input, N_NODES):\n",
    "        # initializer = initializers.Ones()\n",
    "\n",
    "        # build certain features for all 10 boosted particles\n",
    "        #lbn_layer = LBNLayer(self.lorentz_shape, n_particles=self.n_particles, boost_mode=LBN.COMBINATIONS,\n",
    "        lbn_layer = LBNLayer(self.lorentz_shape, n_particles=self.n_particles, boost_mode=self.boost_mode,\n",
    "                    #features=[\"E\", \"pt\", \"eta\", \"phi\", \"m\", \"pair_cos\", \"pair_dr\", \"pair_dy\"])\n",
    "                    features=[\"E\", \"pt\", \"eta\", \"phi\", \"m\", \"pair_cos\", \"pair_dr\", \"pair_dy\"])\n",
    "        # Classifier model\n",
    "        self.model1 = Sequential()\n",
    "        self.model1.add(lbn_layer)\n",
    "        if n_input < 1024:\n",
    "            self.model1.add(Dense (n_input, activation = 'relu', input_dim  = n_input))\n",
    "        else:\n",
    "            self.model1.add(Dense (1024, activation = 'relu', input_dim  = n_input))\n",
    "        for i in range(self.n_layers):\n",
    "            self.model1.add(Dense (self.N_NODES, activation = 'relu'))\n",
    "        self.model1.add(Dense (self.n_outputsC, activation = 'softmax',input_dim = self.N_NODES))      \n",
    "        self.model1.summary()\n",
    "        return self.model1.weights\n",
    "     \n",
    "    # Performs the epochs loop and the actual training.\n",
    "    # Monitors the training and validation loss functions, both for the classifier and the adversary.\n",
    "    # Returns the classifier categorical accuracy.\n",
    "    def fit (self, X, Y, X_val, Y_val, show_loss = False):       \n",
    "        losses = []\n",
    "        losses_val = []\n",
    "\n",
    "        self.means = np.mean ( X, axis = 0)\n",
    "        self.sigmas = np.std ( X, axis = 0)\n",
    "\n",
    "#         for iEpoch in tqdm.tqdm(range(self.nEpochs)):\n",
    "#                 l, l_val = self._train (X, Y, X_val, Y_val)\n",
    "#                 losses.append ( l )\n",
    "#                 losses_val.append ( l_val )\n",
    "\n",
    "#         losses = np.array(losses)               \n",
    "#         losses_val = np.array(losses_val)\n",
    "        \n",
    "        X_indices = np.arange(X.shape[0])\n",
    "        X_val_indices = np.arange(X_val.shape[0])\n",
    "        np.random.shuffle(X_indices)\n",
    "        np.random.shuffle(X_val_indices)\n",
    "        X_shuffled = X[X_indices]\n",
    "        Y_shuffled = Y[X_indices]\n",
    "        #X_val_shuffled = X[X_val_indices]\n",
    "        #Y_val_shuffled = Y[X_val_indices]\n",
    "        X_val_shuffled = X_val\n",
    "        Y_val_shuffled = Y_val\n",
    "\n",
    "        batches = self.n_batch\n",
    "        for iEpoch in tqdm.tqdm(range(self.nEpochs)):\n",
    "            if batches != 0:\n",
    "                step = int(len(X_shuffled)/batches)\n",
    "                step_val = int(len(X_val_shuffled)/batches)\n",
    "                for batch in range(batches):\n",
    "                    if batch != batches - 1:\n",
    "                        l, l_val = self._train (X_shuffled[batch*step:(batch+1)*step,], Y_shuffled[batch*step:(batch+1)*step,], X_val_shuffled, Y_val_shuffled)\n",
    "                        #losses.append ( l )\n",
    "                        #losses_val.append ( l_val )\n",
    "                    else:\n",
    "                        l, l_val = self._train (X_shuffled[batch*step:,], Y_shuffled[batch*step:,], X_val_shuffled, Y_val_shuffled)\n",
    "                        #losses.append ( l )\n",
    "                        #losses_val.append ( l_val )\n",
    "            else :\n",
    "                l, l_val = self._train (X_shuffled, Y_shuffled, X_val_shuffled, Y_val_shuffled)\n",
    "            losses.append ( l )\n",
    "            losses_val.append ( l_val )  \n",
    "\n",
    "        losses = np.array(losses)               \n",
    "        losses_val = np.array(losses_val)\n",
    "               \n",
    "        plt.plot (losses, color = \"c\", label='Training set')\n",
    "        plt.plot (losses_val, color ='tab:blue', label = \"Validation set\")\n",
    "        plt.xlabel (\"Epoch\"); plt.ylabel (\"Loss\")\n",
    "        plt.legend(frameon=False)\n",
    "        plt.show()\n",
    "        \n",
    "        ca = tf.keras.metrics.CategoricalAccuracy()\n",
    "        ca.update_state(Y, self.predict_proba(X))\n",
    "        \n",
    "        return ca.result().numpy()\n",
    "\n",
    "    def save_weights(self, model_name):\n",
    "        self.model1.save_weights(model_name+'_weights_1')\n",
    "    \n",
    "    def load_weights(self, model_name):\n",
    "        self.model1.load_weights(model_name+'_weights_1')\n",
    "        \n",
    "    def save_model(self, model_name):\n",
    "        self.model1.save(\"saved_models/\"+model_name+\"_1\")\n",
    "\n",
    "    def reset_optimizers(self):\n",
    "        self.optimizer  = tf.optimizers.Adam (self.learning_rate)\n",
    "        \n",
    "    def set_epochs(self, epochs):\n",
    "        self.nEpochs = epochs\n",
    "        \n",
    "    # Applies a pre-processing to the input features and returns the classifier representation.\n",
    "    #@tf.function\n",
    "    def predict_proba (self, X):\n",
    "        ppX = (X - self.means)/(self.sigmas)\n",
    "        return  tf.clip_by_value ( self.model1 (ppX) , 1e-7, 1. - 1e-7 )\n",
    "\n",
    "\n",
    "    @tf.function\n",
    "    def _train (self, X, Y, X_val, Y_val):\n",
    "        Y_true = tf.cast (Y, tf.float32)\n",
    "        Y_true_val = tf.cast (Y_val, tf.float32)\n",
    "\n",
    "        with tf.GradientTape() as gt:\n",
    "            #gt.watch ( self.weightsC )\n",
    "            Y_hat = self.predict_proba (X)\n",
    "            Y_hat_val = self.predict_proba (X_val) #N3(N1(x)) validation set\n",
    "            \n",
    "            ## Training set\n",
    "            # Use the categorical cross-entropy as loss function for the classifier\n",
    "            cce = tf.keras.losses.CategoricalCrossentropy()\n",
    "            loss = tf.reduce_mean ( cce( Y_true, Y_hat ) )\n",
    "            \n",
    "            ## Validation set\n",
    "            cce_val = tf.keras.losses.CategoricalCrossentropy()\n",
    "            loss_val = tf.reduce_mean (cce_val( Y_true_val, Y_hat_val ) )\n",
    "            \n",
    "            # Compute the gradient of the overall loss with respect to the classifier weights\n",
    "            gradients = gt.gradient ( loss, self.weights )\n",
    "\n",
    "        # Apply the gradients\n",
    "        self.optimizer.apply_gradients ( zip(gradients, self.weights) )\n",
    "        \n",
    "        return loss, loss_val"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "id": "56f968c7-f597-4362-8077-0091116d0949",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(5, 16) (120467, 5, 16)\n"
     ]
    }
   ],
   "source": [
    "lorentz_shape = (inputs.shape[1], inputs.shape[2])\n",
    "print(lorentz_shape, inputs.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "270faf23-c06e-46c2-b801-40b74b8ee17b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_9\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " LBN (LBNLayer)              (None, 1815)              2100      \n",
      "                                                                 \n",
      " dense_72 (Dense)            (None, 1024)              1859584   \n",
      "                                                                 \n",
      " dense_73 (Dense)            (None, 137)               140425    \n",
      "                                                                 \n",
      " dense_74 (Dense)            (None, 137)               18906     \n",
      "                                                                 \n",
      " dense_75 (Dense)            (None, 137)               18906     \n",
      "                                                                 \n",
      " dense_76 (Dense)            (None, 137)               18906     \n",
      "                                                                 \n",
      " dense_77 (Dense)            (None, 137)               18906     \n",
      "                                                                 \n",
      " dense_78 (Dense)            (None, 137)               18906     \n",
      "                                                                 \n",
      " dense_79 (Dense)            (None, 4)                 552       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 2,097,191\n",
      "Trainable params: 2,097,191\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "1815.0\n"
     ]
    }
   ],
   "source": [
    "#LBNdnn = SimpleLBNNeuralNetwork(lorentz_shape=lorentz_shape, nEpochs=300, learning_rate=0.0001, N_NODES=150, n_layers=6, n_particles=5, boost_mode=LBN.COMBINATIONS, n_features=lbn.n_features)\n",
    "LBNdnn = SimpleLBNNeuralNetwork(lorentz_shape=lorentz_shape, nEpochs=30, nBatch=5, learning_rate=0.0001, N_NODES=137, n_layers=6, n_particles=30, boost_mode=LBN.PAIRS, n_features=lbn.n_features)\n",
    "print(lbn.n_features)\n",
    "# Save initial set of weights (before training) to re-initialize the ADNN in later steps.\n",
    "# Useful if we want to restart always from the same starting point during the optimization studies.\n",
    "LBNdnn.save_weights(\"my_simpleLBNDNN_model_polVBSvsWW_RECO_new\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "63f73e91-f4f2-44ef-8baa-b762ce41b801",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 30/30 [00:45<00:00,  1.50s/it]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x600 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "0.5174114"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "LBNdnn.fit(inputs.astype(np.float32), labels.astype(np.float32), inputs_val.astype(np.float32), labels_val.astype(np.float32))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "268e9453-3393-46d8-8774-f1faa849ff1b",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(30445, 5, 16) (30445, 4)\n",
      "[[0. 0. 1. 0.]\n",
      " [0. 0. 1. 0.]\n",
      " [0. 0. 1. 0.]\n",
      " ...\n",
      " [0. 0. 1. 0.]\n",
      " [0. 0. 1. 0.]\n",
      " [0. 0. 1. 0.]]\n",
      "[0, 7642, 15279, 22837, 30445]\n",
      "[0, 7642, 15279, 22837, 30445]\n"
     ]
    }
   ],
   "source": [
    "Y_predict_train = LBNdnn.predict_proba(inputs.astype(np.float32))\n",
    "Y_predict_val = LBNdnn.predict_proba(inputs_val.astype(np.float32))\n",
    "\n",
    "#print(Y_predict_val)\n",
    "print(inputs_val.shape, labels_val.shape)\n",
    "#print(inputs_val, labels_val)\n",
    "print(labels_val[baseline_targets[2]:baseline_targets[3]])\n",
    "print(baseline_targets)\n",
    "# for itarget, target in enumerate(targets):\n",
    "#     for ilabel, label in enumerate(labels_val):\n",
    "#         if label[itarget] == 1:\n",
    "#             #print(ilabel, label, inputs_val[...,ilabel])\n",
    "#             print(LBNdnn.predict_proba(inputs_val[...,ilabel].astype(np.float32)))\n",
    "#baseline_targets.insert(0,0)\n",
    "print(baseline_targets)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "66ab7ce2-9739-480f-88e6-2730bda10ed2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tf.Tensor(\n",
      "[[0.14102815 0.17965978 0.2591327  0.42017934]\n",
      " [0.3955581  0.30935824 0.24789244 0.04719126]\n",
      " [0.3751087  0.3235972  0.11226892 0.1890252 ]\n",
      " ...\n",
      " [0.19974153 0.11834439 0.47931612 0.202598  ]\n",
      " [0.0196231  0.13329843 0.21541455 0.63166386]\n",
      " [0.30157873 0.2654733  0.2246183  0.20832969]], shape=(30445, 4), dtype=float32)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x600 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABIQAAAJBCAYAAAA3J24LAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/MnkTPAAAACXBIWXMAABcSAAAXEgFnn9JSAACCFklEQVR4nOzdeZhcVZn48e8bSMBEBNKdID/ZFAVNEBAFMUTFZRRQ3EARRc0wOm6IiqCj44Kj48Igoo6MCwIqiAsCiiLihrIqi4ISEERBULTTCWvYAv3+/ji3oFKp6rW6qjv1/TxPPbfr3HvPfatuBbrePuc9kZlIkiRJkiSpd8zodgCSJEmSJEnqLBNCkiRJkiRJPcaEkCRJkiRJUo8xISRJkiRJktRjTAhJkiRJkiT1GBNCkiRJkiRJPcaEkCRJkiRJUo8xISRJkiRJktRjTAhJkiRJkiT1GBNCkiRJkiRJPcaEkCRJkiRJUo8xISRJkiRJktRjTAhJkiRJkiT1GBNCkiRJkiRJPcaEkCSNQkTsHhEZETmWfRPtuxMiYkl1/eu7cf3J1O33VpIkSZqqTAhJmhYi4svVF/vlEbHeGM67tjrv+5MZ31QUEVtFxOERcXi3Y1nbrK3vbZUcPDwidu92LJIkSZpcJoQkTRdfqbZzgReP5oSIeCbw2IbzJ8NdwB+rx1SyFfCh6jGc2yixXzfZAa1FtmJ07+10s4TymnbvbhiSJEmabCaEJE0LmXkRsLR6+q+jPK123D+BH7Y9qEpm/iYzH5+Zj5+sa0ymzDytiv853Y5FkiRJUmeYEJI0ndRG+TwvIh413IERsQGwb/X0a5l5/6RGJkmSJEnTiAkhSdPJ14FVlP92LRnh2P2AOdXPxwFExOyI2D8ivhYRv4uIZRFxb0T8PSJOj4g9xxPUaAoXR8TjI+KkiPhHRNwTEX+OiM9FxCYj9D0zIl4UEV+KiEsi4uaIuC8iBiLix9XriSbnXQ/8ou55NjxOqNs3YlHpiNg6Iv6vqsl0d0TcHhGXRcQHI+IRo3lfIuKxEXFcRNxYve83VbWhhk3ujWSKv7fjukbd+ZtFxKcj4sqIWFn3eb20at95mHNfEBHfjYi/VefdEhG/iog3R8SshmOXVPfpmVXTh5q8rq2Gez8lSZI0vUSmC69Imj4i4hRgH+BPmfm4YY47H1gEXJCZu1VtS4Djq0MSuB2YCcyuO/VTmXlok/52p0oCZGaMdl+1fw/gdKBWDPtOYF1gfeBm4D+pklbD9V25HQhgg7q27wCvzMyhuvMuBrYGNq6a/tkQ1rcy8+3VsUso78sNmblVk/hfAXytLv47gFl1z28Enp+ZVw0T+7OB7wMPr85/WPUeAPwd2CUz/9Z47ZFMg/d2XNeozt2hOrd2nQeq8zeq+gD4amYuaTjvYZT7tW9d8+3VNWvnXQTslZm3VOfsB3yGUqNrJrCS8l7W2zkzb0SSJElrBUcISZpuatPGHhsRz2h2QERsS0kG1R8PcAtwJLAYeHhmbpSZc4D/Rymkuwp4V0S8qF3BRsRmwLcoCYsrgKdm5gaU0Ut7Ur7kHzVMF3cBXwT+BdgwMzfMzEcAfcDbKV/0Xw4cVH9SZu4MvKzu+SMbHm8fZfw7ASdW8Z8PbF9dfzbwIkrSZXPgjIh4+DBdfRf4OfCE6vw5lFFcd1De/4+PJp6G2KbDezuua1Q+RUkGXQY8DZiZmXMpya5tgEOBK5uc9yVKMujPwKtr16XcsxdX7btSJcqq1/CtzHwkcEHVdGST12UySJIkaS3iCCFJ00pEzABuADajyeiI6phPAu+mjHDYNDMbRzq06vtQ4H+An2Xmcxv27c44RghFxDHAm4HlwILMHGjYvx3lC//MZuePIuZ9KSNMrsvMxzbsaxlXw3FLaDFCKCJ+BOwB/AnYITPvatj/JOA3lFE5h2Xmkc2uX22f22QUzNuAzwJ3A48YS62n6fDeTvAad1FGUi3KzAtH2d/TgV8BA8BTmiVxqkTa1ZTE2ZMy83d1+86hTBv7cGYePo6XJEmSpGnCEUKSppUqoXBC9XTfxlEpEbEO8Jrq6bdHmwyq1FYie1rVz4RUtWH2q55+oTFhAZCZfwBOmcBlajFvHRGPnEA/a4iIjYDnV0//pzEZBJCZvwVOrZ7uP0x3H2tMBlW+V20fBrScAtgktmn93o7yGrdW203H0N+/VduTWo3oycybeChR9/xmx0iSJGntZ0JI0nR0PKUGUG3aUb09eegL9Fca9hERm0TEhyPiwohYHhH3x0OFj2vL2s/mobotE/FoSk0WKNOlWhluHxGxQUQcFhG/rIoR31cXc32SZrMJxttoJx6qOfPTYY77SbXdPiJmtjjm1y3a/17389wWxzQzbd7bCVzjB9X2qxHxqYh4ZkTMZni7Vdt/i1Jku+kDqI2A23K8r0uSJEnT27ojHyJJU0tm/rma2vIs4EBWT/wcWG2vzswL6s+LiKcBZ1KK8tbcSflSnsA6QH/VPgcYnGCo8+t+Hq5g8k2tdkTENsDPWD1ZcBdl9EhtxE1tNa05tNdY41+XkqRpLLJMZt7R7MTMvD8eWmSrVTKpHbGtoRPv7QSv8W7gsZTP+SHV44GI+B1lZNGXmhTi/n/V9hHVYyQjJZgkSZK0lnKEkKTpqpYEWlR96SYi5gEvrNqPqz84ItYFTqYkg34H7EWpWbNBZm5SFdTdtf6UyQt9TI6nJBOupxQf7svMOZk5v4q5fsn2qRLzdNGJ93bc18jMWzPz2cDTgSMoRb3vB54MfBC4NiIap+nVpjq+OTNjFI8l43xdkiRJmuZMCEmarr7LQzVWaqOCDqCMMrmfsux2vadRpsc8ALwwM3/UZNRKu+vE1Ne1eVTLo1rsi4jNeWi1tP0z85TMXNFw2GTVtoHV4x9uylRt3/1AY3yTZcq/t+26Rmael5nvyczFlITmi4HfU+ouHRcRm9Qd/o9q61QwSZIkDcuEkKRpKTPvAb5RPX1tVQT6X6vnP8jMxmlLm1fbZU2m2dQ8t0X7eP2FhxIkzxrmuGe3aN+87ufftjhmuJgfLOIcdfOyxuCyuj6eM8xxtRguz8xV47jOeEyH93ai11hDZt6Tmd/noWXv1wcW1x1yfrV9IeNTe12ONpMkSVrLmRCSNJ3Vpo1tCnwAeGL1/Lgmx95WbTdpGFEBPLgU98HtDC4zE/h29fRNEdHfeExELAD2bdHFbXU/79Dk3A2A9w8Twu11P280bLBNZOatwI+rp4c1K2gcETsA+1RPTx7rNcZrmry3475GRKwbEcP9P/ruup/rV2/7UrXdLiLePMz5RMSciJjV0Fx7XRsNd64kSZKmPxNCkqatzLyMUg8ISkII4GZK4ehG5wErKSMfvl1Xd2idiHg+cA6lsHS7fRy4g1Ks+icR8ZTquhERzwN+xOorTdW7Cvhr9fNxEfHk2o6qQPY5DL8a2jXAfdXPrx/nKKH3A6soxY1/HBFPrK4/IyL2orzX6wLXAV8cR/8TMdXf24lcYzNKjaD3R8STqhpYtXO3B06snq4Eflnbl5m/pNQtAvh8RHw6Ih5Td+56EbFrRBwB3MDqxbkB/lBt94qI4abiSZIkaZozISRpuquNEqr99+yrmflA40GZeRtwaPX0GcAfI+IOyipjZwEb8tCUs7bJzL8C+wP3AjsCF0fE7ZQv8j+m1Dw6pMW5Q8BbKbV5FgKXRMTKiFgJXABsC+w3zLXvAr5ePT0CuDMiboiI6yPiyFHGfxnwGkryYzFwRUTcVsX/Q8qqVjcCe2fmnaPps12m+ns70WsAjwE+Qpm6d09ELI+Ie4HLgd0p92RJk7pEbwKOpSQ/3wFcFxF3RMQKSoLsQuAwoI81k6BfBe6hJAD/Wi1Tf331GK6OlCRJkqYZE0KSpruTKF9ga5pNFwMgM78AvIAyMuNOysiWvwGfo0zp+f1kBJiZPwR2Ar5JKYY8i7I0+/8CT6LUw2l17g8oCawfUoporwsMUkaBPDkzfzbC5d8KHM5Dr20LSsHhNaZYDRPDtygJjS9SRgKtR0ly/A74ELBdZl412v7aaaq/txO4xt+AFwGfBi6ijHx7OOV9Xwp8nvK+n9Lkdd2XmW+gFLQ+gXLP1qnOH6B8/v8L2L6xnlZmXkupyfR9YBklabRl9VgXSZIkrTWilGGQJEmSJElSr3CEkCRJkiRJUo8xISRJkiRJktRjTAhJkiRJkiT1GBNCkiRJkiRJPcaEkCRJkiRJUo8xISRJkiRJktRjTAhJkiRJkiT1GBNCkiRJkiRJPcaEkCRJkiRJUo8xISRJkiRJktRj1u12AN0SEf8AZgM3djsWSZIkqU02B+7KzEd2OxBJ0tQWmdntGLoiIm5fb731Nth66627HYokSZLUFtdddx333nvvHZn5iG7HIkma2np2hBBw49Zbb73gyiuv7HYckiRJUlssXLiQpUuXOgJekjQiawhJkiRJkiT1GBNCkiRJkiRJPcaEkCRJkiRJUo8xISRJkiRJktRjTAhJkiRJkiT1GBNCkiRJkiRJPcaEkCRJkiRJUo8xISRJkiRJktRjTAhJkiRJkiT1GBNCkiRJkiRJPcaEkCRJkiRJUo8xISRJkiRJktRjJiUhFBF9ETEQERkRfxpnHxtHxGci4oaIuLfaHh0RG7U5XEmSJEmSpJ4yWSOEPgX0j/fkiOgHfgMcDNwPnA7cAbwd+HVEzG1DjJIkSZIkST2p7QmhiHgO8DrgyxPo5mjgscCpwLaZuV9mbgd8DtgGOGqicUqSJEmSJPWqddvZWUQ8DPgisBQ4Evj3cfSxKbA/cB/wlsy8v273YcArgQMi4t2ZOTDxqCVJkiRNSREzgL5uhzFOy8kc6nYQktRKWxNCwIeAxwDPBFaNs489KCOXzs3Mf9bvyMx7I+IM4EBgL+CE8YcqSZIkaYrrA6brH4HnA8u6HYQktdK2KWMRsT3wLuD4zDx3Al3tUG0va7G/1r79BK7R0zKHuO++ZVPu4R9QJEmSpKJaoOf6CZy/e9VHRsRtEbH+MMceU3fsCQ37lrRo/1nV/t5R9Pu18b6O4Uz0PZqOIuL66nVvNYpja5+BcyY/sumv9t52O45OassIoShDOY8FbgXePcHutqi2N7XYX2vfcoLX6VmrVi3nggvmdzuMNSxaNMCsWfO6HYYkSZK0tnkE8ELglMYdETETeMU4+nwj8HvggxHxncxcbXXpiNgNeBMwCBwyjv4lTbJ2jRB6G7AzcFhmLp9gXw+vtne12L+y2m4wms4i4spmD2DrCcYpSZIkSZPlCcBz2tDP74EHgANa7N+TMjWv1QyNpqoE0IeB9Sl1ZB8UEbOALwEBvDMzB8cYs6QOmPAIoYjYAvgo8MvMPGHCEUmSJElSawsoo06mkn7Kwjptk5lXt6mrAeBvwJ4RsXFm3tKw/wBgCPgGsNMY+z6SsiDQsyNiSd33wfdR7tPZmXniuCOXNKnaMULo88AsynDAdriz2s5usX9Otb1jNJ1l5sJmD+C6iQYqSZIkqeMGyVw2pR6TkKBqVR8nIhZFxOkRcUNE3BsR/4iI30TEJyLi4U26AjiR8p1ttalhEfEIYG/gHErSaEyqFaFfT0kofSoi5kfEE4D3UmZ8jPk7YhT7R8RPImJ5RNxT1Xb5dkSMOGKq7vxvRsQ1EbEyIu6o3qO3VOVOmp3z6og4LyL+WV3zxoj4aUS8teHYWVU/F1fx3VXF94OIeOVYX+90ERHbRcSJEfHn6v1ZFhG/i4ijq5XCa8eN5/0/vPq8L4mIJ0fEjyLi1ohYUd33zarj5kTEEdX7fU9E/CEi9m3SX6120gkRsWm1/WdE3B0Rl0XEa8fx+jePiP+NiOuqa6+o7vmisfY1lbSjhtALKbWDvhAR9e21omWPqiti9crM/McI/f212m7WYn+t/Yaxhanh7LzzUmbO7O/Y9VatGuTiixd07HqSJEnSdBcRewOnU6Zi/Qa4ANgIeBzwHuALPPQH9nqnU0pvvJrVp3ftQ/nedtJ4Y8rMiyPis8A7gM8Am1OST+/LzL+Mpa+IWAc4GXg5cB9wPvDPqs8XVP3+bIRu1qOMdlpOGbV1GWVK3CLKYIZdgCUN5xwBHArcC/yKkuB7JGUho8dW59WcBOxLGaBwLnA78ChgMaX8yTfH8pqng4h4MnAe5bNyBfA9ygCOxwBvp3y+bq4OH8/7X/NUymf4D8CPKSPWXg7sEBG7AD+h1BL+FWVU3jOBb0fEnpn54yb9zQUuqmI6B9gYeBbw1Yh4TGYePsrX/zTgh9X5f6x+ngc8H9gjIl6dmd8aTV9TTbuWnd+IcjOaWb9uX8vK9nUur7athivW2q8YVWQalZkz+y3oLEmSJE1th1Jmeeybmd+t3xERO1O+hK8hM1dGxOnAqyJiy8ys/XH9AOAeSrHpvSYQ1/uBlwK1ETKXAUePo5/3UhIAS4EX1ieUImJDYMdR9HF/FcsPM3NV3fnzgDOB10XEcZn5q6p9fUpN3DuAHRquuS7wtLrnj6Ykg24AnlxfP7fq50ljfcHTxMGU7/KHZuan6ndExOOB2+qaxvT+N3gT8ObM/EJ1zszqnOdSkp//AB6TmSur/f9GWdzqfZQEUqO9KUmkl9adszPwc+ADEfH9zBy2dlY1iu67lMLsB2TmSXX7ngKcDRwbET/PMlpwWpnwlLHMjGYP4NHVIdfVtV8/ii7Pogw5fHpErLYUVkSsR7mpD1A+GJIkSZLUK2p/wf1p447MvDgzhyurcRJlZNGrACLiUcDuwBmZeftEgqq+bNePPHpHZj4wlj6iFKJ+V/X0wMbRRZl5W2b+chSx3J+Zp9cnI6r2ZZSEE8CL63Y9gjKC5Lom17w/M8+ta6q9/79tXEwpM+/JzAtHim+aGu5zd3Vm3lz3fKzvf73zasmg6pxVwOeqp4+nJItW1h1/AmU019Oq5FGjIeBt9edk5sWUkUozgLe0iKPegcCmwNH1yaCqr0uAj1BGhrUq2j6ltWuVsTGLiIMi4uqI+Hh9e/VhOpkyHPCYKitbcwTlw3hiZg50LlpJkiRJ6rpLq+3XI2LnZvVYhvETSoHpV1fPX0X5Pjjhos8RsTFlyljNy8fRzVMoM08uz8xftyGmHSPi3RHx+Yg4PiJOAN5c7X5c7bjqe+VNwI5R6jA9Zphur6ZMvXtBRBwWEf9vonFOE7XP3eer+jwjzjQa7fvf4OwmbX+uttdn5jX1O6qk4w3ATMoUska/y8w/Nmk/udo+fbjXUHletT21xf5awnCXUfQ15bRryth49APbUrJtjd4B7EqZ03p1RFwCLAS2A64FDulQjJIkSZI0VbwPeCJl1sTewC0RcR7wfcofze9pdWJm3h8R3wQOjogdKSMalgM/akNcRwLzKV+09wDeGhFfr0ZjjNbm1XZCi/9UI41OoKx+1soGDc9fR6n98x7gPRFxA/BL4JuZ+eD7k5m3R8QbgC9RBiscERHXAL8Avp6Z548yxhOaNJ+emaeP5vwu+B9KjaTdKa/1zoi4kFJL54TMfHDK2Djf/5pmhc3vHGZf/f71muxrVXf4+mo7moTeVtX2/IaayY06V5C3jbo2Qmg4mTlIybB9jjJS6KXAhsBngV0yc0UXw5MkSZKkjsvMGykjaZ5P+a50IyUx9GXgiojoG6GL2pSXj1MKJn+7cWrPWEXEsyjTav5GqQHzHsr3zC+NZiTJJDiEkoz4PbAnsAkwqyprsm11zGrf7DPz55Ti0a8Gvk6J/7XAmRFxSsOxJ1OKKb8B+A5lVNMbgfMiYrX6OsN4XZPHjmN4jR1VTSl8NmVEzRGUGk/PptSJ+mNE1I/4GfP7X2domDCG2zeZajmTU4CvDvMYqdj5lDRp/0CrekEtU2hVRe/Dh9m/glK86uA2hyZJkiRJ01K11PvZ1YOI2BI4jvIF/T3Au4c59zcRcS1lFA9McLpYVUi5VjvobdUImmMpyZTFlJkfR46yuxur7dYTiYkymABg/8y8smFfy+lgVdLjG9WDiNiVkvDZJyL2yswz645dRilmfGyUYSPPB74FHFIVTG68buO1hh1qMhVlZlJWGjsPoKr3ezQl+fPfwCuqQ8f1/k+SLUdo//so+riJksj6RGZeOtLB082UHCEkSZIkSRpZtWLYJ6un243ilOMoU8V+l5kXTPDyH6TUgzk9M0+r4kng3ynLxn+4SliNxqXArTy0xPh4bVxtb2qy7xVN2prKzIsoo4VgmPc1i7Mo06eglDpZ61W1lw6vnta/P215/9tkx4bRSzW11fDOG0UfP6m2Lx32qGnKhJAkSZKk6aSfiHlT6tGh+iER8c6IeGSTXbUl429ssm81mfmJzOzPzAktkR4R2wOHAbcDBzVc4ypKkmo2cMxo+svMe4FPV0+/0phIiogNI+KZo+iqVnj4TQ3n70sZuURD+xYRsSQiZje0rw88q3p6Y9X2pIh4WVUnp/7YucBT649dm0TEmyLi0U12Nfvcjen9n2QzgM/V39uIeDLl85rA/42ijy9SirG/OyL+vbGQe0SsGxHPj4jRJGOnnG4WlZYkSZKksVra7QC66EPAkRFxOWWxnQB2ALYBVjD66VkTUn0p/jLl++T7MrNZwd//BvYD9oqIV2Tmt0fR9ceAJwEvAa6JiHMpX8Y3B3aijNYYaen5IyhT4j4RES+nJCgeR6m9dCRwaMPxc4HjKStoXUIZ2TIHWERZ4foSHlphakvgu8Bt1bH/oNQQegalUPIZHVx6/rSIuLfFvh9m5kfqnu8UERcN09dbMvOyYfa/Cfi/iFgKXAXcT1kGfgfgHuC/6o4d6/s/mX5QxXhdRPyKUpf42ZRVyT5aLRs/rMy8NSJeDJxBSQ69PyL+ANwCPJLyudyIMoLoD5PxIiaTCSFJkiRJmh7eRvmy/WRKwV4oozOOAo5qkZiZrDh2AS6ixSiLzLw3It4E/Bz4TEScnZm3DtdptRLaPsBrKIWqnwKsD9xM+XI/4oiOzPxVRCymJKSeREmW/Z6ygvVlrJmQuA54F/AcYEH1ulYCf6EkqL5UjV6ier3vpyQVtqUUWb4FuAL4ChOsyTRGOw6z7+qG5xvw0AimZh4xwrU+QEnSPZXyPs2iJM6OBY6sX9p9HO//ZFpOWb38k5Q6T4+gJJSPzswTRttJZl4UEU8E3gm8AKiNVLuZkqA8Dfhp+8LunChTPHtPRFy5YMGCBVdeOWy9r7XSffct44IL5q/WtmjRALNmzeupGFRnaAiWL+92FM319cEMZ7dKkjQaCxcuZOnSpUszc+2oY1KmYw10O4xxmk8pPiypgyJid+AXwFczc0lXg5niHCEkqSSD5s8f+bhuGBiAeSYKJUmSJKmdTAhJkiRJmqqWA1P0r1YjmqLDryWpMCEkSZIkaWrKHAKcdiVJk8CEkCRJkiRJWitk5jmUFfg0AhNCkppbuhT6+zt7zcFBWLCgs9eUJEmSpB5kQkhSc/39FnOWJEmSpLWUazlLkiRJkiT1GBNCkiRJkiRJPcaEkCRJkiRJUo8xISRJkiRJktRjTAhJkiRJkiT1GBNCkiRJkiRJPcZl5yVJkiRNSfHhmAH0dTuOcVqeH8qhbgchSa2YEJIkSZI0VfUBA90OYpzmA8u6HYQkteKUMUmSJEmaYiIiI+L6CZy/e9VHRsRtEbH+MMceU3fsCQ37lrRo/1nV/t5R9Pu18b6O4Uz0PZqOIuL6unu1/zDH7VJ3XDbZv0Z7RLy2av99RMxs0e+uEfFARAxGxLyJv6K1w3T9LJoQkiRJkqS12yOAFzbbUX3xf8U4+nwjcA/wwYh4bJN+dwPeBAwCh4yjf43s1cPsO2CsnWXm14CfAtsB727cX31WvkzJI7wrMx0BN82ZEJIkSZKkqecJwHPa0M/vgQdonSDYkzI177KxdJqZfwI+DKwPfLF+X0TMAr4EBPDOzBwcY8wa2W+B50dEf+OOiFgX2A9YCtw7xn7fCNwNvD8iHtew7z2UZNFPM/OrYw9ZU401hCRJkiRNJwsoo06mkn7Kl++2ycyr29TVAPA3YM+I2Dgzb2nYfwAwBHwD2GmMfR8J7A88OyKWZOYJVfv7KPfp7Mw8cdyRazgnUd7//YDPN+x7HqWG1dHAh8bSaWb+OSI+BBxBSeo9CyAitgXeT0kWvWkigWvqMCEkSZIkaToZzA9Nrakq8eFof5+lvssNmblVQ/siynSeJwGPBG4B/gr8HPhoZt7ZpLsTgT0oU8MeHM0TEY8A9gbOoSSNxiQz74+I1wMXAZ+KiDMpo43eC9zFOBIHERHAK4EDKQmqOcA/gN8AX8zMn43y/BdX5z+KkvC6CjgB+ELm6qu/Vee8Cngz8DhgQ0pB8D8Cp2Xm5+uOnQW8HvhX4DHAwyhJtz8AJ2bmN8f6msfp+5RkzwGsmRA6AEhK0mhMCaHKpymJvt0j4t+A4yjJofWA92TmdWPpbLSf2arO1auBFwFPBDaljHC6Ajim2Xtb1bZ6HSVxtS7wweo6dwPfAw7NzNsiYj7wEcrUyb6qz3dn5jkN/S0BjqeMfjsZ+BiwO2Uk3O+A/87MM8f4+p9AGV31HEqi7lbgF8BHMvPKsfTVbk4ZkyRJkqRpICL2Bs6lfGG+GTiVMnVoLuUL5xrThyqnAytZs+bMPpQvuieNN6bMvBj4bBXDZyg1ZmYBH8zMv4ylr4hYB/gWZbTSM4DLgdOAm4AXAG8bRTfrVec/l5JIOoOSsFpISZwc1+ScIyhJs6dU1zwVuBbYHjis4diTqn62rfr9HiW5sZjOjpy5mxLnrhGxda0xIh5OSYadl5l/HU/HmXk/8AbKVMP/oYwMegYlIXLUWPoa42d2K+BYyn24nvLe/g7YFTg5Ig4f5lIvBc6iTFM8i5JIej3wvWpa3YXA86tYfgfsDJwVEU9s0d/WwK8pyaWzgUuApwE/iIh/HdWLByLiJdXrfR1lZOP3gb9QkrO/iYhnjLavyeAIIUmSJEmaHg6l/FF/38z8bv2OiNgZWN7spMxcGRGnA6+KiC0z84Zq1wGUwtCnAHtNIK73U76Qv7J6fhllutJYvRd4OWX63QvrE0oRsSGw4yj6uL+K5YeZuaru/HnAmcDrIuK4zPxV1b4+JdF0B7BDwzXXpSQBas8fDewL3AA8OTOX1+1bn5I86KQTKYmGVwP/VbW9FJhd7Ru3zLw0Io4G3lX1/QDwhipZNBZj+cwuA/4F+FlmZt1xj6aMJvpARJyQmdc3uc5BwIsy84fVORsAFwDPBH5Z/Xxg7TMRER+hfG4PpbyHjQ4Avgb8W+01R8QLKcnV/42IszNz2FF1EbEV5T6sonyef1q3bw9KcujEiHhsZt43XF+TxRFCkiRJkjQ91Jb5/mnjjsy8ODPvGObckyijJ14FEBGPokyFOSMzb59IUJm5ktULS78jMx8YSx/VVKx3VU8PbBxdlJm3ZeYvRxHL/Zl5en0yqGpfRkk4QRlBU/MIyqii65pc8/7MPLeuqfb+/7Y+GVQde09mXjhSfG32c8qom/qRXwdQRsd8pw39H0WZegbw1cy8ZBx9jPozm5nLM/On9cmgqv0vwH9T8hd7t7jON2rJoOqcOyij1QA2Aw5u+EwcSXltz2zR352Uz/GDCbDM/AEleTqbMmVwJO+gTHl8b30yqOrrLOD/gM0po9+6woSQJEmSJE0Pl1bbr0fEzhExlu9zP6HUuqklD15F+T444aLPEbEx5ctvzcvH0c1TgI2AyzPz122IaceIeHdEfD4ijq9qzby52v3g6lmZOUCZkrZjRHwiIh4zTLdXU6bevSAiDouI/zfROCeiqoV0MrBN9Xl4JKVOzZlNioePx4cpSUQoK5ptMI4+xvyZjYjFEfH+iPi/untX+0w1rnxWc3aTtj9X20sa34/MvA1YQalT1LS/Fu/hydX26S1fwEOeV21PbbG/lmzcZRR9TQqnjEmSJEnS9PA+SrHdvavHLRFxHtXUk8y8p9WJVQHobwIHR8SOlJEky4EftSGuIynFck+mFK9+a0R8vaovNFqbV9sxFSxuVI00OoFSFLmVxsTG64BvUmravCcibqBMM/pmZj74/mTm7RHxBkqB5SOAIyLiGkqB4K9n5vmjjPGEJs2nZ+bpozm/wYnAIZRE3w3AOrQnyfcM4N8oI5B+DbyEMkrn4DF2NerPbDUt8FTg2cP01yop1Wz61p3D7Kvt72ux74YW7ddX29EkA7eqXb/ULW+pVe2vSecIIUmSJEmaBjLzRspImucDnwNupHzJ/jJwRUS0+nJbUyse/XFKweRvN06tGquIeBZlRbC/UYoqv4fyPfNLVQ2eTjuEkgz6PbAnsAkwKzODUggaHhr1AkBm/hx4LCWp8nVK/K8FzoyIUxqOPZmyutgbKNOyNgLeCJwXEZ8aZYyva/LYcQyvsT6e31JWUHtlFfOtwA+HO2ckEbEeJekVlPpKb6KsDPbWqu7PWOIby2f2k5Rk0C8p0xn7gXWre/f8WngtLjXUon2kfZOplm/56giPCY+IGy9HCEmSJEnSNFHVNDm7ehARW1JWzno2JRnz7mHO/U1EXEsZxQMTHElSFVKu1Q56WzWC5lhKYmIxZRrZkaPs7sZqu/WwR43spdV2/yZLerecDlbVUfpG9SAidqUkfPaJiL3qlxqv6hEdCxxbLVn/fMrqaIdUBauHXUq8SnC004mU0TubAF/OzHsn2N/7Kcmz79cKQUfEuylJnC9HxFPGUlx6DJ/Zl1KKV7+oSV2r4abyTYYtR2j/+yj6uInyeX5XY82pqcKEkCRJkqTppD8+3O7v0xPWtSkfmXlDRNRGVmw3ilOOo6ysdGNmXjDBy3+QUtPl9Mw8rYonI+LfKUt7fzgivlO3qtlwLqWMbtkhInbJzN+MM6aNq+1NTfa9YrSdZOZFEfF1SiHq7SgrlDU7LinLl/+QMjJpITBsQmgSfAN4J2X0zNcm0lFEbEdJ0twBvLVu11coib6nV9f6n/FeY5jP7MbA7S2KnI/63rXJ8yJio8y8taG9tpLeeaPo4yeUhNBLKQnEKccpY5IkSZKmk6WU4shT6bF0Ul9xJSLeWRUOblRbMv7GJvtWk5mfyMz+zJzQEukRsT1wGHA7Zcnv+mtcRZn+Mxs4ZjT9VaNaPl09/Uo1iqT+ehtGRKsVoepdU23f1HD+vpSEBg3tW0TEkoiY3dC+PvCs6umNVduTIuJlVZ2i+mPnAk+tP7aTMvP6zJxX3dfRJCqaqgo+fxmYCfxnZj6YVKsSX28E7gMOr5ZUH02fY/nMXgNsHBH7NfbBQ/eiUx4OHFU/7TEi9qQkpu4Gjh9FH5+qjj0yIl7WuDMi1ouIfSNiszbFPGaOEJIkSZKk6eFDlC+XlwPXUkaE7ABsQ1kxabTTsyakLnGwLvC+zGxWtPe/gf2AvSLiFZn57VF0/THgSZQCxtdExLmUhNvmwE6UERcjLT1/BGVK3Cci4uWUJMPjKHVsjqSMjqo3l/Ll/vMRcQllZNEcYBFlyfRLeGiVqC2B7wK3Vcf+g1JD6BmUYsdndGHp+XZ6C7ArpabN5xt3ZuZVEfEJysiwY3goqTOcsXxmP06Z/vbNiHgr5V7sADyekix85/he1ricBLwM2D0ifk1ZjewZlPgPrk+WtZKZf4qI/SkjuL4bEX+i1HtaCTyK8pmeQ/nMj9jfZHCEkCRJkiRND2+jrIY1m1IweQ/gfuAoYPvMvLaDcewCXAT8X7MDqhE/tVE6n4mIjUbqtKo1sw+wpOr7KZQv5ZsBPwCOHkUfv6LUL/o5pe7MCymjWvahSZKDsqrZu4BzgC2q6y2mrDL1TuCZdTV5LqLU17mUUmPn5VWMV1AKa+8zUnxTVTVK5WOUz9MbqiXtm/kYJcm2Z+NInhZG/ZnNzJOAF1De5x2r4/9OmVr2/bG/qgn5E/A0yr19Pg993vfOzFFP/8rM71EKuB8DJPAvlNc4HziDMuKoIyMMm4ky8qv3RMSVCxYsWHDllZ2e3tl99923jAsumL9a26JFA8yaNa+nYlCdZctg/ur3g4EBmNfh+zFV4pAkaZpauHAhS5cuXZqZC7sdSzvEh2MeZYTIdDQ/P5TLuh2EpNGLiCWUEWMfzszDuxvN5HPKmCRJkqSpajnlL+nT0ZRcVUiSatqWEIqIQyhD655I+Y/2+pQ5lb8E/iczfz+Gvq6n9TJvAE/IzKvHH60kSZKkqS4/lEOAo2wkaRK0c4TQ+ygFka4AasmfhcBrgFdGxMsy8wdj7POrLdpvG1+IkiRJkiRJamdC6MXApZl5T31jRLyFUrzr2IjYrCoUNiqZuaSN8UlqYSiHWD67ofHuwVL/vpPuHizl5ur05ZDV7yVJkiRNusw8ATihy2F0TNsSQpl5fov2Y6rpZFsDCygjiCRNIcvvWcH8dzc0Hr+gK7HQEMfAPSuYxybdiUWSJEmS1lKd+sP7qmp7X4euJ0mSJEmSpBYmfZWxiHgNsC1wbfUYy7mHUUYW3QtcCZyW6dKNkiRJkiRJE9H2hFCVxFlIKTD9hOrnvwP7Z+YDY+zuiIbnn46It2XmcROPVJIkSZIkqTdNxgih5wPPqXt+A/DazLx0DH18H/gFcCllmcnHAAcCb6cUp16emd8bTUcRcWWLXVuPIR6p5yx95Xn0b7ZNR685eNM1LPjm4o5eU5IkSZJ6UdsTQpn5XICI2Ah4IvBB4JcR8f7M/O9R9nFwQ9OVwLsi4mrgS8AngVElhCSNT//D5jJvzrzOXvRhg529niRJkiT1qEkrKp2Zt2bmucBelJE+H4mInSfY7VeAAWDbiNhqlHEsbPYArptgLJIkSZIkSdPSpBeVzsxVEfEt4MnA3sDFE+hrKCKuA+YDmwLXtyVITQmrVk2N0SEzZ/YR0akF+CRJktRKBDOAvm7HMU7LMxnqdhCS1MqkJ4QqtW/67Zh/snG1XdmGvjSFXHzxgm6HAMCiRQPMmtXhqVKSJElqpo8yQ2A6mk+phypJU1KnhkE8s9pOaJpWRCykLGF/F3D1RIOSJEmSpKkoIjIirp/A+btXfWRE3BYR6w9z7DF1x57QsG9Ji/afVe3vHUW/Xxvv6xjORN+j6SQirq+7R5P1OLzbr3MqiIhzqvdjq27HMtnaMkIoInYDNgDOzsyhuvaZwJuA1wB3A9+q23cQcBBwWma+t659L+CezPx5wzW2B74JBHBsZt7XjtglSZIkaS33COCFwCmNO6rvbK8YR59vBH4PfDAivpOZf2rodzfKd8FB4JBx9K/VnQL0N7TtCOxAGXhxXsO+zYCbmvTzumr7XeDOhn2/m1CEmnbaNWXsccDxwGBEXAosp3xYn0ip9XMPsCQzb6w7p58y2mfThr52AT4UETcAl1NGAz0G2KmK9xzgP9oUtyRJkiRNRU8AVrWhn98DC4ADaJIQAvakTM27jPKda1Qy808R8WHg48AXgefU9kXELMrq0AG8MzOnRrHQaSwzD21sq0b07ACcl5lLRtNPRNQSQodm5vXtik/TU7sSQr8EPkaZGrY9JdlzH6Xo8ynAZxszxsP4MbA5sDOwG7AhcDsl43kScHxmPtCmuNUlM2f2sWhR96eDr1o1OGVqF6m5wbtXwMruTr/vm93HDAuNS5I0VSzgoRqlU0U/sLSdHWZmu0pkDAB/A/aMiI0z85aG/QcAQ8A3GENCqHIksD/w7IhYkpknVO3vo9ynszPzxHFHLmlStSUhlJl/Af5zjOccDhzepP1C4MJ2xKWpK2KGhZs1Kgu+ubjbITBw6ADz5vh5lSRpihjMnFrFmiMmo89I4IbM3KqhfRHwbuBJwCOBW4C/Aj8HPpqZjdOAAE4E9qBMDftiXV+PoKwEfQ4laTQmmXl/RLweuAj4VEScSRlt9F7KTI83jbXPiAjglcCBlATVHOAfwG+AL2bmz0Z5/our8x9FSXhdBZwAfKG+zEndOa8C3kyZ/bIhpSD4HyklTj5fd+ws4PXAv1JmsjyMknT7A3BiZn5zrK95OomILSkzdp5DmZZ2D3AzZQDHUZn5x7pjXwDsAzyNch/WAf5EKSXzqcy8t6HvJZSZRx8Gvk4ZffZsYBZwPvCuzFwaEetS/g0sAbagTI37dP19qvrbCvgLZQDLi4CPAC+jLHb1F8oots80fh5GeP1zgcMon69HUwbCXFq99h+Mtp+polOrjEmSJEmSJiAi9gZOp0zF+g1wAbARJYnxHuALrFkXhuqclcCrqUsIUb6sr0+ZiTEumXlxRHwWeAfwGcpsj1nA+6qBA6MWEesAJwMvp3zRPh/4Z9XnC6p+h00IAetRRjstp4zauoySpFoEfJ5SomRJwzlHAIcC9wK/ooxAeyRl9stjq/NqTgL2Be4AzqXMZnkUsBh4OKXu7VopIjanvJ9zgWuBMylJni2BN1AGdvyx7pSvUBJmfwCuoCTadgH+G3hORDyvxeyfR1M+3/8EfkoZbbYH8OSqtvAXgN2BXwB/Bp4F/G9E3JeZX27S33qUhOnW1XYWJaF1FGXK3ZJRvv5tqng2p8yG+jGllvKuwBkRcVhmHjmavqYKE0KSJEmSND0cSlkpet/M/G79jojYmZIEWUNmroyI04FXRcSWmXlDtesAygiPU4C9JhDX+4GXUkbmQEkaHD2Oft5LSQYtBV5Yn1CKiA0pRZRHcn8Vyw8z88EaTBExj5LAeF1EHJeZv6ra1wfeRknw7NBwzXUpo1tqzx9NSQbdADw5M5fX7VufMmprbfZ6SjLofzPzbfU7ImILYGbD8W+kTBu8u+64DSgJuxdSEpTNVqB7LfAJSlIxqxFcx1ESNz+jjPh6XGYuq/p8DiVR859As4TQrpSE1ONq9awiYmtK8u91EXF6Zp4+3AuvkpWnUJJB76aMcBqq9j0WOBv4RESclZl/GK6vqcSiGJIkSZI0PdTmsP+0cUdmXpyZdwxz7kmUkUWvAoiIR1FGWZyRmbdPJKjMXMnqI4/eMda6r9VUrHdVTw9sHF2Umbdl5i9HEcv9mXl6fTKoal9GSThBme5T8wjKCJLrmlzz/sw8t66p9v7/tj4ZVB17T1X+ZG023Ofvr5l5XUPb9+qTQVXbHcA7q6f196Hen4EPZmZW5yTw6WrfAsrn68Fpo9U0wt8CWw6zVPyh9cXNq1g/Uj09qMU59famLJr13cz8n/ppZlW95HdRRku9YRR9TRmOEJI0ZfStP5eBIxoar1oKfY0rbE6ewbsGWXCMhcYlSdKUdCll9bGvR8RHgEvHUP/kJ5RaN6+m1GZ5FWWAwISLPkfExpQpYzUvp0ynGounUKa/XZ6Zv25DTDsCz6NMZ5pNSYZtUO1+XO24zByIiJuAHSPiE8CXMvPPLbq9mjL17gURcRhwUmb+faKxTiOXVtuPRcQDwE8z857hToiIx1FGnz2WUg9qBuVeQN19aHBOY0KPkiSCsvLeOU3O+TNlhNamlOlc9VZk5k+anHMy8H/AooiYMcK/pedV21Nb7K993ncZpo8px4SQpCljRsxg3l0NjQ/rBws6S5IkQVm964mU0Qp7A7dExHnA9ykFjVt+Oa8KQH8TOLhKlhxAmWL2ozbEdSQwn/IFew/grRHx9cy8eAx9bF5trxv2qBFUI41OoKx+1soGDc9fR6n98x7gPRFxA6UQ8Tcz88H3JzNvj4g3UIoRHwEcERHXUGrZfD0zzx9ljCc0aR5x2tIUcAIlMfIK4Azgnoi4GDgLOC4z/1E7sJrmdSRlNFCrsuuN96FmjQLnmXln6ZJ/tBh9VqudtV6TfTc0aSMzb4uIWymJyI1pMeWyslW1PSkihqu51bm/ZLeBU8YkSZIkaRrIzBspI2meD3wOuJGSGPoycEVE9I3QRe2L7McpBZO/3WQkxphExLMoK4L9jbKq2Hso3zO/VNXg6bRDKMmg3wN7ApsAszIzgG2rY1ZLUGTmzykjWF5NWd1qBqWOzZkRcUrDsSdTVhd7A/AdSjLhjcB5EfGpUcb4uiaPHcfwGrsiMx/IzP0oq7d9GLgYeCqlSPQ11Qp4NftR7sVNlLpLj+Kh+1BL2rRKFA03UmfUK4K1WS13chbw1WEe3+tKdOPkCCFJkiRJmiYy835KAduz4cFlwI+jLM/9HkrB21bn/iYirqWM4oEJTherCinXage9rRpBcywlmbKYMo1stKsu3Vhtt55ITJSC0gD7Z+aVDfse0+qkqo7SN6oHEbErJeGzT0TslZln1h27DDgWOLYaCfN8ylLqh1QFqxuv23itVomQaSEzf0up2XN4RDwCOJwyEuhoHpoyVbsPb87MHzZ00fI+TJItmjVWsW8E3A3cOkIfN1XbYxsLuk9njhCSJEmSpGmqWjHsk9XT7UZxynGUqTG/y8wLJnj5D1LqwJyemadV8STw75Rl4z9cJaxG41LKl/IdImIidVg2rrY3Ndn3itF2kpkXUUYLwTDvaxZnAbWkx8LRXmNtUCXS3gskq79PbbkPbdJXrUTWqLYq3oWjKIJeq0H00mGPmmZMCEmSJEmaTvojmDeVHnSobkhEvDMiHtlkV23J+Bub7FtNZn4iM/szc0JLpEfE9sBhwO00rNKUmVdRklSzgWNG019m3stDK0l9pTGRFBEbRsQzR9HVNdX2TQ3n70sZuURD+xYRsSQiZje0rw88q3p6Y9X2pIh4WVWnqP7YuZSpUw8euzaKiNdERLPk2J6U6V/1r712H/69GkVV6+PplM9Npx1ZP6UyIh5NSWgCfH4U538XWAq8OiI+EBGr1SqKYreI2K1tEXeAU8YkSZIkTSdLux1AF32I8sX2cuBaypfwHYBtgBWMfnrWhETEDErdonWB92XmGkWAKXVl9gP2iohXZOa3R9H1xygrRb2EUpPmXMrKaJtT6tb8hFLseThHUKbEfSIiXk5JTDyOUnvpSODQhuPnAscDn4+ISygjWuYAiyjLrF/CQytLbUlJDNxWHfsPypSjZ1AKJJ+xli89vw/wtYi4jlKj6W7g0ZRk2BDw/rpjPwssAd4C7B4RV1DqCC0GPsWa92EyXQTMAv4UET8HZgLPoSQsT8zMViuHPagqyv4S4MfAfwEHVa9pgJIQ3pFSWP2dwKiKi08FjhCSJEmSpOnhbZTVsGZTRmXsAdwPHAVsn5nXdjCOXShftP+v2QHViJ/aKJ3PRMRGI3Va1Ufah5JIuIiSxHkZsBnwA0qNmpH6+BUl6fBzSq2aF1Kmr+1D85Eg1wHvoixlvkV1vcWUlaneCTyzei1UMb2fMr1tW+DlVYxXUApr7zNSfNPcUZT38A7g6ZTpU/Mp9ZOempnfqR2YmddQ3pszKAmTFwEPB96YmZ0eIXQvpcbWN4BdKTWfbqQkpZaMtpPq39eTKJ+Bm6q+XkZJyP4WeCsTrMvVaVGmePaeiLhywYIFC668cth6X2ul++5bxgUXzF+tbdGiAWbN6r2lvX0vimV/vYr5xy9YrW3gX5cyb4sndDiQZTB/9fvBwADM69z9WLZyGfOPXD2GgUMHmDentz4TkqTpaeHChSxdunRpZq4VdUyq6VgD3Y5jnOZnsqzbQUi9KiK2Av4C/DIzd+9uNFOTU8akbhsaguXLuxvD8hXdvb4kSZIkqaNMCEndtnz5mqNiOm02wyxQKkmS1DXLKVNSpqMu/8VPkoZnQkiSJEnSlJTJEDjtSpImgwkhSZIkSZK0VsnM6ykr8akFVxmTJEmSJEnqMY4QkqaipUuhv79z17t7EBpWGWPjuZ27viRJkiSpo0wISVNRf39Hl1pnZZO2GQ4glCRJkqS1ld/4JEmSJEmSeowJIUmSJEmSpB5jQkiSJEmSJKnHmBCSJEmSJEnqMSaEJEmSJEmSeoyrjEmSJEmaks45J2YAfd2OY5yW7757DnU7CElqxYSQJEmSpKmqDxjodhDjNB9Y1u0gJKkVp4xJkiRJ0hQTERkR10/g/N2rPjIibouI9Yc59pi6Y09o2LekRfvPqvb3jqLfr433dQxnou/RdBQR19fdq9rjgYgYiIgfRcQLWpx3QnXskhb7XxYR90XE/RHxmib7Z0fEwRFxdkTcHBH3RsQdEbG06vtFEbFOm1/utBYRW1Xv+TndjqUVRwhJkiRJ0trtEcALgVMad0TETOAV4+jzjcDvgQ9GxHcy808N/e4GvAkYBA4ZR/8a3neBO6ufZwNPAPYA9oiI92bmJ0bbUUTsC5wMBPCazDy5Yf9uwHeATYF7gIuBvwPrAVsDr6seS4GFE3hN6jBHCEmSJEnS1PME4Dlt6Of3wAPAAS3270mZmnfZWDqtEkAfBtYHvli/LyJmAV+iJBjemZmDY4xZIzs0M5dUj1dk5hOBt1b7/isiNhlNJxHxCkoyCOBVTZJBOwE/oySD/gfYJDOfkZmvzMyXZub2wBbA0cBjJv6y1EmOEJIkSZI0nSygjDqZSvopoyPaJjOvblNXA8DfgD0jYuPMvKVh/wHAEPANYKcx9n0ksD/w7IhYkpknVO3vo9ynszPzxHFHrjHJzGMi4j+AzYGnAacPd3xEvBI4kXL/98/M7zbsn1HtXw/4QGZ+tMV1bwTeGRHe62nGhJAkSZKk6WRw991zShVrPuecaHufEZHADZm5VUP7IuDdwJOARwK3AH8Ffg58NDPvZE0nUqYTvYK60TwR8Qhgb+AcStJoTDLz/oh4PXAR8KmIOJMy2ui9wF2UKWNjEhEBvBI4kJKgmgP8A/gN8MXM/Nkoz39xdf6jKAmPq4ATgC9krr76W3XOq4A3A48DNqQUBP8jcFpmfr7u2FnA64F/pYyIeRgl6fYH4MTM/OZYX3ObDVASQsN+14+IVwFfo7w3r8jM05scthdlpNpfgY+PdOHMvHQsgUbEXsA7KcnDfmA58GfgJ5n54brjNgJeQ5n2+HjK5/5OytS1T2XmT5r0fQ7wTODRwFOBQynT2W4BTgLen5n3RcTWwEcpo/EeDvwaeHtmXtHQ3+HAhyj3/Srgv6p+Z1TnfCAzLxrj66/FtZiHCuifBfxXZv51LH2Nl1PGJEmSJGkaiIi9gXOBFwE3A6cCvwXmAu+hfKlu5nRgJfDqhvZ9KFO+ThpvTJl5MfDZKobPAF8GZgEfzMy/jKWvqijxtyijlZ4BXA6cBtwEvAB42yi6Wa86/7mURNIZlITVQuDzwHFNzjmCkjR7SnXNU4Frge2BwxqOPanqZ9uq3+9REiaLGUcCrJ2qBN821dOrhjnuAEoy6AFgnxbJICjTCQG+k5kPtCvOKoa3Aj8EngX8iVIT6Q/AlsDhDYfvSvmMbUOVpKu2zwN+HBEHDnOpt1Pu7a2UZMssyj39ckQ8jnIPd6QkVP8E7A78Ypgpd4uAXwGbAT+q4ngu8MuIeN4oXjoAEfEW4ALgZcANlH+jy4F/Ay6JiCeMtq+JcISQJEmSJE0Ph1L+qL9vk+k9O1O+UK4hM1dGxOnAqyJiy8y8odp1AKVI8CmU0SDj9X7gpZSROVDqER09jn7eC7ycMv3uhfUJpYjYkPLFfST3V7H8MDNX1Z0/DzgTeF1EHJeZv6ra16ckmu4Admi45rqUqVe1548G9qV8gX9yZi6v27c+ZdRWx0XEwygjeT4ObAB8PzOvbHH4fpREyirgZZl55jBd71Btf9uuWOu8G0hg18y8pNZYjdZ6ZsOxfwSe1jgCJyKeREnkfDoivt1idNzrq3Mvqc55JPA7yoijnYFjgfdlZlbX/mq17y2UEUGN3gD8N2VEUFZ9vhk4BjghIrbOzLuHe+ERUUtw3Qy8uH5kVUT8WxXT8ZRE2KRyhJAkSZIkTQ/zqu1PG3dk5sWZeccw555EKfL8KoCIeBRlNMQZmXn7RILKzJWsXlj6HWMdUVJNxXpX9fTAxtFFmXlbZv5yFLHcn5mn1yeDqvZllIQTlOlkNY+gjCq6rsk178/Mc+uaau//b+uTQdWx92TmhSPF10Z/qS07T5medylltM3hDL9q3B6UPMBnR0gGQZnGBC1qdkXEV6ol5+sfi0cZ/zzg1vpkEEAW5zS0/aXZdKzM/C1ltNYjKK+9maPrr5GZ/6CMIAvKff9gLbFTbY+sDm1MStXcABxeO6c67/8o08Y2pYy6G8l/AOsAb2qcZpeZXwG+Dzy1SnhNKkcISZIkSdL0cCllJMjXI+IjwKWN9XCG8RNKjZJXU0aSvIqSGJhwIeCI2Bh4R13TyylT28biKcBGwOWZ+es2xLQjZSTMlpRl2YMyegZKnSAAMnMgIm4CdoyITwBfysw/t+j2asrUuxdExGHASZn594nGOk71y86vS5nC9DTgEOBGmk+NAzgf2I1SBPq8zPz+BGJ4HSWxUe8c4LxRnHspsDgivgIcNcyIJuDB6YTPoUzZ2pSSzIGH7uXjmp0HnN2krXZ/z2lMHNbt27RFf9/NzPubtJ9MqSn0dIb5N1UV6n4OJYn34xaH1aaF7sLkjM56kAkhqcGqVR1etOL+wVK2rs7MHKL9pQklSZI0zb0PeCKlEPTewC0RcR5lRMGJmXlPqxOrAtDfBA6ukiUHUKaY/agNcR0JzKd8Kd4DeGtEfL2qLzRam1fb6yYSSDXS6ATK6metbNDw/HXANyl1mN4TETcAvwS+mZkPvj+ZeXtEvAH4EqXu0BERcQ3wC+DrmXn+KGM8oUnz6cPU8mnm0My8vqHfrSj1bb4SEddn5s+bnHcspZbOR4BvR8QLhinUXRsF1bQ2VWY+mE+IiC8AbxxD/G+l1M05EDgwIv5Jec9PBU6pH2EWEZsBP+ChKWzNNN7TmmbF0u9stS8z7ywzxx5MODW6oUX79dX2/7XYX9NPKV4NcF91reGOnVQmhKQGF1+8oPMXPX31p4seWMEsWtUxkyRJUi/KzBsj4inAsykrLj2Th5JD746IpzVOZWpwEnAwZYTQ9sD/NRkhMSYR8SzKl/q/UYoq70dJmHwpInZuMZpiMh1CSQb9nlKn5jLglsxcFRG1osSrfQvPzJ9HxGMp7+kelKl0rwVeGxHfzcx96449OSJ+Spl29jzKPXgj8MaIOCoz38XIXtek7XpGWCZ+JJl5fUR8Evhfyn1ulhAiMz8aERtQ3p/vRcS/tJjudjllNNGTmEDh8RYxXBERCyjv916U9/wV1ePCiNg9M++rDj+Wkgz6LiUR90fgjswcioh/p0xXbJVZGW4E3WhH17VTrWzPnZTXM5xhR021Q9tqCEXEIRFxakRcGxG3RcS9EXFDRHwtIp44jv42jojPVH3U+jq6WnJOkiRJknpOVdfm7Mw8ODN3ALaifPF/HGWEy3Dn/oayetYeVdOEpotVhZRrtYPeVtUiOpYyZWhHVp9GNpIbq+3WE4mJUlAaYP/MPCszB+qSXo9pdVJm3p6Z38jM12bmFpTpVzcB+1TLo9cfuywzj83MV1CWQN8TuB04JCIWjhRgZkaTx+Fjf6lN1eogtZpCVYvhPcD/AXOAM6tRY41qo6NeXk3Zaquq7tLpmfnvmbkNsB1wBeW9fz1ARMwB/gX4J7BfZv6mqidVS+a0vKeTZMsR2keaQjhIKeQ+BPxrZi4Z5nF6m2JuqZ0jhN5H+TBdQcnGQlna7zXAKyPiZZn5g9F0FBH9wIXAYylz+E6v+no7sGeV+V7RxtglSZIkTQ/955wz5SbXT/rUjlYy84ZqVMizKV+oR3IcZbWyGzPzggle/oOUxMPpmXlaFU9WozZ+B3w4Ir6TD61qNpxLKUuD7xARu1TJq/HYuNre1GTfcMWWV5OZF0XE1ymFqLejrFDW7LgEzoqIH1JGJi2kAyM7hlFLkDRbcavRWynf4V8LnB0Rz8jMq+v2n0lZvv4JlPfho+0MtFFmXhkRn6ckGWuf5Q0pA1lubixUHhEzeSgB2Ckvi4h3NymaXlthb9j6SdXUzXMoSdnn0KRAfCe1c5WxFwMbZ+ZTM/Nl1WNbyodsJnBstWzfaBxNSQadCmybmftl5nbA54BtgKPaGLckSZKk6WMppTjyVHosndRXXImId1bLZjeqjWC5scm+1WTmJzKzPzMntIJRRGwPHEYZGXNQwzWuAj5JKeZ8zGj6y8x7gU9XT78SEauNxIiIDSOi1cpP9a6ptm9qOH9fSuKDhvYtImJJRMxuaF+fh1auurFqe1JEvKyqU1R/7FxKQeEHj+2GqobQu6unI60gVktmHUj53j0P+GnVR23/EGWAx73ARyLiiIjYsLGfiOgDth1DnLMj4uDG2T9VweXa6LXa+zgA3AZsFxG71R27DuUzts1or9smW9GwHH2VAH0aZRTTSNPAoCxbPwQcHxG7N+6MiIdHxIER8bCJBjuSto0QalVAKzOPiYhDKEP/FlBGELUUEZtSMqv3AW9pmHN6GCXzdkCVlRtoS/DqWTNn9rFoUXc/RqsGruHiP412dUZJkiT1sA8BR0bE5ZSpX0GprbINsIKHlsyeVNUX9y9Tvk++LzObFe79b0o9ob0i4hWZ+e1RdP0xSr2alwDXRMS5lITA5sBOlJXSRlp6/ghKUuETEfFySoLocZRVzI6kjI6qNxc4Hvh8RFxCGVk0h7Ka1TzgEkrCBMq0oO8Ct1XH/oOyMtozKEWNz2hRi2cyHBkRzVYZmwX8BvjUaDrJzAciYn/ge5T37WcR8fTa6mmZeWlEPBf4DuX7+MER8WvK1Kj1q+vuQBkEcjXl/RrJLOAz1Wu4lFI/aRawM+VeX0+pQ1UbUXME5fP0y4j4OeWz/lRgE8qy828dzWttky8D/xERL6PkNh5bxb0KWJKZd43UQWaeFxFvpdR6+kVE/IHyOV1FSTjtSClqfSpw9yS8hgd1qqh0bc7mfcMeVexBGbl0bmb+s35HZt4bEWdQsph7UarHS+MWMYNZs+Z1N4h1OryqmSRJkqart1G+Lz2ZUrcGykiKoyhLdzdLzExWHLsAF1Hq0Kyh+u72Jkp9o89ExNmZeetwnVZf/vehjEo5kJLEWR+4mbLKVNNrNfTxq4hYTEkgPImSLPs9sA+lwHRjQug64F2U6TsLqte1klKL52OUZejvrY69CHg/ZXretpQlxm+hJAa+wgRrMo3RPnU/J3AH5fV9G/h8XUHmEWXmfVWC40eUItk/iYhnZuZgtf+8iNiaUtfnRZTpXE+jjBz6W3XNU4AfjLKI+J2UJM5zKMmk7Sm5gr9SalD9b32JmMz8WETcRKlJtRslSXIeZcriTqN9nW1yASWB+BFKEfIAfgZ8cCxTMDPzCxFxEeU17V71dRfl/TyJkgy6rZ2BNzPpCaGIeA3lH8u11WMktaXkLmux/zLKfxy2n3h0kgCGhoCVq099H1wW5T9JnTQYNE7B7xtq79xWSZKk6SAz1yiUlJlfB74+yvPPofXKS82O/yZl6fXG9hNo+EN8Zn6GMsJjpD5/MZYYqnOGgK9Wj5GObdp3Zl5ESTY007jC2B1UCbVRXO8flETTf4907GTJzK3Ged4SYMkw+++mJCZa7b8L+Gz1mJAqaXQMo5xOWJ3zNeBrTXZdQZOBIpm5+zB9ndDsnLr9w35mq1Fgzx3umOq46xnm85+Zv2OYe9IJbU8IRcRhlEJacyjFpxZShpPt36TwUjNbVNtmRcDq21tV95Y0RiuWB/zPstXaFvxPNyLpB1aPY2DFIPM26UYskiRpClgOzO92EOM03PLvktR1kzFC6Pmsno29AXhtZl46yvMfXm1bjU1YWW03GE1nEdGqwvtElzOU1AkrVsCy7Nz17m4yhW9oaM02SZI06XbfPYdo/GuRJKkt2p4QysznAlQVw59Imdf3y4h4f2Z2bWidpGlq8W5AB+sszeahtRlqblkBGzhMSZIkSdLaY9JqCFUFw86NiL2ACynL1J2dmRePcGqtUvrsFvvnVNs7RhnHwmbt1cihBaPpQ5IkSZIk9abMPBw4vMthtN2kF5XOzFUR8S1KJfy9gZESQn+ttpu12F9rv6EN4Ulq4bxLVrDNFnM7es3Ba1awYHFnrylJkiRJvahTy87X5nuMZn3vy6ttq+Xjau1XTCgiScOa2zfEvNH8i22nQWv1SJIkSVIndCoh9Mxqe90ojj0LGAKeHhHzM3OgtiMi1qOMMnoAOLPtUUrqrrlNRgctvQr6O1hU+qZr4PuLO3c9SZIkSeqCtiSEImI3yqpfZ2fmUF37TOBNwGuAu4Fv1e07CDgIOC0z31trz8ybI+Jk4NXAMRHxysy8v9p9BGWU0VfrE0WS1hIzZqzZ1t8/urGF7dJslTFJkiRJWsu0a4TQ44DjgcGIuBRYDvRTVhnbFLgHWJKZN9ad0w9sW+1v9A5gV2Af4OqIuARYCGwHXAsc0qa4JUmSJEmSek6TP8ePyy+BjwF/BLYHXg7sBqwAPgc8MTO/PdrOMnMQ2KU6dxbwUmBD4LPALpm5ok1xS5IkSZIk9Zy2jBDKzL8A/znGcw5nmGXbqqTPwdVDkiRJkiRJbdKpotKSWhhqsrDW8hXBuh3817liebsGC0qSJEmSpgMTQlKX3XJrrNH2hN36uO22TkbRZHUvSZIkSdJay2EBkiRJkiRJPcaEkCRJkiRJUo8xISRJkiRJktRjrCEkTUHn/+AW5m87r2PXG1w5yIJjnrBa29y+pR27viRJkiSps0wISVPQ3I2HmNe5fBDMTpgzuFrTDMcPSpIkSdJay698kiRJkiRJPcaEkCRJkiRJUo8xISRJkiRJktRjTAhJkiRJkiT1GItKS102lENrtK2451bWXbmsYzEM3jU48kGSJEmSpLWGCSGpy2699/Y12nY77QXcNnNFF6KRJEmSJPUCp4xJkiRJkiT1GBNCkiRJkiRJPcaEkCRJkiRJUo+xhpA0Bf1gr5PYdocndzWGvtl9Xb2+JEmSJGnymBCSpqCN19uIeXPmdTsMSZIkSdJayiljkiRJkiRJPcaEkCRJkiRJUo8xISRJkiRJktRjTAhJkiRJkiT1GBNCkiRJkiRJPcaEkCRJkiRJUo8xISRJkiRJktRjTAhJkiRJkiT1GBNCkiRJkiRJPcaEkCRJkiRJUo8xISRJkiRJktRjTAhJkiRJkiT1GBNCkiRJkiRJPcaEkCRJkiRJUo8xISRJkiRJktRjTAhJkiRJkiT1mHW7HYDUTUNDsHx5d2O45dYZsE53Y5AkSZIk9RYTQuppy5fD/PndjWHDDR/N6ad3NwZJkiRJUm9xypgkSZIkSVKPmXBCKCJmR8RLIuIrEfHHiLgnIlZGxOUR8cGIePgY+7s+InKYx+MnGrMkSZIkSVIva8eUsVcBX65+vgr4PvAIYBHwYWD/iHhmZg6Msd+vtmi/bVxRStPIUN7Cffct62oMM2f2EeEgQkmSJElaG7UjIbQK+BJwdGZeVWuMiE2BHwJPAo6mJI5GLTOXtCE2acyWLoX+/s5db+Cv17HsjtXblt+5Fxdc0LkYmlm0aIBZs+Z1NwhJkiRJ0qSYcEIoM79Kk9E8mXlzRLwVuAB4WUTMysz7Jno9abL198O8DuZB7r/zgTUSQpIkSZIkTabJng9yebVdD+ib5GtJkiRJkiRpFCZ72fnHVNtVwIqxnBgRhwFbA/cCVwKnZWZ3i6pIkiRJkiStBSY7IfT2antWZt47xnOPaHj+6Yh4W2YeN5ZOIuLKFru2HmM80qRYZ8ZGvKShXtBV+59H36O26VgMq1YNcvHFCzp2PUmSJElSd01aQigi9gL+jTI66ANjOPX7wC+AS4FllFFGB1KSS8dGxPLM/F6bw5W6JmIGt61avW3ddeZa0FmSJEmSNGkmJSEUEY8HTgQCOCwzLx/hlAdl5sENTVcC74qIqymrmX0SGHVCKDMXtojxSsAhEZIkSZIkqee0vah0RDwKOAvYGDgqMz/Tpq6/AgwA20bEVm3qU5IkSZIkqee0NSEUEXOBs4EtgeOBQ9vVd2YOAddVTzdtV7+SJEmSJEm9pm0JoYh4OPAjyjSsU4E3ZGa2q//KxtV2ZZv7lSRJkiRJ6hltSQhFxHqUuj67AD8G9s/MB9rRd901FgLbAncBV7ezb0mSJEmSpF4y4YRQRKwDnAw8GzgXeFlm3jfCOQdFxNUR8fGG9r0i4tlNjt8e+A6lSPWxI/UvSZIkSZKk1tqxythBwEurnweBYyKi2XGHZuZg9XM/ZbRPYy2gXYAPRcQNwOWU0UCPAXaqYj0H+I82xCxJkiRJktSz2pEQ2rju55e2PAoOpySMhvNjYHNgZ2A3YEPgduA84CTg+HZPRZMkSZIkSeo1E04IZebhlGTPhM/JzAuBCycakyRJkiRJklpr67LzkiRJkiRJmvraMWVMkibN4EgTTdt9veXrwMr+1dqGhjobgyRJkiRNNhNCkqa0BQs6fcVtgGWrtazY/xo22arTcUiSJEnS5HHKmCRJkiRJUo8xISRJkiRJktRjTAhJkiRJkiT1GGsISZoy+vpgYKC7MVxzxXUsfu7Wq7WtuOcWlq1c1uKMzuib3ceMMIcvSZIkqT1MCEmaMmbMgHnzuhvD4NwH1mhbfNoL4ewOL3fWYODQAebN6fKbI0mSJGmt4Z+bJUmSJEmSeowjhNQVQ0OwfHm3o4DB7g76kCRJkiSpK0wIqSuWL4f587sdhSRJkiRJvcmEkCTVmbveRmu0Ld3/fPq32bhjMQzeNciCYxZ07HqSJEmSeo8JIUmq02wlr/715zJvTn8XopEkSZKkyWFRaUmSJEmSpB7jCCFNGUuXQv8UGITR19ftCCRJkiRJmlwmhDRl9PfDvHndjkKSJEmSpLWfU8YkSZIkSZJ6jAkhSZIkSZKkHmNCSJIkSZIkqceYEJIkSZIkSeoxJoQkSZIkSZJ6jAkhSZIkSZKkHmNCSJIkSZIkqceYEJIkSZIkSeoxJoQkSZIkSZJ6jAkhSZIkSZKkHmNCSJIkSZIkqceYEJIkSZIkSeox63Y7gF6SOcSqVcu7HQarVg12OwRJkiRJktRFJoQ6aNWq5VxwwfxuhyFJkiRJknqcU8YkSZIkSZJ6jAkhSZIkSZKkHmNCSJIkSZIkqcdYQ6jLdt55KTNn9nc7DGbO7Ot2CJIkSZIkqUNMCHXZzJn9zJo1r9thSJIkSZKkHuKUMUmSJEmSpB5jQkiSJEmSJKnHTDghFBGzI+IlEfGViPhjRNwTESsj4vKI+GBEPHwcfW4cEZ+JiBsi4t5qe3REbDTReCVJkiRJknpdO0YIvQo4DTgQeAD4PnAu8Gjgw8DFETF/tJ1FRD/wG+Bg4H7gdOAO4O3AryNibhtiliRJkiRJ6lntSAitAr4ELMjMBZn5iszcA9gW+C3weODoMfR3NPBY4FRg28zcLzO3Az4HbAMc1YaYJUmSJEmSetaEE0KZ+dXMfGNmXtXQfjPw1urpyyJi1kh9RcSmwP7AfcBbMvP+ut2HAcuAA8Yy4kiSJEmSJEmrm+yi0pdX2/WAvlEcvwclpnMz85/1OzLzXuAMYB1gr3YGKUmSJEmS1EsmOyH0mGq7ClgxiuN3qLaXtdhfa99+IkFJkiRJkiT1snUnuf+3V9uzqhE+I9mi2t7UYn+tfcvRBhARV7bYtfVo+5AkSZIkSVqbTNoIoYjYC/g3yuigD4zytNoS9Xe12L+y2m4wgdAkSZIkSZJ62qSMEIqIxwMnAgEclpmXj3DKpMnMhc3aq5FDCzocjiRJkiRJUte1fYRQRDwKOAvYGDgqMz8zhtPvrLazW+yfU23vGGd4kiRJkiRJPa+tCaGImAucTanxczxw6Bi7+Gu13azF/lr7DWOPTpIkSZIkSdDGhFBEPBz4EWUa1qnAGzIzx9hNbWrZTi3219qvGHuEkiRJkiRJgjYlhCJiPeB7wC7Aj4H9M/OBcXR1FjAEPD0i5je5xt7AA8CZE4tYkiRJkiSpd004IRQR6wAnA88GzgVelpn3jXDOQRFxdUR8vL49M2+u+poFHBMR9UWvjwDmASdm5sBE45YkSZIkSepV7Vhl7CDgpdXPg5RETrPjDs3MwernfmBbYNMmx70D2BXYB7g6Ii4BFgLbAdcCh7QhZkmSJEmSpJ7VjoTQxnU/v7TlUXA4JWE0rMwcjIhdquNfUvX5T+CzwIcy89ZxxilJkiRJkiTakBDKzMMpyZu2nZOZK4CDq4ckSZIkSZLaqK3LzkuSJEmSJGnqMyEkSZIkSZLUY0wISZIkSZIk9RgTQpIkSZIkST3GhJAkSZIkSVKPacey89L0NTQEy5d3N4blK7p7fY1sxQpYlp273t2Da7YNDXXu+pIkSZLWeiaEetBUyIEMNvm+2xXLl8P8+d2NYTbw7u6GoBEs3g3o4Ie22WfilhWwwSadi0GSJEnSWs2EUA+aCjkQSZIkSZLUPSaE1NOGcojls7sbw2CXry9JkiRJ6j0mhNTTlt+zgvlO15IkSZIk9RgTQtJUtPHcbkfQu+Y2ee+XXgX9HSwqfdM18P3FnbueJEmSpJ5jQkgALF0K/f3djaGvr7vXn1JmzOh2BL2r2Xvf3w/zOhhDs1XGJEmSJKmNTAgJqL7vdvILryRJkiRJ6hoTQlKDpa88j/7NtulqDH2zHS4lSZIkSZo8JoSkBv0Pm8u8OQ6XkiRJkiStvSxUIkmSJEmS1GNMCEmSJEmSJPUYE0KSJEmSJEk9xoSQJEmSJElSj7GotKSmVq0a7HYIAMyc2UeEuWtJkiRJaicTQpKauvjiBd0OAYBFiwaYNctV3yRJkiSpnfyzuyRJkiRJUo8xISRJkiRJktRjTAhJkiRJkiT1GGsISWLmzD4WLRrodhisWjU4ZWoXSZIkSdLazISQJCJmWLhZkiRJknqIU8YkSZIkSZJ6jAkhSZIkSZKkHmNCSJIkSZIkqceYEJIkSZIkSeoxJoQkSZIkSZJ6jAkhSZIkSZKkHmNCSJIkSZIkqceYEJIkSZIkSeoxJoQkSZIkSZJ6jAkhSZIkSZKkHmNCSJIkSZIkqce0JSEUEU+OiP+IiFMj4qaIyIjIcfZ1fe38Fo/HtyNmSZIkSZKkXrVum/r5APDiNvVV89UW7be1+TqSJEmSJEk9pV0JoQuBK4CLq8f1wHoT6TAzl0w4KkmSJEmSJK2hLQmhzPxk/fOIaEe3kiRJkiRJmgTtGiEkSWutwcEOX2/5OrCyf7W2oaHOxiBJkiRp7TZlE0IRcRiwNXAvcCVwWmYu625UknrRggWdvuI2wOr/uVux/zVsslWn45AkSZK0tpqyCSHgiIbnn46It2XmcWPpJCKubLFr6/GFJUmSJEmSNL1NxYTQ94FfAJdS/kT+GOBA4O3AsRGxPDO/18X4JKnjVtxzC8tWdn+QZN/sPmbEjG6HIUmSJGmCplxCKDMPbmi6EnhXRFwNfAn4JDDqhFBmLmzWXo0c6vhEEEkaj8WnvRDO7nAxoyYGDh1g3px53Q5DkiRJ0gRNuYTQML4CfBTYNiK2yszruxyPpLVQXx8MDHQ3hmuuuI7Fz3VWqyRJkqTJM20SQpk5FBHXAfOBTYHruxuRpLXRjBkwr8sDYAbnPtDdACRJkiSt9aZbIYiNq+3KrkYhSZIkSZI0jU2bEUIRsRDYFrgLuLrL4UjSpJm73kZrtC3d/3z6t9l4zYMn0eBdgyw4xlJrkiRJ0tqoKwmhiDgIOAg4LTPfW9e+F3BPZv684fjtgW8CARybmfd1Ml5J6qRmq3j1rz+XeXP6uxCNJEmSpLVRWxJCEfEC4AN1TbOq9ovq2j6SmT+sfu6njPbZtKGrXYAPRcQNwOWU0UCPAXaqYj0H+I92xCxJkiRJktSr2jVCaB7w1CbtT204ZiQ/BjYHdgZ2AzYEbgfOA04Cjs9Mq61KkiRJkiRNQFsSQpl5AnDCGI4/HDi8SfuFwIXtiEmSJEmSJEnNTbdVxiRJkiRJkjRBJoQkSZIkSZJ6jAkhSZIkSZKkHmNCSJIkSZIkqceYEJIkSZIkSeoxJoQkSZIkSZJ6jAkhSZIkSZKkHrNutwPoKUNDa7YtH+z8XRgMoH/1tqEhzA9KkiRJktQbTAh10i0r1mx7wgK4rdOB9APLVm9asQI26W96tCRJkiRJWrs4JESSJEmSJKnHmBCSJEmSJEnqMSaEJEmSJEmSeow1hLrt/PNg/jadveY1t8Dizl5yDUNDsHx5l4MAljep6yRJkiRJ0lrOhFC3zZ0L8+Z19pqD0dnrNbN8Ocyf3+0oYDbw7m4HIUmSJElSZzllTJIkSZIkqceYEJIkSZIkSeoxJoQkSZIkSZJ6jDWENHUsXQr9/Z295t2DcPyC1ds2ntvZGCRJkiRJ6jATQpo6+vs7X2B7ZZO2GQ6ckyRJkiSt3fzmK0mSJEmS1GNMCEmSJEmSJPUYE0KSJEmSJEk9xoSQJEmSJElSjzEhJEmSJEmS1GNMCEmSJEmSJPUYE0KSJEmSJEk9xoSQJEmSJElSjzEhJEmSJEmS1GNMCEmSJEmSJPUYE0KSJEmSJEk9Zt1uB6ApYsUKWJadu97gYOeuJUmSJEmSVmNCSMXi3QCTNJp6Vq3q/udy5sw+IhxQKUmSJGntYUJI0pR28cULuh0CixYNMGvWvG6HIUmSJElt45+8JUmSJEmSeowJIUmSJEmSpB7jlLFeNHfumm1Lr4L+DhaVbqavr7vXlyRJkiSpR5gQ6kUzmgwM6+8HS6Soy2bO7GPRooGuxrBq1eCUqFskSZIkSZOpLQmhiHgy8C/ALtXjUQCZGePsb2PgcOAlwCOBfwCnAYdn5q0TDljSlBQxw+LNkiRJktQB7Roh9AHgxe3oKCL6gQuBxwJ/Bk4HFgJvB/aMiKdl5op2XEuSJEmSJKkXtauo9IXAR4AXAZsC906gr6MpyaBTgW0zc7/M3A74HLANcNTEQpUkSZIkSeptbRkhlJmfrH8eMa6ZYkTEpsD+wH3AWzLz/rrdhwGvBA6IiHdnZncLjUiSJEmSJE1TU23Z+T0oMZ2bmf+s35GZ9wJnAOsAe3UhNkmSJEmSpLXCVFtlbIdqe1mL/ZcBBwLbdyYcTZahHGL5Xcu7HQaDdw12OwRJkiRJkjpuqiWEtqi2N7XYX2vfcrQdRsSVLXZtPdo+1H7L71rO/CPndzsMSZIkSZJ60lSbMvbwantXi/0rq+0GHYhFkiRJkiRprTTVRgi1XWYubNZejRxa0OFwJEmSJEmSum6qjRC6s9rObrF/TrW9owOxSJIkSZIkrZWm2gihv1bbzVrsr7Xf0IFY1GFL37KU/tn93Q6Dvtl93Q5BkiRJkqRJNdUSQpdX251a7K+1X9GBWNRh/bP7mTdnXrfDkCRJkiRprTfVpoydBQwBT4+I1Zagioj1gL2BB4AzuxCbJEmSJEnSWqErCaGIOCgiro6Ij9e3Z+bNwMnALOCYiKgfwXQEMA84MTMHOhetJEmSJEnS2qUtU8Yi4gXAB+qaZlXtF9W1fSQzf1j93A9sC2zapLt3ALsC+wBXR8QlwEJgO+Ba4JB2xCxJkiRJktSr2lVDaB7w1CbtT204ZkSZORgRuwCHAy8BXgr8E/gs8KHMvHUigUqSJEmSJPW6tiSEMvME4IQxHH84JeHTav8K4ODqIUmSJEmSpDaaaquMSZKaGFwxA5Z1+JorA1b2r9Y2NNTZGCRJkiRNDhNCkjQNLFg8twtX7acxC7XiLYNsskEXQpEkSZLUVlNt2XlJkiRJkiRNMhNCkiRJkiRJPcaEkCRJkiRJUo+xhpAkTTF9c5MB5q3eeN75MLezdYSu+fttLH7u1qs3plWlJUmSpLWBCSFJmmJmzIB5DK7euHjbjscx+LA1i0pz663A/I7HIkmSJKm9nDImSZIkSZLUY0wISZIkSZIk9RgTQpIkSZIkST3GGkKSNNX09cHAQLejgN//GZ6zetOKe25h2cplzY/vkL7ZfcwI/54hSZIkTYQJIUmaambMgHnzRj5usm18yxpNi097IZw92OTgzhk4dIB5c6bA+yNJkiRNY/6JVZIkSZIkqceYEJIkSZIkSeoxJoQkSZIkSZJ6jDWEJElNzV1vozXalu5/Pv3bbNyxGAbvGmTBMQs6dj1JkiSpV5gQkiQ11Wwlr/715zJvTn8XopEkSZLUTk4ZkyRJkiRJ6jGOEOqgoaE125avCNbt8F0Y7O6K0ZIkSZIkqctMCHXQLbfGGm1P2K2P227rQjCSJEmSJKlnmRASAIMrB2F2du56dzlMSZIkSZKkbjEhJAAWHPMEmGOSRpIkSZKkXmBRaUmSJEmSpB7jCKEuO/8HtzB/23kdvebgysEyIqjew5Z3NAZJkiRJktQ9JoS6bO7GQ8zrbD6o1ApyepgkSZIkST3LhJAAWPqWpfTP7u9qDH2z+7p6fUmSJEmSeoUJIQHQP7ufeXM6PVRJkiRJkiR1g0WlJUmSJEmSeowJIUmSJEmSpB5jQkiSJEmSJKnHmBCSJEmSJEnqMSaEJEmSJEmSeowJIUmSJEmSpB5jQkiSJEmSJKnHmBCSJEmSJEnqMSaEJEmSJEmSeowJIUmSJEmSpB7TtoRQRDwsIv4rIq6JiHsi4u8RcVxEPGqM/VwfETnM4/HtilmSJEmSJKkXrduOTiJifeDnwK7AzcD3gK2AfwVeGBG7Zuafx9jtV1u03zbeOCVJkiRJktSmhBDwfkoy6ELgeZl5J0BEHAJ8CjgO2H0sHWbmkjbFJkmSJEmSpDoTTghFxCzgoOrpW2vJIIDMPCoiXgc8MyKenJmXTvR6kqTeNnjXYLdDAKBvdh8zwlJ8kiRJmp7aMUJoN2BD4LrM/G2T/acA2wN7AyaEJEkTsuCYBd0OAYCBQweYN2det8OQJEmSxqUdCaEdqu1lLfbX2rcfS6cRcRiwNXAvcCVwWmYuG1eEkiRJkiRJelA7EkJbVNubWuyvtW85xn6PaHj+6Yh4W2YeN5ZOIuLKFru2HmM8kiRJkiRJa4V2FD94eLW9q8X+ldV2g1H2933gZZQE0mxgO+AoYD3g2Ih48TjjlCRJkiRJEu1bZaxtMvPghqYrgXdFxNXAl4BPUpa1H21/C5u1VyOHpkYhCklT2qpV3S9iPHNmH9GDBYz7ZvcxcOhAt8Ng8K7BKVO7SJIkSWqHdiSEaquKzW6xf061vWOC1/kK8FFg24jYKjOvn2B/kjQqF1/c/UTAokUDzJrVewWMZ8QMCzdLkiRJk6Adf27+a7XdrMX+WvsNE7lIZg4B11VPN51IX5IkSZIkSb2sHQmhy6vtTi3219qvaMO1Nq62K4c9SpIkSZIkSS21Y8rY+cBtwNYRsWNm/q5h/77V9oyJXCQiFgLbUopXXz2RviRJ4zO4YgYs624MfX0wo/fKKUmSJEltNeGEUGbeFxH/C/wn8PmIeF5mrgSIiEOA7YFfZualtXMi4iDgIOC0zHxvXftewD2Z+fP6a0TE9sA3gQCOzcz7Jhq3JDUzc2YfixZ1t4jxqlWDU6JuUTMLFs/tdggMDMA8ywpJkiRJE9KuVcY+CjwXWARcGxHnUpaNfyrlb8kHNhzfTxnt01gLaBfgQxFxA2Uq2l3AYyjTztYFzgH+o00xS9IaImb0ZPFmSZIkSb2lLYPuM/Me4FnARyhJnJdQEkInADtl5p9H2dWPgeOA24HdKNPNHgucB7wBeG5m3t2OmCVJ09TQULcjkCRJkqa9do0QokrUfLB6jHTs4cDhTdovBC5sV0ySpLXQihWwSX+3o5AkSZKmtbYlhCRJa5e+uckA3Z0+N0g/C7iqqzFIkiRJayMTQpKkpmbMgHkMdjsMSZIkSZPAhXslSZIkSZJ6jCOEJEnN9fWVNd676ZpbYHF3Q5AkSZLWRiaEJEnNzZgB87pbQ4jB6O71JUmSpLWUU8YkSZIkSZJ6jAkhSZIkSZKkHmNCSJIkSZIkqcdYQ0iSpoFVq6bG8u8zZ/YR4d8SJEmSpOnOhJAkTQMXX7yg2yEAsGjRALNmdbnQtCRJkqQJ88+8kiRJkiRJPcaEkCRJkiRJUo8xISRJkiRJktRjrCEkSVPMzJl9LFo00O0wWLVqcMrULpqKBu/qfqHvvtl9zLDItyRJksbBhJAkTTERMyzcPA0sOKb7ybKBQweYN8fPiiRJksbOhJAkaVoZXDEDlnX4misDVvav3viw5TAjOxuIJEmS1CYmhCRJ08qCxXO7cNV+1shCHTYP5nR/2pgkSZI0HhYekCRJkiRJ6jGOEJIkaRyWvuUq+ud1bsrY4F2DU6JukSRJktYOJoQkSRqH/jn9zJvT7SgkSZKk8TEhJEmasvrmJgM0rKK19Cro729+wiQZHIQFDs6RJEnSWsSEUAcN5dAabSvuuZV1V3Z2uZzBuyyCKml6mDED5tHw36z+pDFHJEmSJGlsTAh10K333r5G226nvYDbZq7oQjSSNE0NdiGpPRiUlcbqDA3h2gySJEmarkwISZKml67M3Wqy7PyKFbBJZ6euSZIkSe3inzYlSZIkSZJ6jCOEJEkah8EVM9YYNDSp11sZsHL1EUlDa5amkyRJkkbFhFCX/WCvk9h2hyd3Owz6Zvd1OwRJmlYWLJ7b4SuuOW3tT/96DTO26HAYDfpm9zEjHHAsSZI03ZgQ6rKN19uIeXNcLkfS9LBqVYcLOm8wBH9fulrTzHXmEp1OQFxzCyzu7CVHY/Hxu8Gc7q4cOXDogP8fkyRJmoZMCEmSRu3ii7tR0Hl1ixYNMGtWhxMQg9HZ60mSJEmTzDHekiRJkiRJPcYRQpIkjaBvbjJAw6ikpVdBf+eWnb/mrytY/JRO1y2SJEnS2sqEkCRJI5gxA+bRUKunP2nMEU2modxojbalb7mK/nnZsRgG7xpkwTHdnzYoSZKkiTMhJElqaubMPhYtGuhqDKtWDU6JukVTQbOVvPrn9DNvTheCkSRJ0rRnQkiS1FTEjM4Xbx6Fjq90BnD/IGy4etPMZQN0tNT0YFCWnq8zNITlACVJkjQeJoQkSdNK10YMnb7600WLtmPWbZ0MoB9YtnrTihWwSefqGEmSJGntYUJIkqRpanDFjDVyRJN6vZUBK1dPQA3c2YURW030ze5rOq1OkiRJzZkQkiRpmlqwuNOrjq05Smk75sGc7ieFBg4dYN6cqTfFUZIkaapqW0IoIh4GvBd4JbAFsAI4C/hAZv5tjH1tDBwOvAR4JPAP4DTg8My8tV0xS5KkCVo5NZIwQ0PdjkCSJGl6aUtCKCLWB34O7ArcDHwP2Ar4V+CFEbFrZv55lH31AxcCjwX+TKnasBB4O7BnRDwtM1e0I25J0tQ2FVY6A1h13wAXX7Ld6m1/OA/W6dwInfuvu40NX7j66Jzbb+8js8vTpI5Z2t3rV1a8ZZBNNuh2FJIkSdNHu0YIvZ+SDLoQeF5m3gkQEYcAnwKOA3YfZV9HU5JBpwL7Zeb9VV+fBd4GHAUsaVPckqQpbKqudAZw8Z8Wd/yap5+++vOXvGSA226bmu9Pp624ewXLVmZXY7COkSRJmk4mnBCKiFnAQdXTt9aSQQCZeVREvA54ZkQ8OTMvHaGvTYH9gfuAt9SSQZXDKNPRDoiId2dm9/9kLElSF/1+o21Yt4PLzi+jj2dwfseuNxZP//yLyNnLuxrD0rdcRf+c7q761tcHM8xJSZKkUWjHCKHdgA2B6zLzt032nwJsD+wNDJsQAvYAZgDnZuY/63dk5r0RcQZwILAXcMIE45YkaVq77oRbO3zFFZzO/A5fc3SWLLmSW2/t7mipp30p6eiyb02cccpfmLuxBZWmgoVP3rXbIUiSNKx2JIR2qLaXtdhfa9++TX0dOMq+JElSjzjhhIXdDmFKeABYdke3o1DR3SmMkiSNpB0JoS2q7U0t9tfat+xwXwBExJUtdj3+uuuuY+HCzv0Cee89d3P/A6u3rbvOq1hv/Yd1LAZJ0vhkbtvl69/P3Xdf19UYJI3enDndSVJed911AJt35eKSpGmlHQmhh1fbu1rsX1ltR7P2Rzv7GsnQvffeu3Lp0qU3tqGvsdq62l4Hf+nC5dVFdfdePcZ737u8973Le9+7toal0J17vzmtf5eWJOlB7VplbMrKzCk3hrw2amkqxqbJ5b3vXd773uW9713e+97lvZckTQftWIeitqrY7Bb751Tb0cxob2dfkiRJkiRJaqIdCaG/VtvNWuyvtd/Q4b4kSZIkSZLURDsSQpdX251a7K+1X9HhviRJkiRJktREOxJC5wO3AVtHxI5N9u9bbc8YRV9nAUPA0yNifv2OiFgP2JuyouqZ445WkiRJkiSpx004IZSZ9wH/Wz39fETU6vwQEYcA2wO/zMxL69oPioirI+LjDX3dDJwMzAKOiYj6otdHAPOAEzNzYKJxS5IkSZIk9arIzIl3ErE+cA7wVOBm4Fxgy+r5MmDXzPxz3fGHAx8CvpqZSxr66gcuoizVeh1wCbAQ2A64tuprxYSDliRJkiRJ6lHtmDJGZt4DPAv4CHAX8BJKQugEYKf6ZNAo+hoEdgE+Rxkp9FJgQ+CzwC4mgyRJkiRJkiamLSOEJEmSJEmSNH20ZYSQJEmSJEmSpg8TQpIkSZIkST3GhJAkSZIkSVKPMSEkSZIkSZLUY0wISZIkSZIk9RgTQm0QEQ+LiP+KiGsi4p6I+HtEHBcRjxpHXxtHxGci4oaIuLfaHh0RG01C6Jqgdtz7iNgoIl4VESdHxF8i4r6IuCMifh0Rb4+ImZP5GjQ+7fx339Dv4yLi7ojIiPhpu+JV+7T73kfEVhHxherf/70RMRgRF0bEYe2OXRPT5v/f/0tE/DAilkXEqohYHhFnR8RLJyN2jU9EPDki/iMiTo2Im6r/No97iV5/z5MkTSUuOz9BEbE+8AtgV+Bm4FxgK2AXYBmwa2b+eZR99QMXAo8F/gxcAiysHtcAT8vMFW1+CRqndt37iPgo8J9AAr+j3Ot5wG7AesB5wPMz8662vwiNSzv/3Tfp+xfAM4EAfpaZz21HzGqPdt/7iNgTOAV4GHAZcC3QBzwRWJmZj21n/Bq/Nv///h3Apyn/3b8QuBHYHHga5d/+xzLzP9v7CjQeEXE68OLG9syMcfTl73mSpCnFEUIT937KL4cXAttk5n6Z+VTgXZQv9ceNoa+jKb8knApsW/W1HfA5YBvgqHYGrglr171fCRwBbJWZO2XmKzPzOZQvhH8FFlfX0tTRzn/3D4qIfwN2B77cpjjVfm279xHxeMp/71cCizPzKZm5f2Y+D3gU8Mq2R6+JaMu9j4h5wCeAVcCzMnO36r/7u1H+/d8LvDciHjMJr0FjdyHwEeBFwKaU+zNeR+PveZKkKcQRQhMQEbOAAWBDYKfM/G3D/suB7YGnZOalI/S1KXATcD+wRWb+s27fepS/Hs4F/l9mDrT1hWjM2nnvR7jO/sA3gOsz89ETCFltMln3PiI2Aa6i/MX4Y5SRCI4QmkLafe8j4kxgT+AFmXnmJISsNmnz/+9fCJwB/Dgz92iy/3uU5MN+mfntNr0EtUlE3AOsN9YRQv6eJ0maihwhNDG7UX45vK7xl8PKKdV271H0tQflfpxb/0sCQGbeS/nlcR1gr/GHqzZq570fzuXV9v9NsB+1z2Td+89Qpg29ZQKxaXK17d5HxObA84E/mwyaFtr57360I0yWj/I4TQ/+nidJmnJMCE3MDtX2shb7a+3bd7gvTb5O3a/alIF/TLAftU/b731E7AXsR6kb8qcJxKbJ1c57vzvl/8EXRMS6EfGKqtDs/0bEmyJi4wnGqvZq573/DXAr8OyIeGb9joh4BiVReC2lRpHWHv6eJ0mactbtdgDT3BbV9qYW+2vtW3a4L02+Tt2vt1fb702wH7VPW+99RMwBjgH+CHxyYqFpkrXz3i+otndSvvjv2rD/vyNi38z8xdhC1CRp273PzNuqemHfAH4RERdU528GLALOB16bmfdNLGRNMf6eJ0machwhNDEPr7atVn9aWW036HBfmnyTfr8i4k3Acyl/Sf7EePtR27X73n+U8gXgTX4BnPLaee9rI4BeDzweeBWlfsi2wInVz6eNdyl7tV1b/91n5qmU+lHLKdPR9qu2dwBnA38bd6Saqvw9T5I05ZgQkqagiHg6paZMAgdm5t+7HJImQUQ8BTgY+FpmntPlcNRZtf//rgu8MTNPzsxbMvOazHwNcDGlZo01pdZCEfEu4KfAryhThB5ebX8O/BdlFSpJkqRJZUJoYu6strNb7J9Tbe/ocF+afJN2vyJiO8oUsVnA2zPztLGHp0nUlnsfEetSlpe/FTi0LZFpsk3Gf/PvBL7TZP/x1faZTfap89p27yNid+BI4HfAyzPz95m5MjN/D+xbtb8gIvacQLyaevw9T5I05VhDaGL+Wm03a7G/1n5Dh/vS5JuU+xURj6ZMF9gYODwzPze+8DSJ2nXvNwN2pBQM/07EaisYb1RtnxwR5wBk5u5jC1OToJ3/7mvH/DUzs8n+66vt/NGFpknWznv/mmp7WmYO1e/IzAci4lTKfxueAfxojHFq6vL3PEnSlGNCaGJqS4Lv1GJ/rf2KDvelydf2+xURmwI/ATYFPpOZHx5/eJpE7b73j6wezWyEI0Smknbe+9rS5a1WE5tbbe9ssV+d1c57X/vif1uL/bV2V5pbu/h7niRpynHK2MScT/nFbeuI2LHJ/n2r7Rmj6OssYAh4ekSs9hfhiFgP2Bt4ADhz3NGqndp576mWmP4xsDVlqsg72xCjJkdb7n1mXp+Z0ewBPKs67Gd1beq+dv67v4BSUPiREbFtk/21ROBvm+xT57Xz3v+j2j6lxf6dq+31ow1O04K/50mSphwTQhNQrQj0v9XTz1fLRwMQEYdQCkT+MjMvrWs/KCKujoiPN/R1M3AypW7MMVV9kZoj4P+3d+fBklblHce/PwVCEBfCkhA1jimtbAgkRoJKwqLRAA5LCGohiROIYmkkYJmiBEGIQVCJiWJJgElYBCQoOGwxQaTYYhFMQNFEFBUQgQQHEISRZeDJH+ft0LTdM/fe7jtbfz9VXe/0ec97zul+79Lz3HOew+bAWVV1z/y8Gs3GJO99ko2AS4GXAecBbxuxhERrgEnee61dJvwzfznwMSBdW8/pu+a1wCJaUvmT5+nlaBYm/H2/pDu+Jckb+k8k2ZO249yTgPnj1kJ+zpMkrU1cMja+v6ZtDf4q4JYk19C2kP4d4IfAAQP1N6NtK7zlkLYOAbYH9gFuTvIfwG8AWwG3AO+Zh/Fr7iZ1748FXkn7y+By4B8G8skAUFWLJjh2jWeS3/dau0zy3n+UNhvstcC3k1zX1d8eeCZwRFVdPx8vQnMyqXu/hJZIfF/g4u53/a3Ai3lq1tARVfWteXgNmqUkuwNH9hVt0JVf11f2waq6tPu3n/MkSWsNZwiNqaoeoX2g/yCwDNiL9gHxdOC3qup7s2hrKbAdcCLtA8fetG2HPwFsV1X3TXLsGs8E730vT8QzaX8ZfuuIh9YQk/y+19plwj/zHwd2Aw4DlgKvp80UvApYWFUfmuTYNZ5J3ftuBuibgANp286/hPb7fgFtudCu3vs1yua0oF/v0fuLTX/Z5jNpyM95kqQ1TVyZIkmSJEmSNF2cISRJkiRJkjRlDAhJkiRJkiRNGQNCkiRJkiRJU8aAkCRJkiRJ0pQxICRJkiRJkjRlDAhJkiRJkiRNGQNCkiRJkiRJU8aAkCRJkiRJ0pQxICRJkiRJkjRlDAhJkiRJkiRNGQNCkiRJkiRJU8aAkCSNkKSS3DbG9Tt1bVSSB5JsuIK6n+qre/rAuUUjyr/Ulb9vBu2eOcMxH903jt7j0STfS3JqkpfMpJ1xJDm963enNa2fvnt6+kB57x4dPVDeez8XTWDIkiRJ0sQYEJKkVeM5wBuGnUiyPvDGObR5EPAIcNSwQE2SVwPvAJYC75ll218DzugeXwB+Fvgz4MYkr5jDWNUnyZVdoGjB6h6LJEmSppMBIUka7deA10ygna8DTwD7jzi/K7ApcMNsGq2q7wDHABsCJ/efS7IBcAoQ4NCqWjrLMS+pqkXdYy/gpcClwMbASbNsa11yPe3rYuSsrAGf7Op/ft5GJEmSJM2BASFJGqGqbq6q706gqXuALwK7JtlkyPn9gSeBc+bQ9gnATcAuA8uSDgd+Hbisqs6aQ7tPU1UPAe/snr48yQvGbXNtVFXLuq+Lu2dYf2lX/4H5HpskSZI0GwaEJGmEUTmEkrwqyZIkt3f5df4nyfVJjk+y8YjmzgI2YGBpWJLnAAuBK4E7ZzvGqlpOW8r1JPA3SbZI0pvBsoy2ZGwiqur7wH3d0xfC03PkJNkuySVJ7u3Ktu1dm+SPk1yb5MEky5LclOR9K8qr1F23a3fdQ0nuT3JBkl8dUm/DJAcmubDLd/STJD9KcnWSN6/stc2in6E5hFbQ7tNyCCVZkKSAHbsqt/bna+rqXNI9f92INjfqXtuPkzx7JuOQJEmSBhkQkqRZSLIQuAbYA7gbuAC4Efg54DBgsxGXLgEeBt4yUL4PbcnX2XMdU1V9BfhEN4aPA6fSgk9HVdWtc213UJJnAM/qnj46cPr3gGuBBcBlwNW0IBVJTgbOBF5Oe+8uBbYEPgRckWSjEV3u29XdALgYuAvYG7guyTYDdRcAi4HfBm4DLgS+CmwPfGYw2fMY/YzrIVpepv/tnp/PU7mazujKesv/3raC8T4XOLeqfjzh8UmSJGlKrLe6ByBJa5n30oLpf1RV5/ef6JIt3zvsoqp6OMkSYL8kL6qq27tT+9MSQ38O2G2Mcb2fFsTozYa5Afi7Mdob5veBnwEeB24eOPenwGFV9ZH+wiT7AG+nBVl2qqpbuvLnApcAOwB/RXtfB70TeHtVndpdE+A4WuDtdOA3++r+sBvfl6qq+vp/MXAFcGSS06vqtjH7GUuXy2lRkiuBnwfeO2RM/wzcAeyZZIuqumfgfC9QdOqkxiVJkqTp4wwhSZqdzbvj5YMnquorK5mxcTYtyfN+AEmeD+wEXFxVD44zqKp6mKcnlj6kqp4Yp82eJM9LsjdwWld0RlUtG6j2deCjQy4/uDse0wsGdeN9AHgXUMBBI5aOfbkXpOmuKeBI4AfAtkl26Dt3b1Vd3h8M6spvBY6l/b5bOOIlzrifVaG7b6cC6wNv7T/XLWN7NXBTVV2/KsclSZKkdYsBIUmanf/sjp9O8opuGdVMfZGWYLq3bGw/2s/hsZM+d8mqD+kr2nfMJj/Ql9fmftrSuC2Bfxnop+eSwWBMkvVpS7ZgyJK4qrqJlhB7Y2DbIW2eO+Sax2mzqQB+d/B8kh2SvD/JSUlO63L99N6Llw7pY079rAKLgV5+qH692UGnrNrhSJIkaV3jkjFJmp3DgZfRZpssBO5Pci1wEXBWVT0y6sKqWp7kXODgLuHy/rQlZl+YwLhOALYAPgP8AfCuJJ/u8gvNxddoOXig5Qu6i7Yc69oR9b8/pGxTWl6epd0MpmFuA7YBnj/k3O1DynrXAPxir6BbgnYBsMuIawBGJWCecT+rSlXdneQi4A+T7FhVVyXZAPgT4CeMkXNKkiRJAmcISdKsVNUdtMTFrwdOpOV6WUhb4nNTkk1X0kTvP/LHAVsD53WzUeYsyc7AAbRdyt5By33zDOCUJHMN/C+pqkXd46CqOmYFwSBoeZDmolZeZUY+TAsGXUVbhrcZsF5VhXavoC3XW5v8fXfszQrai/a6PltVP1odA5IkSdK6w4CQJM1SVS2vqsuq6uCq2oa2w9UVtCVJh63k2uuBW2izeGDM5WJd7p1e7qB3d7mIFtN2/NqW4cu7VpV7gceAzZI8a0SdBd3xziHnXjTiml75XX1lewNPAHtU1VVdTqFeDqVfXsk4Z9PPqnQ58B1gn25JoMvFJEmSNDEGhCRpTN2OYR/unm41g0v+kRYs+WpVfXnM7o+iBaKWVNXnu/EUbWevx4BjkowKeMyrbubTdd3TNw+eT7IVbbnYQzy1PK3fG4dcsx6wT/e0f8bSJsCDI5Jz/1Q7Y/QzKY91x5EzuLr7eAqwIe0+vwb4ZlX92zyMR5IkSVPGgJAkzUKSQ5P8wpBTvS3j71hZG1V1fFVtVlVjbWeeZGvgL4EHgT8f6OObtCDVRsCnxulnTCd2x6OT/P9MnSTPBj5JW8Z18ojcSzskOWCg7Bjgl2i7bF3TV/5tYJMkb+qvnORQYOeVjHE2/UxKb9bRr6yk3mm0HE6H0N4rt5qXJEnSRBgQkqTZ+QBwZ5IbkvxTkvOSfAv4C+A+WnLnedftbnYqbYbJ4VU1bMnVsbRAyW5JVjZLZl5U1edos1xeAHwjySVJzgO+C+xIm0F01IjLTwIWJ/n3JOck+QYtqfeDwKKBusd1x3OTXN3V/y/a/fjblQxzNv1MykXd8Zwkn02yOMniwUpVtRQ4v3v6KHDmPI1HkiRJU8aAkCTNzrtp25RvBOxKywW0HPgYsHVV3bIKx7EdLaBy0rAKVfUoLck0wMeTPG/VDO2nxnEQbXesG2lBoIXAPcARwC5VtWzEpecBe9ByA+1JCypdCLyyqm4c6ONsYHfa+7Et7d7cRUs0fRErNuN+JqWqLgAOBX5Aez8O7B7DXNEdL6iqe+djPJIkSZo+aSkKJEnSmijJvwKvA3auqitX83AkSZK0jjAgJEnSGipJbxbYf1fVTBKWS5IkSTMycncTSZK0eiQ5npbUendaMukjVu+IJEmStK5xhpAkSWuYJLcBLwRuB06oqtW5U5wkSZLWQQaEJEmSJEmSpoy7jEmSJEmSJE0ZA0KSJEmSJElTxoCQJEmSJEnSlDEgJEmSJEmSNGUMCEmSJEmSJE0ZA0KSJEmSJElTxoCQJEmSJEnSlDEgJEmSJEmSNGUMCEmSJEmSJE0ZA0KSJEmSJElTxoCQJEmSJEnSlDEgJEmSJEmSNGUMCEmSJEmSJE0ZA0KSJEmSJElT5v8ArhbxzwPg0sgAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 900x600 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x600 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x600 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "Y_predict_train = LBNdnn.predict_proba(inputs.astype(np.float32))\n",
    "Y_predict_val = LBNdnn.predict_proba(inputs_val.astype(np.float32))\n",
    "\n",
    "print(Y_predict_val)\n",
    "axis = np.linspace(0,1,20)\n",
    "\n",
    "#targets = ['LL', 'MIX', 'TT', 'BKG']\n",
    "predictions = targets\n",
    "colors = ['r', 'g', 'b', 'y']\n",
    "\n",
    "Y_targets = {}\n",
    "Y_predictions = {}\n",
    "\n",
    "for itarget, target in enumerate(targets):\n",
    "    Y_targets[target] = LBNdnn.predict_proba( inputs_val.astype(np.float32)[baseline_targets[itarget]:baseline_targets[itarget+1]] )\n",
    "    Y_predictions[target] = {}\n",
    "    for iprediction, prediction in enumerate(predictions):\n",
    "        Y_predictions[target][prediction] = Y_targets[target][:,iprediction]\n",
    "\n",
    "for prediction in predictions:\n",
    "    for target, mycolor in zip(targets, colors):\n",
    "        plt.hist(Y_predictions[target][prediction].numpy(), bins = axis, label = prediction + ' class - ' + target.strip('is') + ' sample', histtype='step', color=mycolor,  density=True, linewidth=2 )\n",
    "    plt.title('Validation dataset')\n",
    "    plt.xlabel(prediction + ' Probability')\n",
    "    plt.legend(frameon=False, bbox_to_anchor=(1.05, 1.))\n",
    "    #plt.yscale('log')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "ac511b3c-889a-41a0-8484-76841bf345c2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classifier categorical accuracy =  0.48306782722943015\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score\n",
    "from scipy.stats import ks_2samp, anderson_ksamp\n",
    "\n",
    "Y_true = labels_val\n",
    "Y_pred = LBNdnn.predict_proba(inputs_val.astype(np.float32))\n",
    "\n",
    "# The classifier categorical accuracy is a good and simple metrics to evaluate the overall classification performance\n",
    "Y_true_max = np.argmax(Y_true, axis=1)\n",
    "Y_pred_max = np.argmax(Y_pred, axis=1)\n",
    "print(\"Classifier categorical accuracy = \", accuracy_score(Y_true_max, Y_pred_max))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "b7dda20d-13a1-4f21-9483-e988bf0984fc",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import classification_report, confusion_matrix\n",
    "import itertools\n",
    "#from matplotlib.backends.backend_pdf import PdfPages\n",
    "\n",
    "def plot_confusion_matrix(cm, classes,\n",
    "                          normalize=False,\n",
    "                          title='Confusion matrix',\n",
    "                          cmap=plt.cm.Blues):\n",
    "    \"\"\"\n",
    "    This function prints and plots the confusion matrix.\n",
    "    Normalization can be applied by setting `normalize=True`.\n",
    "    \"\"\"\n",
    "    plt.imshow(cm, interpolation='nearest', cmap=cmap)\n",
    "    plt.title(title)\n",
    "    plt.colorbar()\n",
    "    tick_marks = np.arange(len(classes))\n",
    "    plt.xticks(tick_marks, classes, rotation=45, fontsize = 16)\n",
    "    plt.yticks(tick_marks, classes, fontsize = 16)\n",
    "\n",
    "    thresh = cm.max() / 1.2\n",
    "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
    "        plt.text(j, i, cm[i, j],\n",
    "                 horizontalalignment=\"center\",\n",
    "                 color=\"white\" if cm[i, j] > thresh else \"black\",fontsize=10)\n",
    "\n",
    "    plt.xlabel(\"Predicted label\", fontsize=16)\n",
    "    plt.ylabel(\"True label\", fontsize=16)\n",
    "\n",
    "    \n",
    "    plt.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "21632233-f418-4238-af28-48a0d7c934ba",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x600 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "C = confusion_matrix(Y_true_max, Y_pred_max, normalize=\"true\")\n",
    "C = np.around(C, decimals=3)\n",
    "#target_names = ['LL', 'MIX', 'TT','BKG']\n",
    "\n",
    "plt.figure()\n",
    "plot_confusion_matrix(C, classes=targets, title='')\n",
    "plt.show(1) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "5066e1de-7e63-4bf8-8669-8be4b182c04a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['E', 'beta', 'eta', 'gamma', 'm', 'p', 'pair_cos', 'pair_dr', 'pair_ds', 'pair_dy', 'phi', 'pt', 'px', 'py', 'pz']\n",
      "20\n",
      "n features = 910.0\n",
      "feature shape =  (120606, 910)\n",
      "tf.Tensor(\n",
      "[[ 2.3696016e+01  2.8708834e+01  2.2615669e+01 ... -6.3419662e-02\n",
      "  -6.6889167e-01  7.1485871e-01]\n",
      " [ 3.1475433e+01  6.3175903e+01  4.6979534e+01 ... -4.7365878e-02\n",
      "  -4.9957144e-01  5.3390259e-01]\n",
      " [ 2.1237879e+01  3.5391109e+01  4.2443199e+01 ... -7.8393072e-02\n",
      "  -8.2681721e-01  8.8363707e-01]\n",
      " ...\n",
      " [ 1.5219990e+01  1.7237530e+01  2.5030783e+01 ... -5.1940527e-02\n",
      "  -5.4782045e-01  5.8546734e-01]\n",
      " [ 1.7962805e+01  2.1165071e+01  2.8516485e+01 ... -6.3210577e-02\n",
      "  -6.6668624e-01  7.1250170e-01]\n",
      " [ 2.8903145e+01  2.1932037e+01  4.2018242e+01 ... -4.4852030e-02\n",
      "  -4.7305772e-01  5.0556678e-01]], shape=(120606, 910), dtype=float32) 120606\n"
     ]
    }
   ],
   "source": [
    "lbn = LBN(20, boost_mode=LBN.PAIRS)\n",
    "\n",
    "# run build, which creates trainable variables and an eager callable in eager mode\n",
    "lbn.build(inputs.shape, features=[\"E\", \"pt\", \"eta\", \"phi\", \"m\", \"pair_cos\", \"pair_dr\", \"pair_dy\"])\n",
    "\n",
    "# show available features\n",
    "print(lbn.available_features)\n",
    "print(lbn.n_particles)\n",
    "# build certain features for all 10 boosted particles\n",
    "features = lbn(inputs)\n",
    "print('n features =', lbn.n_features)\n",
    "print('feature shape = ', features.shape)\n",
    "# print features (10 x E, 10 x pt, 10 x eta, 10 x phi, 10 x m, 45 x pair_cos)\n",
    "print(features, len(features))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "18420f8b-b656-4795-9c84-c5aa67916db3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_2\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " LBN (LBNLayer)              (None, 910)               1400      \n",
      "                                                                 \n",
      " dense_16 (Dense)            (None, 910)               829010    \n",
      "                                                                 \n",
      " dense_17 (Dense)            (None, 150)               136650    \n",
      "                                                                 \n",
      " dense_18 (Dense)            (None, 150)               22650     \n",
      "                                                                 \n",
      " dense_19 (Dense)            (None, 150)               22650     \n",
      "                                                                 \n",
      " dense_20 (Dense)            (None, 150)               22650     \n",
      "                                                                 \n",
      " dense_21 (Dense)            (None, 150)               22650     \n",
      "                                                                 \n",
      " dense_22 (Dense)            (None, 150)               22650     \n",
      "                                                                 \n",
      " dense_23 (Dense)            (None, 4)                 604       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 1,080,914\n",
      "Trainable params: 1,080,914\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "910.0\n"
     ]
    }
   ],
   "source": [
    "# LBN pairs mode\n",
    "LBNdnn = SimpleLBNNeuralNetwork(lorentz_shape=lorentz_shape, nEpochs=300, learning_rate=0.0001, N_NODES=150, n_layers=6, n_particles=20, boost_mode=LBN.PAIRS, n_features=lbn.n_features)\n",
    "print(lbn.n_features)\n",
    "# Save initial set of weights (before training) to re-initialize the ADNN in later steps.\n",
    "# Useful if we want to restart always from the same starting point during the optimization studies.\n",
    "LBNdnn.save_weights(\"my_simpleLBNDNN_model_polVBSvsWW_new_PAIRS\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "4e858e23-fe81-4ecc-8c49-a356c89fc07f",
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 27%|██▋       | 81/300 [01:01<03:05,  1.18it/s]2023-04-13 13:08:19.969398: W tensorflow/core/common_runtime/bfc_allocator.cc:462] Allocator (GPU_0_bfc) ran out of memory trying to allocate 418.67MiB (rounded to 439005952)requested by op ReluGrad\n",
      "If the cause is memory fragmentation maybe the environment variable 'TF_GPU_ALLOCATOR=cuda_malloc_async' will improve the situation. \n",
      "Current allocation summary follows.\n",
      "Current allocation summary follows.\n",
      "2023-04-13 13:08:19.969642: I tensorflow/core/common_runtime/bfc_allocator.cc:1010] BFCAllocator dump for GPU_0_bfc\n",
      "2023-04-13 13:08:19.969665: I tensorflow/core/common_runtime/bfc_allocator.cc:1017] Bin (256): \tTotal Chunks: 237, Chunks in use: 232. 59.2KiB allocated for chunks. 58.0KiB in use in bin. 2.0KiB client-requested in use in bin.\n",
      "2023-04-13 13:08:19.969672: I tensorflow/core/common_runtime/bfc_allocator.cc:1017] Bin (512): \tTotal Chunks: 80, Chunks in use: 68. 49.8KiB allocated for chunks. 43.8KiB in use in bin. 34.4KiB client-requested in use in bin.\n",
      "2023-04-13 13:08:19.969678: I tensorflow/core/common_runtime/bfc_allocator.cc:1017] Bin (1024): \tTotal Chunks: 9, Chunks in use: 9. 12.5KiB allocated for chunks. 12.5KiB in use in bin. 9.3KiB client-requested in use in bin.\n",
      "2023-04-13 13:08:19.969684: I tensorflow/core/common_runtime/bfc_allocator.cc:1017] Bin (2048): \tTotal Chunks: 9, Chunks in use: 9. 25.0KiB allocated for chunks. 25.0KiB in use in bin. 23.5KiB client-requested in use in bin.\n",
      "2023-04-13 13:08:19.969689: I tensorflow/core/common_runtime/bfc_allocator.cc:1017] Bin (4096): \tTotal Chunks: 10, Chunks in use: 10. 50.0KiB allocated for chunks. 50.0KiB in use in bin. 45.1KiB client-requested in use in bin.\n",
      "2023-04-13 13:08:19.969695: I tensorflow/core/common_runtime/bfc_allocator.cc:1017] Bin (8192): \tTotal Chunks: 0, Chunks in use: 0. 0B allocated for chunks. 0B in use in bin. 0B client-requested in use in bin.\n",
      "2023-04-13 13:08:19.969702: I tensorflow/core/common_runtime/bfc_allocator.cc:1017] Bin (16384): \tTotal Chunks: 13, Chunks in use: 13. 385.5KiB allocated for chunks. 385.5KiB in use in bin. 384.4KiB client-requested in use in bin.\n",
      "2023-04-13 13:08:19.969724: I tensorflow/core/common_runtime/bfc_allocator.cc:1017] Bin (32768): \tTotal Chunks: 4, Chunks in use: 3. 205.5KiB allocated for chunks. 154.8KiB in use in bin. 89.1KiB client-requested in use in bin.\n",
      "2023-04-13 13:08:19.969730: I tensorflow/core/common_runtime/bfc_allocator.cc:1017] Bin (65536): \tTotal Chunks: 36, Chunks in use: 36. 3.22MiB allocated for chunks. 3.22MiB in use in bin. 3.18MiB client-requested in use in bin.\n",
      "2023-04-13 13:08:19.969736: I tensorflow/core/common_runtime/bfc_allocator.cc:1017] Bin (131072): \tTotal Chunks: 3, Chunks in use: 3. 545.2KiB allocated for chunks. 545.2KiB in use in bin. 294.6KiB client-requested in use in bin.\n",
      "2023-04-13 13:08:19.969742: I tensorflow/core/common_runtime/bfc_allocator.cc:1017] Bin (262144): \tTotal Chunks: 4, Chunks in use: 4. 1.85MiB allocated for chunks. 1.85MiB in use in bin. 1.85MiB client-requested in use in bin.\n",
      "2023-04-13 13:08:19.969748: I tensorflow/core/common_runtime/bfc_allocator.cc:1017] Bin (524288): \tTotal Chunks: 23, Chunks in use: 23. 14.20MiB allocated for chunks. 14.20MiB in use in bin. 12.78MiB client-requested in use in bin.\n",
      "2023-04-13 13:08:19.969754: I tensorflow/core/common_runtime/bfc_allocator.cc:1017] Bin (1048576): \tTotal Chunks: 6, Chunks in use: 6. 9.51MiB allocated for chunks. 9.51MiB in use in bin. 8.52MiB client-requested in use in bin.\n",
      "2023-04-13 13:08:19.969761: I tensorflow/core/common_runtime/bfc_allocator.cc:1017] Bin (2097152): \tTotal Chunks: 72, Chunks in use: 70. 180.01MiB allocated for chunks. 175.38MiB in use in bin. 173.52MiB client-requested in use in bin.\n",
      "2023-04-13 13:08:19.969768: I tensorflow/core/common_runtime/bfc_allocator.cc:1017] Bin (4194304): \tTotal Chunks: 21, Chunks in use: 17. 129.37MiB allocated for chunks. 103.94MiB in use in bin. 89.79MiB client-requested in use in bin.\n",
      "2023-04-13 13:08:19.969774: I tensorflow/core/common_runtime/bfc_allocator.cc:1017] Bin (8388608): \tTotal Chunks: 101, Chunks in use: 97. 961.60MiB allocated for chunks. 915.55MiB in use in bin. 890.41MiB client-requested in use in bin.\n",
      "2023-04-13 13:08:19.969780: I tensorflow/core/common_runtime/bfc_allocator.cc:1017] Bin (16777216): \tTotal Chunks: 32, Chunks in use: 27. 747.36MiB allocated for chunks. 661.04MiB in use in bin. 603.71MiB client-requested in use in bin.\n",
      "2023-04-13 13:08:19.969786: I tensorflow/core/common_runtime/bfc_allocator.cc:1017] Bin (33554432): \tTotal Chunks: 22, Chunks in use: 18. 933.37MiB allocated for chunks. 734.25MiB in use in bin. 627.02MiB client-requested in use in bin.\n",
      "2023-04-13 13:08:19.969804: I tensorflow/core/common_runtime/bfc_allocator.cc:1017] Bin (67108864): \tTotal Chunks: 24, Chunks in use: 20. 2.13GiB allocated for chunks. 1.78GiB in use in bin. 1.78GiB client-requested in use in bin.\n",
      "2023-04-13 13:08:19.969810: I tensorflow/core/common_runtime/bfc_allocator.cc:1017] Bin (134217728): \tTotal Chunks: 16, Chunks in use: 15. 2.55GiB allocated for chunks. 2.36GiB in use in bin. 2.02GiB client-requested in use in bin.\n",
      "2023-04-13 13:08:19.969816: I tensorflow/core/common_runtime/bfc_allocator.cc:1017] Bin (268435456): \tTotal Chunks: 16, Chunks in use: 13. 5.87GiB allocated for chunks. 4.85GiB in use in bin. 4.18GiB client-requested in use in bin.\n",
      "2023-04-13 13:08:19.969822: I tensorflow/core/common_runtime/bfc_allocator.cc:1033] Bin for 418.67MiB was 256.00MiB, Chunk State: \n",
      "2023-04-13 13:08:19.969836: I tensorflow/core/common_runtime/bfc_allocator.cc:1039]   Size: 268.21MiB | Requested Size: 46.24MiB | in_use: 0 | bin_num: 20, prev:   Size: 21.96MiB | Requested Size: 21.96MiB | in_use: 1 | bin_num: -1, next:   Size: 21.96MiB | Requested Size: 21.96MiB | in_use: 1 | bin_num: -1\n",
      "2023-04-13 13:08:19.969845: I tensorflow/core/common_runtime/bfc_allocator.cc:1039]   Size: 368.81MiB | Requested Size: 1.84MiB | in_use: 0 | bin_num: 20, prev:   Size: 418.67MiB | Requested Size: 418.67MiB | in_use: 1 | bin_num: -1, next:   Size: 1.84MiB | Requested Size: 1.84MiB | in_use: 1 | bin_num: -1\n",
      "2023-04-13 13:08:19.969853: I tensorflow/core/common_runtime/bfc_allocator.cc:1039]   Size: 410.08MiB | Requested Size: 105.20MiB | in_use: 0 | bin_num: 20, prev:   Size: 105.20MiB | Requested Size: 105.20MiB | in_use: 1 | bin_num: -1, next:   Size: 9.25MiB | Requested Size: 9.25MiB | in_use: 1 | bin_num: -1\n",
      "2023-04-13 13:08:19.969857: I tensorflow/core/common_runtime/bfc_allocator.cc:1046] Next region of size 14458814464\n",
      "2023-04-13 13:08:19.969877: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272000000 of size 1280 next 1\n",
      "2023-04-13 13:08:19.969882: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272000500 of size 256 next 2\n",
      "2023-04-13 13:08:19.969886: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272000600 of size 256 next 3\n",
      "2023-04-13 13:08:19.969890: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272000700 of size 256 next 4\n",
      "2023-04-13 13:08:19.969894: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272000800 of size 256 next 5\n",
      "2023-04-13 13:08:19.969898: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272000900 of size 256 next 6\n",
      "2023-04-13 13:08:19.969902: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272000a00 of size 256 next 7\n",
      "2023-04-13 13:08:19.969906: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272000b00 of size 256 next 10\n",
      "2023-04-13 13:08:19.969909: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272000c00 of size 256 next 18\n",
      "2023-04-13 13:08:19.969913: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272000d00 of size 256 next 19\n",
      "2023-04-13 13:08:19.969924: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272000e00 of size 256 next 21\n",
      "2023-04-13 13:08:19.969928: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272000f00 of size 256 next 25\n",
      "2023-04-13 13:08:19.969937: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272001000 of size 256 next 26\n",
      "2023-04-13 13:08:19.969941: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272001100 of size 256 next 23\n",
      "2023-04-13 13:08:19.969945: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272001200 of size 256 next 34\n",
      "2023-04-13 13:08:19.969949: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272001300 of size 256 next 36\n",
      "2023-04-13 13:08:19.969954: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272001400 of size 256 next 47\n",
      "2023-04-13 13:08:19.969995: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272001500 of size 256 next 56\n",
      "2023-04-13 13:08:19.970000: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272001600 of size 256 next 53\n",
      "2023-04-13 13:08:19.970004: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272001700 of size 256 next 54\n",
      "2023-04-13 13:08:19.970008: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272001800 of size 256 next 62\n",
      "2023-04-13 13:08:19.970017: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272001900 of size 256 next 63\n",
      "2023-04-13 13:08:19.970022: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272001a00 of size 768 next 64\n",
      "2023-04-13 13:08:19.970026: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272001d00 of size 256 next 67\n",
      "2023-04-13 13:08:19.970079: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272001e00 of size 256 next 68\n",
      "2023-04-13 13:08:19.970085: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272001f00 of size 768 next 69\n",
      "2023-04-13 13:08:19.970095: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272002200 of size 768 next 72\n",
      "2023-04-13 13:08:19.970099: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272002500 of size 768 next 74\n",
      "2023-04-13 13:08:19.970103: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272002800 of size 768 next 76\n",
      "2023-04-13 13:08:19.970107: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272002b00 of size 1280 next 8\n",
      "2023-04-13 13:08:19.970116: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272003000 of size 4864 next 9\n",
      "2023-04-13 13:08:19.970125: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272004300 of size 256 next 57\n",
      "2023-04-13 13:08:19.970129: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272004400 of size 256 next 58\n",
      "2023-04-13 13:08:19.970133: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272004500 of size 256 next 59\n",
      "2023-04-13 13:08:19.970165: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272004600 of size 256 next 60\n",
      "2023-04-13 13:08:19.970170: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272004700 of size 3840 next 52\n",
      "2023-04-13 13:08:19.970179: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272005600 of size 4864 next 55\n",
      "2023-04-13 13:08:19.970184: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272006900 of size 485120 next 1041\n",
      "2023-04-13 13:08:19.970211: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f327207d000 of size 606976 next 66\n",
      "2023-04-13 13:08:19.970217: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272111300 of size 546048 next 65\n",
      "2023-04-13 13:08:19.970226: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272196800 of size 2560 next 458\n",
      "2023-04-13 13:08:19.970230: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272197200 of size 256 next 657\n",
      "2023-04-13 13:08:19.970257: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272197300 of size 256 next 109\n",
      "2023-04-13 13:08:19.970262: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272197400 of size 512 next 100\n",
      "2023-04-13 13:08:19.970271: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272197600 of size 512 next 661\n",
      "2023-04-13 13:08:19.970275: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272197800 of size 768 next 104\n",
      "2023-04-13 13:08:19.970314: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272197b00 of size 1280 next 444\n",
      "2023-04-13 13:08:19.970319: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272198000 of size 6144 next 484\n",
      "2023-04-13 13:08:19.970328: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272199800 of size 167936 next 71\n",
      "2023-04-13 13:08:19.970333: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32721c2800 of size 90112 next 70\n",
      "2023-04-13 13:08:19.970360: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32721d8800 of size 90112 next 73\n",
      "2023-04-13 13:08:19.970366: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32721ee800 of size 90112 next 75\n",
      "2023-04-13 13:08:19.970374: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272204800 of size 90112 next 77\n",
      "2023-04-13 13:08:19.970378: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f327221a800 of size 90112 next 79\n",
      "2023-04-13 13:08:19.970382: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272230800 of size 256 next 78\n",
      "2023-04-13 13:08:19.970410: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272230900 of size 256 next 80\n",
      "2023-04-13 13:08:19.970414: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272230a00 of size 256 next 81\n",
      "2023-04-13 13:08:19.970423: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272230b00 of size 768 next 478\n",
      "2023-04-13 13:08:19.970427: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272230e00 of size 768 next 469\n",
      "2023-04-13 13:08:19.970454: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272231100 of size 1024 next 453\n",
      "2023-04-13 13:08:19.970459: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272231500 of size 256 next 486\n",
      "2023-04-13 13:08:19.970468: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272231600 of size 256 next 488\n",
      "2023-04-13 13:08:19.970472: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272231700 of size 256 next 492\n",
      "2023-04-13 13:08:19.970500: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272231800 of size 256 next 481\n",
      "2023-04-13 13:08:19.970509: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272231900 of size 512 next 482\n",
      "2023-04-13 13:08:19.970514: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272231b00 of size 512 next 485\n",
      "2023-04-13 13:08:19.970518: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272231d00 of size 256 next 480\n",
      "2023-04-13 13:08:19.970546: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272231e00 of size 256 next 83\n",
      "2023-04-13 13:08:19.970555: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272231f00 of size 2560 next 84\n",
      "2023-04-13 13:08:19.970559: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272232900 of size 256 next 212\n",
      "2023-04-13 13:08:19.970563: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272232a00 of size 256 next 158\n",
      "2023-04-13 13:08:19.970593: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272232b00 of size 256 next 372\n",
      "2023-04-13 13:08:19.970602: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272232c00 of size 256 next 127\n",
      "2023-04-13 13:08:19.970606: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272232d00 of size 256 next 129\n",
      "2023-04-13 13:08:19.970632: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272232e00 of size 6912 next 213\n",
      "2023-04-13 13:08:19.970638: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272234900 of size 256 next 226\n",
      "2023-04-13 13:08:19.970647: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272234a00 of size 256 next 228\n",
      "2023-04-13 13:08:19.970651: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272234b00 of size 4096 next 489\n",
      "2023-04-13 13:08:19.970679: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272235b00 of size 768 next 494\n",
      "2023-04-13 13:08:19.970684: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272235e00 of size 768 next 499\n",
      "2023-04-13 13:08:19.970693: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272236100 of size 768 next 500\n",
      "2023-04-13 13:08:19.970697: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272236400 of size 768 next 502\n",
      "2023-04-13 13:08:19.970724: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272236700 of size 768 next 505\n",
      "2023-04-13 13:08:19.970734: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272236a00 of size 768 next 506\n",
      "2023-04-13 13:08:19.970738: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272236d00 of size 256 next 531\n",
      "2023-04-13 13:08:19.970742: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272236e00 of size 256 next 524\n",
      "2023-04-13 13:08:19.970768: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272236f00 of size 512 next 528\n",
      "2023-04-13 13:08:19.970773: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272237100 of size 512 next 529\n",
      "2023-04-13 13:08:19.970782: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272237300 of size 768 next 465\n",
      "2023-04-13 13:08:19.970786: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272237600 of size 768 next 460\n",
      "2023-04-13 13:08:19.970814: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272237900 of size 768 next 466\n",
      "2023-04-13 13:08:19.970818: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272237c00 of size 768 next 556\n",
      "2023-04-13 13:08:19.970828: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272237f00 of size 256 next 371\n",
      "2023-04-13 13:08:19.970832: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272238000 of size 256 next 378\n",
      "2023-04-13 13:08:19.970861: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272238100 of size 256 next 317\n",
      "2023-04-13 13:08:19.970876: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272238200 of size 256 next 380\n",
      "2023-04-13 13:08:19.970880: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272238300 of size 256 next 381\n",
      "2023-04-13 13:08:19.970884: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272238400 of size 256 next 327\n",
      "2023-04-13 13:08:19.970927: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272238500 of size 30464 next 290\n",
      "2023-04-13 13:08:19.970957: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f327223fc00 of size 30464 next 865\n",
      "2023-04-13 13:08:19.970967: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272247300 of size 40704 next 11\n",
      "2023-04-13 13:08:19.970972: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272251200 of size 2412288 next 12\n",
      "2023-04-13 13:08:19.970976: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f327249e100 of size 2412288 next 13\n",
      "2023-04-13 13:08:19.971003: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32726eb000 of size 2412288 next 14\n",
      "2023-04-13 13:08:19.971012: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272937f00 of size 2412288 next 15\n",
      "2023-04-13 13:08:19.971016: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3272b84e00 of size 7236608 next 16\n",
      "2023-04-13 13:08:19.971020: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f327326ba00 of size 9648640 next 17\n",
      "2023-04-13 13:08:19.971049: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3273b9f400 of size 9648640 next 20\n",
      "2023-04-13 13:08:19.971077: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32744d2e00 of size 9648640 next 24\n",
      "2023-04-13 13:08:19.971081: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3274e06800 of size 9648640 next 22\n",
      "2023-04-13 13:08:19.971103: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f327573a200 of size 9648640 next 32\n",
      "2023-04-13 13:08:19.971108: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f327606dc00 of size 9648640 next 35\n",
      "2023-04-13 13:08:19.971116: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32769a1600 of size 614400 next 495\n",
      "2023-04-13 13:08:19.971120: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276a37600 of size 1204224 next 885\n",
      "2023-04-13 13:08:19.971123: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b5d600 of size 30464 next 388\n",
      "2023-04-13 13:08:19.971148: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b64d00 of size 30464 next 156\n",
      "2023-04-13 13:08:19.971156: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b6c400 of size 30464 next 412\n",
      "2023-04-13 13:08:19.971160: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b73b00 of size 256 next 508\n",
      "2023-04-13 13:08:19.971163: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b73c00 of size 256 next 111\n",
      "2023-04-13 13:08:19.971187: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b73d00 of size 256 next 602\n",
      "2023-04-13 13:08:19.971191: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b73e00 of size 512 next 603\n",
      "2023-04-13 13:08:19.971199: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b74000 of size 512 next 110\n",
      "2023-04-13 13:08:19.971202: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b74200 of size 512 next 514\n",
      "2023-04-13 13:08:19.971226: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b74400 of size 512 next 509\n",
      "2023-04-13 13:08:19.971230: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b74600 of size 2560 next 510\n",
      "2023-04-13 13:08:19.971239: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b75000 of size 1792 next 980\n",
      "2023-04-13 13:08:19.971243: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b75700 of size 256 next 825\n",
      "2023-04-13 13:08:19.971267: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b75800 of size 256 next 671\n",
      "2023-04-13 13:08:19.971271: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b75900 of size 256 next 649\n",
      "2023-04-13 13:08:19.971279: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b75a00 of size 512 next 549\n",
      "2023-04-13 13:08:19.971282: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b75c00 of size 256 next 698\n",
      "2023-04-13 13:08:19.971306: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b75d00 of size 256 next 812\n",
      "2023-04-13 13:08:19.971311: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b75e00 of size 256 next 553\n",
      "2023-04-13 13:08:19.971318: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b75f00 of size 256 next 580\n",
      "2023-04-13 13:08:19.971322: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b76000 of size 512 next 587\n",
      "2023-04-13 13:08:19.971346: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b76200 of size 256 next 637\n",
      "2023-04-13 13:08:19.971350: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f3276b76300 of size 256 next 608\n",
      "2023-04-13 13:08:19.971358: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b76400 of size 512 next 132\n",
      "2023-04-13 13:08:19.971361: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b76600 of size 512 next 149\n",
      "2023-04-13 13:08:19.971385: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b76800 of size 2560 next 619\n",
      "2023-04-13 13:08:19.971390: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b77200 of size 768 next 615\n",
      "2023-04-13 13:08:19.971398: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b77500 of size 768 next 762\n",
      "2023-04-13 13:08:19.971401: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b77800 of size 768 next 616\n",
      "2023-04-13 13:08:19.971424: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b77b00 of size 768 next 614\n",
      "2023-04-13 13:08:19.971429: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b77e00 of size 768 next 611\n",
      "2023-04-13 13:08:19.971437: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b78100 of size 768 next 613\n",
      "2023-04-13 13:08:19.971440: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b78400 of size 4864 next 530\n",
      "2023-04-13 13:08:19.971464: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b79700 of size 5120 next 767\n",
      "2023-04-13 13:08:19.971469: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7ab00 of size 256 next 768\n",
      "2023-04-13 13:08:19.971477: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7ac00 of size 256 next 618\n",
      "2023-04-13 13:08:19.971480: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7ad00 of size 512 next 778\n",
      "2023-04-13 13:08:19.971484: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7af00 of size 256 next 780\n",
      "2023-04-13 13:08:19.971509: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7b000 of size 4608 next 658\n",
      "2023-04-13 13:08:19.971512: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7c200 of size 1536 next 106\n",
      "2023-04-13 13:08:19.971521: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7c800 of size 256 next 583\n",
      "2023-04-13 13:08:19.971524: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f3276b7c900 of size 256 next 883\n",
      "2023-04-13 13:08:19.971548: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7ca00 of size 256 next 645\n",
      "2023-04-13 13:08:19.971553: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7cb00 of size 256 next 997\n",
      "2023-04-13 13:08:19.971560: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7cc00 of size 512 next 656\n",
      "2023-04-13 13:08:19.971564: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7ce00 of size 256 next 803\n",
      "2023-04-13 13:08:19.971567: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7cf00 of size 512 next 639\n",
      "2023-04-13 13:08:19.971570: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7d100 of size 256 next 607\n",
      "2023-04-13 13:08:19.971574: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7d200 of size 512 next 774\n",
      "2023-04-13 13:08:19.971598: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7d400 of size 256 next 878\n",
      "2023-04-13 13:08:19.971603: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7d500 of size 256 next 542\n",
      "2023-04-13 13:08:19.971610: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7d600 of size 256 next 761\n",
      "2023-04-13 13:08:19.971614: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7d700 of size 256 next 672\n",
      "2023-04-13 13:08:19.971617: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7d800 of size 512 next 930\n",
      "2023-04-13 13:08:19.971642: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7da00 of size 256 next 924\n",
      "2023-04-13 13:08:19.971650: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7db00 of size 256 next 928\n",
      "2023-04-13 13:08:19.971653: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7dc00 of size 512 next 714\n",
      "2023-04-13 13:08:19.971657: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7de00 of size 256 next 952\n",
      "2023-04-13 13:08:19.971681: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7df00 of size 512 next 857\n",
      "2023-04-13 13:08:19.971685: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7e100 of size 256 next 906\n",
      "2023-04-13 13:08:19.971693: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7e200 of size 256 next 907\n",
      "2023-04-13 13:08:19.971696: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7e300 of size 256 next 644\n",
      "2023-04-13 13:08:19.971720: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7e400 of size 512 next 920\n",
      "2023-04-13 13:08:19.971724: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7e600 of size 256 next 921\n",
      "2023-04-13 13:08:19.971732: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7e700 of size 256 next 877\n",
      "2023-04-13 13:08:19.971736: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7e800 of size 512 next 829\n",
      "2023-04-13 13:08:19.971760: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7ea00 of size 256 next 666\n",
      "2023-04-13 13:08:19.971764: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7eb00 of size 256 next 582\n",
      "2023-04-13 13:08:19.971771: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7ec00 of size 256 next 914\n",
      "2023-04-13 13:08:19.971775: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7ed00 of size 512 next 862\n",
      "2023-04-13 13:08:19.971799: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7ef00 of size 256 next 776\n",
      "2023-04-13 13:08:19.971803: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7f000 of size 256 next 101\n",
      "2023-04-13 13:08:19.971811: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7f100 of size 256 next 822\n",
      "2023-04-13 13:08:19.971815: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7f200 of size 512 next 949\n",
      "2023-04-13 13:08:19.971838: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7f400 of size 256 next 723\n",
      "2023-04-13 13:08:19.971842: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7f500 of size 256 next 792\n",
      "2023-04-13 13:08:19.971850: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7f600 of size 512 next 853\n",
      "2023-04-13 13:08:19.971853: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7f800 of size 256 next 856\n",
      "2023-04-13 13:08:19.971882: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7f900 of size 256 next 89\n",
      "2023-04-13 13:08:19.971891: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7fa00 of size 512 next 868\n",
      "2023-04-13 13:08:19.971899: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7fc00 of size 256 next 870\n",
      "2023-04-13 13:08:19.971903: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b7fd00 of size 4864 next 575\n",
      "2023-04-13 13:08:19.971906: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b81000 of size 768 next 564\n",
      "2023-04-13 13:08:19.971929: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b81300 of size 768 next 579\n",
      "2023-04-13 13:08:19.971933: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b81600 of size 768 next 470\n",
      "2023-04-13 13:08:19.971941: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b81900 of size 768 next 461\n",
      "2023-04-13 13:08:19.971945: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b81c00 of size 2560 next 456\n",
      "2023-04-13 13:08:19.971969: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b82600 of size 512 next 451\n",
      "2023-04-13 13:08:19.971974: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b82800 of size 512 next 449\n",
      "2023-04-13 13:08:19.971981: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b82a00 of size 4864 next 473\n",
      "2023-04-13 13:08:19.971985: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b83d00 of size 3840 next 408\n",
      "2023-04-13 13:08:19.972021: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b84c00 of size 768 next 447\n",
      "2023-04-13 13:08:19.972026: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b84f00 of size 768 next 445\n",
      "2023-04-13 13:08:19.972033: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b85200 of size 768 next 383\n",
      "2023-04-13 13:08:19.972037: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b85500 of size 768 next 532\n",
      "2023-04-13 13:08:19.972040: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b85800 of size 768 next 915\n",
      "2023-04-13 13:08:19.972044: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b85b00 of size 768 next 88\n",
      "2023-04-13 13:08:19.972047: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b85e00 of size 2560 next 97\n",
      "2023-04-13 13:08:19.972050: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b86800 of size 256 next 348\n",
      "2023-04-13 13:08:19.972075: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b86900 of size 2560 next 570\n",
      "2023-04-13 13:08:19.972079: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b87300 of size 768 next 98\n",
      "2023-04-13 13:08:19.972086: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b87600 of size 768 next 697\n",
      "2023-04-13 13:08:19.972090: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b87900 of size 768 next 886\n",
      "2023-04-13 13:08:19.972132: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b87c00 of size 1280 next 785\n",
      "2023-04-13 13:08:19.972137: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b88100 of size 256 next 561\n",
      "2023-04-13 13:08:19.972146: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f3276b88200 of size 512 next 925\n",
      "2023-04-13 13:08:19.972150: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b88400 of size 256 next 926\n",
      "2023-04-13 13:08:19.972176: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b88500 of size 1792 next 974\n",
      "2023-04-13 13:08:19.972182: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b88c00 of size 256 next 892\n",
      "2023-04-13 13:08:19.972191: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b88d00 of size 256 next 893\n",
      "2023-04-13 13:08:19.972195: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b88e00 of size 256 next 891\n",
      "2023-04-13 13:08:19.972199: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f3276b88f00 of size 512 next 922\n",
      "2023-04-13 13:08:19.972226: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b89100 of size 256 next 923\n",
      "2023-04-13 13:08:19.972235: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b89200 of size 256 next 647\n",
      "2023-04-13 13:08:19.972240: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f3276b89300 of size 512 next 799\n",
      "2023-04-13 13:08:19.972266: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b89500 of size 256 next 940\n",
      "2023-04-13 13:08:19.972272: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b89600 of size 256 next 581\n",
      "2023-04-13 13:08:19.972281: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f3276b89700 of size 512 next 971\n",
      "2023-04-13 13:08:19.972285: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b89900 of size 256 next 972\n",
      "2023-04-13 13:08:19.972310: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b89a00 of size 256 next 498\n",
      "2023-04-13 13:08:19.972316: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b89b00 of size 90112 next 496\n",
      "2023-04-13 13:08:19.972325: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276b9fb00 of size 92416 next 917\n",
      "2023-04-13 13:08:19.972329: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276bb6400 of size 121344 next 296\n",
      "2023-04-13 13:08:19.972356: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276bd3e00 of size 1050624 next 507\n",
      "2023-04-13 13:08:19.972362: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276cd4600 of size 119808 next 830\n",
      "2023-04-13 13:08:19.972372: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276cf1a00 of size 29952 next 136\n",
      "2023-04-13 13:08:19.972376: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276cf8f00 of size 29952 next 864\n",
      "2023-04-13 13:08:19.972404: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276d00400 of size 29952 next 243\n",
      "2023-04-13 13:08:19.972408: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276d07900 of size 29952 next 292\n",
      "2023-04-13 13:08:19.972417: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276d0ee00 of size 56832 next 310\n",
      "2023-04-13 13:08:19.972422: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276d1cc00 of size 30464 next 964\n",
      "2023-04-13 13:08:19.972449: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276d24300 of size 30720 next 376\n",
      "2023-04-13 13:08:19.972453: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276d2bb00 of size 90112 next 501\n",
      "2023-04-13 13:08:19.972462: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276d41b00 of size 90112 next 504\n",
      "2023-04-13 13:08:19.972466: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276d57b00 of size 124160 next 880\n",
      "2023-04-13 13:08:19.972494: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276d76000 of size 242944 next 293\n",
      "2023-04-13 13:08:19.972499: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276db1500 of size 2395648 next 999\n",
      "2023-04-13 13:08:19.972508: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3276ffa300 of size 606208 next 258\n",
      "2023-04-13 13:08:19.972512: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f327708e300 of size 2387200 next 33\n",
      "2023-04-13 13:08:19.972539: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32772d5000 of size 9648640 next 38\n",
      "2023-04-13 13:08:19.972549: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3277c08a00 of size 12061184 next 27\n",
      "2023-04-13 13:08:19.972554: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3278789400 of size 154375680 next 28\n",
      "2023-04-13 13:08:19.972558: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3281ac2a00 of size 38594048 next 31\n",
      "2023-04-13 13:08:19.972585: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3283f91000 of size 9648640 next 29\n",
      "2023-04-13 13:08:19.972594: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32848c4a00 of size 28945920 next 37\n",
      "2023-04-13 13:08:19.972599: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f328645f800 of size 115781888 next 30\n",
      "2023-04-13 13:08:19.972603: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f328d2ca900 of size 9648640 next 39\n",
      "2023-04-13 13:08:19.972633: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f328dbfe300 of size 9648640 next 40\n",
      "2023-04-13 13:08:19.972643: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f328e531d00 of size 28945664 next 41\n",
      "2023-04-13 13:08:19.972647: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32900cca00 of size 28945664 next 42\n",
      "2023-04-13 13:08:19.972651: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3291c67700 of size 91660800 next 44\n",
      "2023-04-13 13:08:19.972678: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32973d1900 of size 9648640 next 46\n",
      "2023-04-13 13:08:19.972683: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3297d05300 of size 3312640 next 61\n",
      "2023-04-13 13:08:19.972693: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f329802df00 of size 7274496 next 1025\n",
      "2023-04-13 13:08:19.972697: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f329871df00 of size 606208 next 266\n",
      "2023-04-13 13:08:19.972724: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32987b1f00 of size 606208 next 224\n",
      "2023-04-13 13:08:19.972729: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3298845f00 of size 606208 next 374\n",
      "2023-04-13 13:08:19.972738: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32988d9f00 of size 606208 next 391\n",
      "2023-04-13 13:08:19.972742: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f329896df00 of size 606208 next 439\n",
      "2023-04-13 13:08:19.972771: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3298a01f00 of size 848128 next 927\n",
      "2023-04-13 13:08:19.972776: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3298ad1000 of size 969216 next 92\n",
      "2023-04-13 13:08:19.972785: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3298bbda00 of size 2444288 next 889\n",
      "2023-04-13 13:08:19.972790: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3298e12600 of size 30720 next 124\n",
      "2023-04-13 13:08:19.972816: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3298e19e00 of size 30720 next 873\n",
      "2023-04-13 13:08:19.972821: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3298e21600 of size 60928 next 808\n",
      "2023-04-13 13:08:19.972830: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3298e30400 of size 122368 next 957\n",
      "2023-04-13 13:08:19.972834: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3298e4e200 of size 10142976 next 828\n",
      "2023-04-13 13:08:19.972860: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32997fa700 of size 2424576 next 419\n",
      "2023-04-13 13:08:19.972871: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3299a4a600 of size 2424576 next 402\n",
      "2023-04-13 13:08:19.972880: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3299c9a500 of size 2424576 next 343\n",
      "2023-04-13 13:08:19.972885: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3299eea400 of size 2503168 next 390\n",
      "2023-04-13 13:08:19.972914: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f329a14d600 of size 53615872 next 43\n",
      "2023-04-13 13:08:19.972920: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f329d46f300 of size 91660800 next 45\n",
      "2023-04-13 13:08:19.972929: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a2bd9500 of size 90112 next 759\n",
      "2023-04-13 13:08:19.972933: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a2bef500 of size 90112 next 617\n",
      "2023-04-13 13:08:19.972937: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a2c05500 of size 90112 next 758\n",
      "2023-04-13 13:08:19.972964: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a2c1b500 of size 90112 next 741\n",
      "2023-04-13 13:08:19.972973: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a2c31500 of size 97536 next 116\n",
      "2023-04-13 13:08:19.972978: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a2c49200 of size 256 next 112\n",
      "2023-04-13 13:08:19.972982: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a2c49300 of size 256 next 113\n",
      "2023-04-13 13:08:19.973010: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a2c49400 of size 256 next 115\n",
      "2023-04-13 13:08:19.973019: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a2c49500 of size 256 next 117\n",
      "2023-04-13 13:08:19.973023: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a2c49600 of size 256 next 114\n",
      "2023-04-13 13:08:19.973027: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a2c49700 of size 256 next 118\n",
      "2023-04-13 13:08:19.973054: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a2c49800 of size 256 next 121\n",
      "2023-04-13 13:08:19.973058: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a2c49900 of size 2424576 next 450\n",
      "2023-04-13 13:08:19.973068: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a2e99800 of size 9796608 next 937\n",
      "2023-04-13 13:08:19.973072: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a37f1400 of size 611072 next 201\n",
      "2023-04-13 13:08:19.973101: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a3886700 of size 2424576 next 1038\n",
      "2023-04-13 13:08:19.973106: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a3ad6600 of size 4619264 next 85\n",
      "2023-04-13 13:08:19.973115: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a3f3e200 of size 9698048 next 894\n",
      "2023-04-13 13:08:19.973119: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a487dd00 of size 7273472 next 884\n",
      "2023-04-13 13:08:19.973146: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a4f6d900 of size 2424576 next 440\n",
      "2023-04-13 13:08:19.973150: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a51bd800 of size 7273472 next 413\n",
      "2023-04-13 13:08:19.973159: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a58ad400 of size 2424576 next 1033\n",
      "2023-04-13 13:08:19.973163: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a5afd300 of size 16971520 next 816\n",
      "2023-04-13 13:08:19.973190: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f32a6b2ca00 of size 17870080 next 384\n",
      "2023-04-13 13:08:19.973194: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a7c37700 of size 2424576 next 360\n",
      "2023-04-13 13:08:19.973203: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a7e87600 of size 2424576 next 351\n",
      "2023-04-13 13:08:19.973208: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a80d7500 of size 4848896 next 234\n",
      "2023-04-13 13:08:19.973235: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32a8577200 of size 38791680 next 425\n",
      "2023-04-13 13:08:19.973240: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32aaa75c00 of size 29093888 next 441\n",
      "2023-04-13 13:08:19.973249: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32ac634c00 of size 31114496 next 50\n",
      "2023-04-13 13:08:19.973253: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32ae3e1100 of size 91660800 next 51\n",
      "2023-04-13 13:08:19.973257: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32b3b4b300 of size 323224320 next 48\n",
      "2023-04-13 13:08:19.973284: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32c6f8b600 of size 439005952 next 49\n",
      "2023-04-13 13:08:19.973294: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32e1236900 of size 3030784 next 632\n",
      "2023-04-13 13:08:19.973298: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32e151a800 of size 9091840 next 633\n",
      "2023-04-13 13:08:19.973335: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32e1dc6300 of size 24059392 next 119\n",
      "2023-04-13 13:08:19.973341: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32e34b8100 of size 9648640 next 601\n",
      "2023-04-13 13:08:19.973350: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32e3debb00 of size 9648640 next 604\n",
      "2023-04-13 13:08:19.973355: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32e471f500 of size 28945664 next 606\n",
      "2023-04-13 13:08:19.973381: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32e62ba200 of size 28945664 next 589\n",
      "2023-04-13 13:08:19.973386: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32e7e54f00 of size 91660800 next 522\n",
      "2023-04-13 13:08:19.973395: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32ed5bf100 of size 9648640 next 86\n",
      "2023-04-13 13:08:19.973399: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f32edef2b00 of size 16971520 next 913\n",
      "2023-04-13 13:08:19.973424: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32eef22200 of size 9698048 next 979\n",
      "2023-04-13 13:08:19.973430: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f32ef861d00 of size 11924480 next 105\n",
      "2023-04-13 13:08:19.973439: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32f03c1100 of size 53066752 next 826\n",
      "2023-04-13 13:08:19.973443: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f32f365cd00 of size 110313984 next 772\n",
      "2023-04-13 13:08:19.973470: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32f9f90f00 of size 9648640 next 355\n",
      "2023-04-13 13:08:19.973476: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32fa8c4900 of size 9648640 next 782\n",
      "2023-04-13 13:08:19.973484: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32fb1f8300 of size 9648640 next 568\n",
      "2023-04-13 13:08:19.973488: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32fbb2bd00 of size 9648640 next 687\n",
      "2023-04-13 13:08:19.973519: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32fc45f700 of size 9648640 next 809\n",
      "2023-04-13 13:08:19.973525: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32fcd93100 of size 9648640 next 547\n",
      "2023-04-13 13:08:19.973534: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f32fd6c6b00 of size 192969728 next 578\n",
      "2023-04-13 13:08:19.973538: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3308ece700 of size 91660800 next 662\n",
      "2023-04-13 13:08:19.973564: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f330e638900 of size 110313984 next 596\n",
      "2023-04-13 13:08:19.973570: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f3314f6cb00 of size 430001664 next 709\n",
      "2023-04-13 13:08:19.973579: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f332e981900 of size 9698048 next 948\n",
      "2023-04-13 13:08:19.973583: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f332f2c1400 of size 80500736 next 958\n",
      "2023-04-13 13:08:19.973609: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3333f86c00 of size 1929728 next 794\n",
      "2023-04-13 13:08:19.973615: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f333415de00 of size 2424576 next 888\n",
      "2023-04-13 13:08:19.973623: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33343add00 of size 7768576 next 837\n",
      "2023-04-13 13:08:19.973628: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3334b16700 of size 288380928 next 552\n",
      "2023-04-13 13:08:19.973659: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f3345e1bf00 of size 77583360 next 800\n",
      "2023-04-13 13:08:19.973664: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f334a819300 of size 38791680 next 636\n",
      "2023-04-13 13:08:19.973674: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f334cd17d00 of size 18184704 next 1006\n",
      "2023-04-13 13:08:19.973678: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f334de6f700 of size 29093888 next 836\n",
      "2023-04-13 13:08:19.973716: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f334fa2e700 of size 2424576 next 866\n",
      "2023-04-13 13:08:19.973721: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f334fc7e600 of size 7273472 next 731\n",
      "2023-04-13 13:08:19.973753: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f335036e200 of size 9698048 next 813\n",
      "2023-04-13 13:08:19.973763: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3350cadd00 of size 2424576 next 597\n",
      "2023-04-13 13:08:19.973767: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3350efdc00 of size 2424576 next 905\n",
      "2023-04-13 13:08:19.973771: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f335114db00 of size 7273728 next 879\n",
      "2023-04-13 13:08:19.973798: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f335183d800 of size 2424576 next 742\n",
      "2023-04-13 13:08:19.973807: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3351a8d700 of size 2424576 next 546\n",
      "2023-04-13 13:08:19.973811: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f3351cdd600 of size 7273728 next 668\n",
      "2023-04-13 13:08:19.973815: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33523cd300 of size 2424576 next 784\n",
      "2023-04-13 13:08:19.973847: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f335261d200 of size 2424576 next 835\n",
      "2023-04-13 13:08:19.973857: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f335286d100 of size 12122880 next 540\n",
      "2023-04-13 13:08:19.973861: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33533fcc00 of size 2424576 next 651\n",
      "2023-04-13 13:08:19.973870: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f335364cb00 of size 2424576 next 804\n",
      "2023-04-13 13:08:19.973897: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f335389ca00 of size 7273728 next 840\n",
      "2023-04-13 13:08:19.973902: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3353f8c700 of size 2424576 next 872\n",
      "2023-04-13 13:08:19.973911: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33541dc600 of size 7273472 next 911\n",
      "2023-04-13 13:08:19.973915: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33548cc200 of size 7273472 next 899\n",
      "2023-04-13 13:08:19.973948: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f3354fbbe00 of size 48489728 next 729\n",
      "2023-04-13 13:08:19.973953: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3357dfa300 of size 23032576 next 686\n",
      "2023-04-13 13:08:19.973962: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f33593f1600 of size 281240064 next 749\n",
      "2023-04-13 13:08:19.973966: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f336a027800 of size 23032576 next 563\n",
      "2023-04-13 13:08:19.973996: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f336b61eb00 of size 2424576 next 876\n",
      "2023-04-13 13:08:19.974001: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f336b86ea00 of size 2424576 next 738\n",
      "2023-04-13 13:08:19.974010: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f336babe900 of size 58188032 next 796\n",
      "2023-04-13 13:08:19.974015: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f336f23ca00 of size 23032576 next 904\n",
      "2023-04-13 13:08:19.974042: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3370833d00 of size 81220096 next 751\n",
      "2023-04-13 13:08:19.974052: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f33755a8f00 of size 110557440 next 479\n",
      "2023-04-13 13:08:19.974056: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f337bf18800 of size 485120 next 726\n",
      "2023-04-13 13:08:19.974060: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f337bf8ef00 of size 439005952 next 908\n",
      "2023-04-13 13:08:19.974088: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f339623a200 of size 386721024 next 143\n",
      "2023-04-13 13:08:19.974098: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33ad308700 of size 1929728 next 931\n",
      "2023-04-13 13:08:19.974102: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f33ad4df900 of size 19297280 next 854\n",
      "2023-04-13 13:08:19.974132: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33ae746d00 of size 41005824 next 467\n",
      "2023-04-13 13:08:19.974137: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33b0e62000 of size 3030784 next 190\n",
      "2023-04-13 13:08:19.974146: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33b1145f00 of size 3030784 next 141\n",
      "2023-04-13 13:08:19.974151: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33b1429e00 of size 5999104 next 463\n",
      "2023-04-13 13:08:19.974177: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33b19e2800 of size 192969728 next 685\n",
      "2023-04-13 13:08:19.974182: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33bd1ea400 of size 192969728 next 851\n",
      "2023-04-13 13:08:19.974191: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33c89f2000 of size 91660800 next 843\n",
      "2023-04-13 13:08:19.974196: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33ce15c200 of size 294279936 next 475\n",
      "2023-04-13 13:08:19.974222: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33dfa01d00 of size 5427200 next 490\n",
      "2023-04-13 13:08:19.974227: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33dff2ed00 of size 606208 next 624\n",
      "2023-04-13 13:08:19.974236: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33dffc2d00 of size 606208 next 625\n",
      "2023-04-13 13:08:19.974240: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33e0056d00 of size 7273472 next 623\n",
      "2023-04-13 13:08:19.974267: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33e0746900 of size 3030784 next 626\n",
      "2023-04-13 13:08:19.974272: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33e0a2a800 of size 3030784 next 627\n",
      "2023-04-13 13:08:19.974324: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33e0d0e700 of size 3030784 next 628\n",
      "2023-04-13 13:08:19.974329: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33e0ff2600 of size 3030784 next 630\n",
      "2023-04-13 13:08:19.974339: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33e12d6500 of size 9091840 next 631\n",
      "2023-04-13 13:08:19.974365: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33e1b82000 of size 12122624 next 629\n",
      "2023-04-13 13:08:19.974371: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33e2711a00 of size 38594048 next 144\n",
      "2023-04-13 13:08:19.974380: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33e4be0000 of size 9648640 next 555\n",
      "2023-04-13 13:08:19.974384: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33e5513a00 of size 9648640 next 521\n",
      "2023-04-13 13:08:19.974388: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33e5e47400 of size 9648640 next 566\n",
      "2023-04-13 13:08:19.974421: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33e677ae00 of size 9648640 next 120\n",
      "2023-04-13 13:08:19.974430: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33e70ae800 of size 9648640 next 567\n",
      "2023-04-13 13:08:19.974435: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33e79e2200 of size 10639104 next 474\n",
      "2023-04-13 13:08:19.974468: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33e8407900 of size 285925376 next 311\n",
      "2023-04-13 13:08:19.974483: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33f94b5900 of size 23032576 next 230\n",
      "2023-04-13 13:08:19.974523: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33faaacc00 of size 16971520 next 752\n",
      "2023-04-13 13:08:19.974535: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33fbadc300 of size 7273472 next 138\n",
      "2023-04-13 13:08:19.974539: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33fc1cbf00 of size 9698048 next 638\n",
      "2023-04-13 13:08:19.974568: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f33fcb0ba00 of size 14546688 next 442\n",
      "2023-04-13 13:08:19.974577: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33fd8eb100 of size 23032576 next 277\n",
      "2023-04-13 13:08:19.974582: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f33feee2400 of size 81220096 next 875\n",
      "2023-04-13 13:08:19.974608: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3403c57600 of size 110313984 next 257\n",
      "2023-04-13 13:08:19.974613: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f340a58b800 of size 91660800 next 544\n",
      "2023-04-13 13:08:19.974622: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f340fcf5a00 of size 91660800 next 728\n",
      "2023-04-13 13:08:19.974627: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f341545fc00 of size 91660800 next 817\n",
      "2023-04-13 13:08:19.974660: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f341abc9e00 of size 91660800 next 882\n",
      "2023-04-13 13:08:19.974666: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3420334000 of size 91660800 next 842\n",
      "2023-04-13 13:08:19.974675: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3425a9e200 of size 91660800 next 890\n",
      "2023-04-13 13:08:19.974679: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f342b208400 of size 91660800 next 586\n",
      "2023-04-13 13:08:19.974707: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3430972600 of size 145764352 next 710\n",
      "2023-04-13 13:08:19.974716: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f3439475600 of size 18184704 next 777\n",
      "2023-04-13 13:08:19.974721: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f343a5cd000 of size 38791936 next 975\n",
      "2023-04-13 13:08:19.974747: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f343cacbb00 of size 2424576 next 732\n",
      "2023-04-13 13:08:19.974753: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f343cd1ba00 of size 2424576 next 724\n",
      "2023-04-13 13:08:19.974761: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f343cf6b900 of size 4848896 next 895\n",
      "2023-04-13 13:08:19.974766: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f343d40b600 of size 4849152 next 664\n",
      "2023-04-13 13:08:19.974792: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f343d8ab400 of size 2412288 next 573\n",
      "2023-04-13 13:08:19.974798: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f343daf8300 of size 2412288 next 849\n",
      "2023-04-13 13:08:19.974807: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f343dd45200 of size 2449152 next 536\n",
      "2023-04-13 13:08:19.974811: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f343df9b100 of size 1929728 next 773\n",
      "2023-04-13 13:08:19.974850: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f343e172300 of size 606208 next 541\n",
      "2023-04-13 13:08:19.974861: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f343e206300 of size 606208 next 900\n",
      "2023-04-13 13:08:19.974872: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f343e29a300 of size 606208 next 660\n",
      "2023-04-13 13:08:19.974911: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f343e32e300 of size 90112 next 590\n",
      "2023-04-13 13:08:19.974921: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f343e344300 of size 90112 next 753\n",
      "2023-04-13 13:08:19.974925: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f343e35a300 of size 920576 next 588\n",
      "2023-04-13 13:08:19.974929: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f343e43af00 of size 2412288 next 781\n",
      "2023-04-13 13:08:19.974957: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f343e687e00 of size 2412288 next 806\n",
      "2023-04-13 13:08:19.974966: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f343e8d4d00 of size 2449152 next 766\n",
      "2023-04-13 13:08:19.974970: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f343eb2ac00 of size 2424576 next 669\n",
      "2023-04-13 13:08:19.974974: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f343ed7ab00 of size 2424576 next 844\n",
      "2023-04-13 13:08:19.975003: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f343efcaa00 of size 12122880 next 996\n",
      "2023-04-13 13:08:19.975012: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f343fb5a500 of size 2424576 next 748\n",
      "2023-04-13 13:08:19.975016: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f343fdaa400 of size 2424576 next 901\n",
      "2023-04-13 13:08:19.975020: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f343fffa300 of size 2412288 next 827\n",
      "2023-04-13 13:08:19.975051: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3440247200 of size 2412288 next 763\n",
      "2023-04-13 13:08:19.975060: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3440494100 of size 2449152 next 789\n",
      "2023-04-13 13:08:19.975064: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f34406ea000 of size 16971520 next 591\n",
      "2023-04-13 13:08:19.975068: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f3441719700 of size 48489728 next 810\n",
      "2023-04-13 13:08:19.975096: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3444557c00 of size 9648640 next 703\n",
      "2023-04-13 13:08:19.975105: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3444e8b600 of size 13383936 next 755\n",
      "2023-04-13 13:08:19.975110: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3445b4ef00 of size 28945664 next 702\n",
      "2023-04-13 13:08:19.975114: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f34476e9c00 of size 38594048 next 919\n",
      "2023-04-13 13:08:19.975142: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3449bb8200 of size 38594048 next 577\n",
      "2023-04-13 13:08:19.975151: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f344c086800 of size 9648640 next 684\n",
      "2023-04-13 13:08:19.975155: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f344c9ba200 of size 9648640 next 790\n",
      "2023-04-13 13:08:19.975159: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f344d2edc00 of size 9648640 next 786\n",
      "2023-04-13 13:08:19.975191: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f344dc21600 of size 9648640 next 903\n",
      "2023-04-13 13:08:19.975200: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f344e555000 of size 9648640 next 462\n",
      "2023-04-13 13:08:19.975204: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f344ee88a00 of size 9648640 next 819\n",
      "2023-04-13 13:08:19.975231: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f344f7bc400 of size 9648640 next 959\n",
      "2023-04-13 13:08:19.975236: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f34500efe00 of size 9648640 next 454\n",
      "2023-04-13 13:08:19.975245: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3450a23800 of size 9648640 next 593\n",
      "2023-04-13 13:08:19.975249: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3451357200 of size 9927424 next 846\n",
      "2023-04-13 13:08:19.975275: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3451cced00 of size 3961088 next 783\n",
      "2023-04-13 13:08:19.975280: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3452095e00 of size 2412288 next 859\n",
      "2023-04-13 13:08:19.975290: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f34522e2d00 of size 2436864 next 788\n",
      "2023-04-13 13:08:19.975294: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3452535c00 of size 90112 next 446\n",
      "2023-04-13 13:08:19.975323: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f345254bc00 of size 90112 next 448\n",
      "2023-04-13 13:08:19.975329: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3452561c00 of size 90112 next 443\n",
      "2023-04-13 13:08:19.975338: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3452577c00 of size 90112 next 535\n",
      "2023-04-13 13:08:19.975342: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f345258dc00 of size 90112 next 534\n",
      "2023-04-13 13:08:19.975373: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f34525a3c00 of size 154375680 next 715\n",
      "2023-04-13 13:08:19.975389: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f345b8dd200 of size 274945024 next 847\n",
      "2023-04-13 13:08:19.975428: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f346bf12600 of size 546048 next 562\n",
      "2023-04-13 13:08:19.975443: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f346bf97b00 of size 90112 next 90\n",
      "2023-04-13 13:08:19.975475: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f346bfadb00 of size 90112 next 96\n",
      "2023-04-13 13:08:19.975484: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f346bfc3b00 of size 90112 next 87\n",
      "2023-04-13 13:08:19.975489: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f346bfd9b00 of size 90112 next 585\n",
      "2023-04-13 13:08:19.975514: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f346bfefb00 of size 90112 next 452\n",
      "2023-04-13 13:08:19.975521: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f346c005b00 of size 791808 next 814\n",
      "2023-04-13 13:08:19.975530: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f346c0c7000 of size 154375680 next 771\n",
      "2023-04-13 13:08:19.975534: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3475400600 of size 154375680 next 823\n",
      "2023-04-13 13:08:19.975561: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f347e739c00 of size 28945664 next 841\n",
      "2023-04-13 13:08:19.975566: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f34802d4900 of size 9648640 next 750\n",
      "2023-04-13 13:08:19.975576: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3480c08300 of size 9648640 next 802\n",
      "2023-04-13 13:08:19.975579: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f348153bd00 of size 9648640 next 764\n",
      "2023-04-13 13:08:19.975613: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3481e6f700 of size 9648640 next 572\n",
      "2023-04-13 13:08:19.975618: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f34827a3100 of size 9648640 next 848\n",
      "2023-04-13 13:08:19.975627: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f34830d6b00 of size 9648640 next 902\n",
      "2023-04-13 13:08:19.975631: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3483a0a500 of size 9648640 next 677\n",
      "2023-04-13 13:08:19.975657: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f348433df00 of size 9648640 next 635\n",
      "2023-04-13 13:08:19.975662: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3484c71900 of size 9648640 next 641\n",
      "2023-04-13 13:08:19.975671: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f34855a5300 of size 9648640 next 675\n",
      "2023-04-13 13:08:19.975675: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3485ed8d00 of size 9648640 next 896\n",
      "2023-04-13 13:08:19.975701: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f348680c700 of size 9648640 next 99\n",
      "2023-04-13 13:08:19.975707: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3487140100 of size 9648640 next 860\n",
      "2023-04-13 13:08:19.975716: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3487a73b00 of size 9648640 next 821\n",
      "2023-04-13 13:08:19.975720: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f34883a7500 of size 9648640 next 663\n",
      "2023-04-13 13:08:19.975750: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3488cdaf00 of size 9648640 next 539\n",
      "2023-04-13 13:08:19.975755: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f348960e900 of size 9648640 next 691\n",
      "2023-04-13 13:08:19.975764: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3489f42300 of size 9648640 next 558\n",
      "2023-04-13 13:08:19.975768: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f348a875d00 of size 9648640 next 898\n",
      "2023-04-13 13:08:19.975794: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f348b1a9700 of size 9648640 next 103\n",
      "2023-04-13 13:08:19.975799: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f348badd100 of size 28945664 next 838\n",
      "2023-04-13 13:08:19.975808: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f348d677e00 of size 28945664 next 833\n",
      "2023-04-13 13:08:19.975813: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f348f212b00 of size 182379008 next 595\n",
      "2023-04-13 13:08:19.975840: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f349a000d00 of size 2424576 next 559\n",
      "2023-04-13 13:08:19.975845: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f349a250c00 of size 2424576 next 95\n",
      "2023-04-13 13:08:19.975855: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f349a4a0b00 of size 58188032 next 863\n",
      "2023-04-13 13:08:19.975859: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f349dc1ec00 of size 9648640 next 725\n",
      "2023-04-13 13:08:19.975895: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f349e552600 of size 13383936 next 869\n",
      "2023-04-13 13:08:19.975904: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f349f215f00 of size 439005952 next 795\n",
      "2023-04-13 13:08:19.975909: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f34b94c1200 of size 439005952 next 805\n",
      "2023-04-13 13:08:19.975913: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f34d376c500 of size 211706880 next 609\n",
      "2023-04-13 13:08:19.975941: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f34e0152900 of size 160621824 next 746\n",
      "2023-04-13 13:08:19.975950: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f34e9a80e00 of size 57891328 next 599\n",
      "2023-04-13 13:08:19.975954: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f34ed1b6800 of size 9648640 next 600\n",
      "2023-04-13 13:08:19.975984: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f34edaea200 of size 9648640 next 557\n",
      "2023-04-13 13:08:19.975990: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f34ee41dc00 of size 9648640 next 605\n",
      "2023-04-13 13:08:19.975999: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f34eed51600 of size 147853568 next 756\n",
      "2023-04-13 13:08:19.976003: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f34f7a52700 of size 485120 next 757\n",
      "2023-04-13 13:08:19.976029: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f34f7ac8e00 of size 439005952 next 525\n",
      "2023-04-13 13:08:19.976035: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3511d74100 of size 154375680 next 537\n",
      "2023-04-13 13:08:19.976045: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f351b0ad700 of size 144517376 next 610\n",
      "2023-04-13 13:08:19.976049: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3523a80000 of size 494002176 next 612\n",
      "2023-04-13 13:08:19.976075: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f354119e000 of size 91660800 next 520\n",
      "2023-04-13 13:08:19.976080: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3546908200 of size 9648640 next 682\n",
      "2023-04-13 13:08:19.976089: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f354723bc00 of size 9648640 next 820\n",
      "2023-04-13 13:08:19.976093: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3547b6f600 of size 9648640 next 554\n",
      "2023-04-13 13:08:19.976119: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f35484a3000 of size 19297024 next 775\n",
      "2023-04-13 13:08:19.976125: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f354970a300 of size 28945664 next 734\n",
      "2023-04-13 13:08:19.976134: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f354b2a5000 of size 9648640 next 727\n",
      "2023-04-13 13:08:19.976138: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f354bbd8a00 of size 9648640 next 824\n",
      "2023-04-13 13:08:19.976165: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f354c50c400 of size 9648640 next 683\n",
      "2023-04-13 13:08:19.976170: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f354ce3fe00 of size 9648640 next 721\n",
      "2023-04-13 13:08:19.976179: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f354d773800 of size 9648640 next 674\n",
      "2023-04-13 13:08:19.976183: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f354e0a7200 of size 9648640 next 910\n",
      "2023-04-13 13:08:19.976210: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f354e9dac00 of size 9648640 next 967\n",
      "2023-04-13 13:08:19.976215: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f354f30e600 of size 48240640 next 526\n",
      "2023-04-13 13:08:19.976224: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f355210fe00 of size 155341312 next 620\n",
      "2023-04-13 13:08:19.976229: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f355b535000 of size 1929728 next 621\n",
      "2023-04-13 13:08:19.976261: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f355b70c200 of size 9698048 next 622\n",
      "2023-04-13 13:08:19.976270: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f355c04bd00 of size 242448640 next 652\n",
      "2023-04-13 13:08:19.976275: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f356a783600 of size 48489728 next 653\n",
      "2023-04-13 13:08:19.976301: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f356d5c1b00 of size 12122624 next 654\n",
      "2023-04-13 13:08:19.976307: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f356e151500 of size 9648640 next 512\n",
      "2023-04-13 13:08:19.976316: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f356ea84f00 of size 9648640 next 513\n",
      "2023-04-13 13:08:19.976320: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f356f3b8900 of size 17072640 next 689\n",
      "2023-04-13 13:08:19.976347: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3570400b00 of size 36367360 next 690\n",
      "2023-04-13 13:08:19.976352: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f35726af700 of size 3030784 next 692\n",
      "2023-04-13 13:08:19.976362: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3572993600 of size 9091840 next 693\n",
      "2023-04-13 13:08:19.976366: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f357323f100 of size 12122624 next 694\n",
      "2023-04-13 13:08:19.976393: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3573dceb00 of size 3030784 next 695\n",
      "2023-04-13 13:08:19.976397: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f35740b2a00 of size 3030784 next 696\n",
      "2023-04-13 13:08:19.976406: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3574396900 of size 2412288 next 648\n",
      "2023-04-13 13:08:19.976410: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f35745e3800 of size 2412288 next 655\n",
      "2023-04-13 13:08:19.976438: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3574830700 of size 4267776 next 699\n",
      "2023-04-13 13:08:19.976447: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3574c42600 of size 3030784 next 700\n",
      "2023-04-13 13:08:19.976451: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3574f26500 of size 3030784 next 701\n",
      "2023-04-13 13:08:19.976455: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f357520a400 of size 3312640 next 477\n",
      "2023-04-13 13:08:19.976483: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3575533000 of size 5779712 next 704\n",
      "2023-04-13 13:08:19.976492: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3575ab6100 of size 3030784 next 705\n",
      "2023-04-13 13:08:19.976497: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3575d9a000 of size 3030784 next 706\n",
      "2023-04-13 13:08:19.976527: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f357607df00 of size 15153920 next 711\n",
      "2023-04-13 13:08:19.976533: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3576ef1a00 of size 3030784 next 712\n",
      "2023-04-13 13:08:19.976542: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f35771d5900 of size 3030784 next 713\n",
      "2023-04-13 13:08:19.976546: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f35774b9800 of size 9092352 next 716\n",
      "2023-04-13 13:08:19.976573: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3577d65500 of size 3030784 next 717\n",
      "2023-04-13 13:08:19.976582: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f3578049400 of size 9091840 next 718\n",
      "2023-04-13 13:08:19.976587: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f35788f4f00 of size 9091840 next 719\n",
      "2023-04-13 13:08:19.976591: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f35791a0a00 of size 28945664 next 108\n",
      "2023-04-13 13:08:19.976618: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f357ad3b700 of size 46819584 next 720\n",
      "2023-04-13 13:08:19.976628: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f357d9e2000 of size 36367360 next 722\n",
      "2023-04-13 13:08:19.976632: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f357fc90c00 of size 442469376 next 735\n",
      "2023-04-13 13:08:19.976636: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359a289800 of size 36367360 next 733\n",
      "2023-04-13 13:08:19.976664: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c538400 of size 614400 next 760\n",
      "2023-04-13 13:08:19.976673: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c5ce400 of size 485120 next 464\n",
      "2023-04-13 13:08:19.976678: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c644b00 of size 606976 next 476\n",
      "2023-04-13 13:08:19.976681: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c6d8e00 of size 546048 next 472\n",
      "2023-04-13 13:08:19.976709: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f359c75e300 of size 512 next 932\n",
      "2023-04-13 13:08:19.976718: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c75e500 of size 256 next 933\n",
      "2023-04-13 13:08:19.976723: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f359c75e600 of size 512 next 934\n",
      "2023-04-13 13:08:19.976727: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c75e800 of size 256 next 935\n",
      "2023-04-13 13:08:19.976754: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c75e900 of size 256 next 938\n",
      "2023-04-13 13:08:19.976763: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c75ea00 of size 256 next 939\n",
      "2023-04-13 13:08:19.976767: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c75eb00 of size 256 next 936\n",
      "2023-04-13 13:08:19.976771: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c75ec00 of size 256 next 941\n",
      "2023-04-13 13:08:19.976801: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c75ed00 of size 256 next 942\n",
      "2023-04-13 13:08:19.976806: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c75ee00 of size 256 next 807\n",
      "2023-04-13 13:08:19.976815: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f359c75ef00 of size 512 next 943\n",
      "2023-04-13 13:08:19.976819: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c75f100 of size 256 next 946\n",
      "2023-04-13 13:08:19.976846: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c75f200 of size 256 next 947\n",
      "2023-04-13 13:08:19.976851: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c75f300 of size 256 next 945\n",
      "2023-04-13 13:08:19.976860: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f359c75f400 of size 512 next 950\n",
      "2023-04-13 13:08:19.976863: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c75f600 of size 256 next 951\n",
      "2023-04-13 13:08:19.976896: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c75f700 of size 256 next 953\n",
      "2023-04-13 13:08:19.976901: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c75f800 of size 256 next 954\n",
      "2023-04-13 13:08:19.976910: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c75f900 of size 256 next 944\n",
      "2023-04-13 13:08:19.976914: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c75fa00 of size 1536 next 916\n",
      "2023-04-13 13:08:19.976941: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c760000 of size 256 next 918\n",
      "2023-04-13 13:08:19.976951: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f359c760100 of size 512 next 955\n",
      "2023-04-13 13:08:19.976955: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c760300 of size 256 next 956\n",
      "2023-04-13 13:08:19.976959: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c760400 of size 256 next 960\n",
      "2023-04-13 13:08:19.976986: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c760500 of size 256 next 961\n",
      "2023-04-13 13:08:19.976995: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f359c760600 of size 512 next 963\n",
      "2023-04-13 13:08:19.977000: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c760800 of size 256 next 965\n",
      "2023-04-13 13:08:19.977003: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c760900 of size 256 next 962\n",
      "2023-04-13 13:08:19.977031: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c760a00 of size 256 next 968\n",
      "2023-04-13 13:08:19.977040: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c760b00 of size 256 next 969\n",
      "2023-04-13 13:08:19.977044: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c760c00 of size 256 next 976\n",
      "2023-04-13 13:08:19.977048: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f359c760d00 of size 512 next 977\n",
      "2023-04-13 13:08:19.977075: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c760f00 of size 256 next 978\n",
      "2023-04-13 13:08:19.977084: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c761000 of size 256 next 981\n",
      "2023-04-13 13:08:19.977089: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c761100 of size 256 next 982\n",
      "2023-04-13 13:08:19.977093: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c761200 of size 256 next 984\n",
      "2023-04-13 13:08:19.977119: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c761300 of size 256 next 986\n",
      "2023-04-13 13:08:19.977128: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c761400 of size 256 next 987\n",
      "2023-04-13 13:08:19.977133: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c761500 of size 256 next 985\n",
      "2023-04-13 13:08:19.977137: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c761600 of size 256 next 988\n",
      "2023-04-13 13:08:19.977164: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f359c761700 of size 512 next 990\n",
      "2023-04-13 13:08:19.977173: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c761900 of size 256 next 991\n",
      "2023-04-13 13:08:19.977177: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c761a00 of size 256 next 989\n",
      "2023-04-13 13:08:19.977181: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c761b00 of size 256 next 993\n",
      "2023-04-13 13:08:19.977208: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c761c00 of size 256 next 551\n",
      "2023-04-13 13:08:19.977217: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c761d00 of size 256 next 994\n",
      "2023-04-13 13:08:19.977221: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c761e00 of size 256 next 995\n",
      "2023-04-13 13:08:19.977225: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c761f00 of size 256 next 992\n",
      "2023-04-13 13:08:19.977253: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c762000 of size 256 next 1000\n",
      "2023-04-13 13:08:19.977262: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c762100 of size 256 next 1001\n",
      "2023-04-13 13:08:19.977266: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c762200 of size 256 next 1002\n",
      "2023-04-13 13:08:19.977292: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c762300 of size 256 next 1003\n",
      "2023-04-13 13:08:19.977297: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c762400 of size 256 next 1004\n",
      "2023-04-13 13:08:19.977306: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c762500 of size 256 next 1005\n",
      "2023-04-13 13:08:19.977310: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c762600 of size 256 next 1007\n",
      "2023-04-13 13:08:19.977314: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c762700 of size 256 next 1008\n",
      "2023-04-13 13:08:19.977341: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c762800 of size 256 next 1009\n",
      "2023-04-13 13:08:19.977350: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c762900 of size 256 next 1010\n",
      "2023-04-13 13:08:19.977354: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c762a00 of size 256 next 1011\n",
      "2023-04-13 13:08:19.977358: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c762b00 of size 256 next 1013\n",
      "2023-04-13 13:08:19.977385: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c762c00 of size 256 next 1014\n",
      "2023-04-13 13:08:19.977394: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c762d00 of size 256 next 1015\n",
      "2023-04-13 13:08:19.977398: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c762e00 of size 256 next 1017\n",
      "2023-04-13 13:08:19.977402: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c762f00 of size 256 next 973\n",
      "2023-04-13 13:08:19.977429: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c763000 of size 256 next 1018\n",
      "2023-04-13 13:08:19.977438: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c763100 of size 256 next 1021\n",
      "2023-04-13 13:08:19.977443: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c763200 of size 256 next 1022\n",
      "2023-04-13 13:08:19.977447: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c763300 of size 256 next 1023\n",
      "2023-04-13 13:08:19.977475: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c763400 of size 256 next 1019\n",
      "2023-04-13 13:08:19.977484: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c763500 of size 256 next 1026\n",
      "2023-04-13 13:08:19.977488: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c763600 of size 256 next 1028\n",
      "2023-04-13 13:08:19.977492: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c763700 of size 256 next 1027\n",
      "2023-04-13 13:08:19.977519: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c763800 of size 256 next 1032\n",
      "2023-04-13 13:08:19.977528: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c763900 of size 256 next 1034\n",
      "2023-04-13 13:08:19.977532: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c763a00 of size 256 next 1035\n",
      "2023-04-13 13:08:19.977536: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c763b00 of size 256 next 1036\n",
      "2023-04-13 13:08:19.977563: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c763c00 of size 256 next 1037\n",
      "2023-04-13 13:08:19.977572: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c763d00 of size 256 next 1040\n",
      "2023-04-13 13:08:19.977576: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c763e00 of size 256 next 379\n",
      "2023-04-13 13:08:19.977580: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c763f00 of size 256 next 428\n",
      "2023-04-13 13:08:19.977606: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c764000 of size 256 next 429\n",
      "2023-04-13 13:08:19.977611: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c764100 of size 256 next 414\n",
      "2023-04-13 13:08:19.977620: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c764200 of size 256 next 415\n",
      "2023-04-13 13:08:19.977624: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c764300 of size 256 next 389\n",
      "2023-04-13 13:08:19.977650: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c764400 of size 256 next 392\n",
      "2023-04-13 13:08:19.977655: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c764500 of size 256 next 278\n",
      "2023-04-13 13:08:19.977664: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c764600 of size 256 next 191\n",
      "2023-04-13 13:08:19.977668: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c764700 of size 256 next 153\n",
      "2023-04-13 13:08:19.977695: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c764800 of size 256 next 162\n",
      "2023-04-13 13:08:19.977704: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c764900 of size 256 next 334\n",
      "2023-04-13 13:08:19.977708: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c764a00 of size 256 next 363\n",
      "2023-04-13 13:08:19.977712: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c764b00 of size 256 next 155\n",
      "2023-04-13 13:08:19.977744: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c764c00 of size 256 next 411\n",
      "2023-04-13 13:08:19.977749: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c764d00 of size 256 next 220\n",
      "2023-04-13 13:08:19.977753: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c764e00 of size 256 next 274\n",
      "2023-04-13 13:08:19.977762: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c764f00 of size 256 next 241\n",
      "2023-04-13 13:08:19.977766: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c765000 of size 256 next 131\n",
      "2023-04-13 13:08:19.977792: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c765100 of size 256 next 222\n",
      "2023-04-13 13:08:19.977798: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c765200 of size 256 next 340\n",
      "2023-04-13 13:08:19.977806: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c765300 of size 256 next 353\n",
      "2023-04-13 13:08:19.977810: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c765400 of size 256 next 288\n",
      "2023-04-13 13:08:19.977840: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c765500 of size 256 next 385\n",
      "2023-04-13 13:08:19.977845: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c765600 of size 256 next 123\n",
      "2023-04-13 13:08:19.977854: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c765700 of size 256 next 300\n",
      "2023-04-13 13:08:19.977858: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c765800 of size 256 next 262\n",
      "2023-04-13 13:08:19.977890: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c765900 of size 256 next 393\n",
      "2023-04-13 13:08:19.977896: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c765a00 of size 256 next 301\n",
      "2023-04-13 13:08:19.977904: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c765b00 of size 256 next 416\n",
      "2023-04-13 13:08:19.977909: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c765c00 of size 256 next 91\n",
      "2023-04-13 13:08:19.977940: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c765d00 of size 256 next 373\n",
      "2023-04-13 13:08:19.977946: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c765e00 of size 256 next 291\n",
      "2023-04-13 13:08:19.977954: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c765f00 of size 256 next 253\n",
      "2023-04-13 13:08:19.977958: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c766000 of size 256 next 322\n",
      "2023-04-13 13:08:19.977984: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c766100 of size 256 next 434\n",
      "2023-04-13 13:08:19.977990: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c766200 of size 256 next 239\n",
      "2023-04-13 13:08:19.977998: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c766300 of size 256 next 282\n",
      "2023-04-13 13:08:19.978003: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c766400 of size 256 next 342\n",
      "2023-04-13 13:08:19.978028: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c766500 of size 256 next 173\n",
      "2023-04-13 13:08:19.978034: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c766600 of size 256 next 430\n",
      "2023-04-13 13:08:19.978042: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c766700 of size 256 next 214\n",
      "2023-04-13 13:08:19.978046: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c766800 of size 256 next 170\n",
      "2023-04-13 13:08:19.978072: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c766900 of size 256 next 206\n",
      "2023-04-13 13:08:19.978077: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c766a00 of size 256 next 421\n",
      "2023-04-13 13:08:19.978086: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c766b00 of size 256 next 437\n",
      "2023-04-13 13:08:19.978090: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f359c766c00 of size 256 next 221\n",
      "2023-04-13 13:08:19.978116: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c766d00 of size 256 next 265\n",
      "2023-04-13 13:08:19.978121: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f359c766e00 of size 256 next 354\n",
      "2023-04-13 13:08:19.978130: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c766f00 of size 256 next 172\n",
      "2023-04-13 13:08:19.978134: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f359c767000 of size 256 next 134\n",
      "2023-04-13 13:08:19.978159: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c767100 of size 256 next 254\n",
      "2023-04-13 13:08:19.978164: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c767200 of size 768 next 550\n",
      "2023-04-13 13:08:19.978173: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c767500 of size 768 next 548\n",
      "2023-04-13 13:08:19.978177: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] Free  at 7f359c767800 of size 51968 next 308\n",
      "2023-04-13 13:08:19.978203: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c774300 of size 90112 next 471\n",
      "2023-04-13 13:08:19.978208: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c78a300 of size 90112 next 468\n",
      "2023-04-13 13:08:19.978217: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c7a0300 of size 90112 next 459\n",
      "2023-04-13 13:08:19.978221: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c7b6300 of size 90112 next 457\n",
      "2023-04-13 13:08:19.978225: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c7cc300 of size 90112 next 455\n",
      "2023-04-13 13:08:19.978252: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c7e2300 of size 90112 next 538\n",
      "2023-04-13 13:08:19.978261: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c7f8300 of size 147456 next 736\n",
      "2023-04-13 13:08:19.978266: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359c81c300 of size 3030784 next 737\n",
      "2023-04-13 13:08:19.978270: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359cb00200 of size 9648640 next 516\n",
      "2023-04-13 13:08:19.978317: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359d433c00 of size 9648640 next 517\n",
      "2023-04-13 13:08:19.978322: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359dd67600 of size 9648640 next 518\n",
      "2023-04-13 13:08:19.978331: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f359e69b000 of size 28945664 next 135\n",
      "2023-04-13 13:08:19.978336: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f35a0235d00 of size 29996800 next 743\n",
      "2023-04-13 13:08:19.978363: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f35a1ed1400 of size 36367360 next 744\n",
      "2023-04-13 13:08:19.978368: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f35a4180000 of size 124254720 next 745\n",
      "2023-04-13 13:08:19.978378: I tensorflow/core/common_runtime/bfc_allocator.cc:1066] InUse at 7f35ab7ffa00 of size 609224192 next 18446744073709551615\n",
      "2023-04-13 13:08:19.978382: I tensorflow/core/common_runtime/bfc_allocator.cc:1071]      Summary of in-use Chunks by size: \n",
      "2023-04-13 13:08:19.978414: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 232 Chunks of size 256 totalling 58.0KiB\n",
      "2023-04-13 13:08:19.978426: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 29 Chunks of size 512 totalling 14.5KiB\n",
      "2023-04-13 13:08:19.978453: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 39 Chunks of size 768 totalling 29.2KiB\n",
      "2023-04-13 13:08:19.978459: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 1024 totalling 1.0KiB\n",
      "2023-04-13 13:08:19.978469: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 4 Chunks of size 1280 totalling 5.0KiB\n",
      "2023-04-13 13:08:19.978473: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 2 Chunks of size 1536 totalling 3.0KiB\n",
      "2023-04-13 13:08:19.978506: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 2 Chunks of size 1792 totalling 3.5KiB\n",
      "2023-04-13 13:08:19.978516: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 7 Chunks of size 2560 totalling 17.5KiB\n",
      "2023-04-13 13:08:19.978521: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 2 Chunks of size 3840 totalling 7.5KiB\n",
      "2023-04-13 13:08:19.978548: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 4096 totalling 4.0KiB\n",
      "2023-04-13 13:08:19.978554: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 4608 totalling 4.5KiB\n",
      "2023-04-13 13:08:19.978564: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 5 Chunks of size 4864 totalling 23.8KiB\n",
      "2023-04-13 13:08:19.978568: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 5120 totalling 5.0KiB\n",
      "2023-04-13 13:08:19.978597: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 6144 totalling 6.0KiB\n",
      "2023-04-13 13:08:19.978606: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 6912 totalling 6.8KiB\n",
      "2023-04-13 13:08:19.978611: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 4 Chunks of size 29952 totalling 117.0KiB\n",
      "2023-04-13 13:08:19.978640: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 6 Chunks of size 30464 totalling 178.5KiB\n",
      "2023-04-13 13:08:19.978651: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 3 Chunks of size 30720 totalling 90.0KiB\n",
      "2023-04-13 13:08:19.978655: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 40704 totalling 39.8KiB\n",
      "2023-04-13 13:08:19.978687: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 56832 totalling 55.5KiB\n",
      "2023-04-13 13:08:19.978697: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 60928 totalling 59.5KiB\n",
      "2023-04-13 13:08:19.978702: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 30 Chunks of size 90112 totalling 2.58MiB\n",
      "2023-04-13 13:08:19.978730: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 92416 totalling 90.2KiB\n",
      "2023-04-13 13:08:19.978741: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 97536 totalling 95.2KiB\n",
      "2023-04-13 13:08:19.978746: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 119808 totalling 117.0KiB\n",
      "2023-04-13 13:08:19.978773: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 121344 totalling 118.5KiB\n",
      "2023-04-13 13:08:19.978779: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 122368 totalling 119.5KiB\n",
      "2023-04-13 13:08:19.978789: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 124160 totalling 121.2KiB\n",
      "2023-04-13 13:08:19.978794: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 147456 totalling 144.0KiB\n",
      "2023-04-13 13:08:19.978824: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 167936 totalling 164.0KiB\n",
      "2023-04-13 13:08:19.978834: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 242944 totalling 237.2KiB\n",
      "2023-04-13 13:08:19.978838: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 4 Chunks of size 485120 totalling 1.85MiB\n",
      "2023-04-13 13:08:19.978877: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 3 Chunks of size 546048 totalling 1.56MiB\n",
      "2023-04-13 13:08:19.978883: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 11 Chunks of size 606208 totalling 6.36MiB\n",
      "2023-04-13 13:08:19.978893: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 2 Chunks of size 606976 totalling 1.16MiB\n",
      "2023-04-13 13:08:19.978898: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 611072 totalling 596.8KiB\n",
      "2023-04-13 13:08:19.978926: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 2 Chunks of size 614400 totalling 1.17MiB\n",
      "2023-04-13 13:08:19.978936: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 791808 totalling 773.2KiB\n",
      "2023-04-13 13:08:19.978941: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 848128 totalling 828.2KiB\n",
      "2023-04-13 13:08:19.978971: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 920576 totalling 899.0KiB\n",
      "2023-04-13 13:08:19.978977: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 969216 totalling 946.5KiB\n",
      "2023-04-13 13:08:19.978986: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 1050624 totalling 1.00MiB\n",
      "2023-04-13 13:08:19.978991: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 1204224 totalling 1.15MiB\n",
      "2023-04-13 13:08:19.979029: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 4 Chunks of size 1929728 totalling 7.36MiB\n",
      "2023-04-13 13:08:19.979062: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 2387200 totalling 2.28MiB\n",
      "2023-04-13 13:08:19.979072: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 2395648 totalling 2.28MiB\n",
      "2023-04-13 13:08:19.979078: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 13 Chunks of size 2412288 totalling 29.91MiB\n",
      "2023-04-13 13:08:19.979106: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 28 Chunks of size 2424576 totalling 64.74MiB\n",
      "2023-04-13 13:08:19.979113: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 2436864 totalling 2.32MiB\n",
      "2023-04-13 13:08:19.979122: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 2444288 totalling 2.33MiB\n",
      "2023-04-13 13:08:19.979127: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 3 Chunks of size 2449152 totalling 7.01MiB\n",
      "2023-04-13 13:08:19.979162: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 2503168 totalling 2.39MiB\n",
      "2023-04-13 13:08:19.979172: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 18 Chunks of size 3030784 totalling 52.03MiB\n",
      "2023-04-13 13:08:19.979177: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 2 Chunks of size 3312640 totalling 6.32MiB\n",
      "2023-04-13 13:08:19.979203: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 3961088 totalling 3.78MiB\n",
      "2023-04-13 13:08:19.979209: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 4267776 totalling 4.07MiB\n",
      "2023-04-13 13:08:19.979218: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 4619264 totalling 4.41MiB\n",
      "2023-04-13 13:08:19.979223: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 2 Chunks of size 4848896 totalling 9.25MiB\n",
      "2023-04-13 13:08:19.979251: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 5427200 totalling 5.18MiB\n",
      "2023-04-13 13:08:19.979260: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 5779712 totalling 5.51MiB\n",
      "2023-04-13 13:08:19.979265: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 5999104 totalling 5.72MiB\n",
      "2023-04-13 13:08:19.979291: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 7236608 totalling 6.90MiB\n",
      "2023-04-13 13:08:19.979298: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 7 Chunks of size 7273472 totalling 48.56MiB\n",
      "2023-04-13 13:08:19.979307: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 7274496 totalling 6.94MiB\n",
      "2023-04-13 13:08:19.979312: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 7768576 totalling 7.41MiB\n",
      "2023-04-13 13:08:19.979339: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 5 Chunks of size 9091840 totalling 43.35MiB\n",
      "2023-04-13 13:08:19.979350: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 9092352 totalling 8.67MiB\n",
      "2023-04-13 13:08:19.979355: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 74 Chunks of size 9648640 totalling 680.92MiB\n",
      "2023-04-13 13:08:19.979382: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 5 Chunks of size 9698048 totalling 46.24MiB\n",
      "2023-04-13 13:08:19.979393: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 9796608 totalling 9.34MiB\n",
      "2023-04-13 13:08:19.979398: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 9927424 totalling 9.47MiB\n",
      "2023-04-13 13:08:19.979424: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 10142976 totalling 9.67MiB\n",
      "2023-04-13 13:08:19.979430: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 10639104 totalling 10.15MiB\n",
      "2023-04-13 13:08:19.979440: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 12061184 totalling 11.50MiB\n",
      "2023-04-13 13:08:19.979445: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 3 Chunks of size 12122624 totalling 34.68MiB\n",
      "2023-04-13 13:08:19.979474: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 12122880 totalling 11.56MiB\n",
      "2023-04-13 13:08:19.979484: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 2 Chunks of size 13383936 totalling 25.53MiB\n",
      "2023-04-13 13:08:19.979488: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 15153920 totalling 14.45MiB\n",
      "2023-04-13 13:08:19.979518: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 3 Chunks of size 16971520 totalling 48.56MiB\n",
      "2023-04-13 13:08:19.979529: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 17072640 totalling 16.28MiB\n",
      "2023-04-13 13:08:19.979534: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 19297024 totalling 18.40MiB\n",
      "2023-04-13 13:08:19.979561: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 5 Chunks of size 23032576 totalling 109.83MiB\n",
      "2023-04-13 13:08:19.979567: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 24059392 totalling 22.94MiB\n",
      "2023-04-13 13:08:19.979577: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 11 Chunks of size 28945664 totalling 303.65MiB\n",
      "2023-04-13 13:08:19.979603: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 28945920 totalling 27.60MiB\n",
      "2023-04-13 13:08:19.979609: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 2 Chunks of size 29093888 totalling 55.49MiB\n",
      "2023-04-13 13:08:19.979619: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 29996800 totalling 28.61MiB\n",
      "2023-04-13 13:08:19.979624: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 31114496 totalling 29.67MiB\n",
      "2023-04-13 13:08:19.979651: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 4 Chunks of size 36367360 totalling 138.73MiB\n",
      "2023-04-13 13:08:19.979661: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 4 Chunks of size 38594048 totalling 147.22MiB\n",
      "2023-04-13 13:08:19.979667: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 2 Chunks of size 38791680 totalling 73.99MiB\n",
      "2023-04-13 13:08:19.979694: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 38791936 totalling 36.99MiB\n",
      "2023-04-13 13:08:19.979699: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 41005824 totalling 39.11MiB\n",
      "2023-04-13 13:08:19.979709: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 46819584 totalling 44.65MiB\n",
      "2023-04-13 13:08:19.979714: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 48240640 totalling 46.01MiB\n",
      "2023-04-13 13:08:19.979742: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 48489728 totalling 46.24MiB\n",
      "2023-04-13 13:08:19.979752: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 53066752 totalling 50.61MiB\n",
      "2023-04-13 13:08:19.979757: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 57891328 totalling 55.21MiB\n",
      "2023-04-13 13:08:19.979787: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 58188032 totalling 55.49MiB\n",
      "2023-04-13 13:08:19.979798: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 2 Chunks of size 81220096 totalling 154.92MiB\n",
      "2023-04-13 13:08:19.979803: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 14 Chunks of size 91660800 totalling 1.19GiB\n",
      "2023-04-13 13:08:19.979830: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 2 Chunks of size 110313984 totalling 210.41MiB\n",
      "2023-04-13 13:08:19.979836: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 115781888 totalling 110.42MiB\n",
      "2023-04-13 13:08:19.979846: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 124254720 totalling 118.50MiB\n",
      "2023-04-13 13:08:19.979878: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 144517376 totalling 137.82MiB\n",
      "2023-04-13 13:08:19.979884: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 145764352 totalling 139.01MiB\n",
      "2023-04-13 13:08:19.979894: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 147853568 totalling 141.00MiB\n",
      "2023-04-13 13:08:19.979899: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 5 Chunks of size 154375680 totalling 736.12MiB\n",
      "2023-04-13 13:08:19.979932: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 155341312 totalling 148.14MiB\n",
      "2023-04-13 13:08:19.979943: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 160621824 totalling 153.18MiB\n",
      "2023-04-13 13:08:19.979947: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 182379008 totalling 173.93MiB\n",
      "2023-04-13 13:08:19.979975: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 3 Chunks of size 192969728 totalling 552.09MiB\n",
      "2023-04-13 13:08:19.979985: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 242448640 totalling 231.22MiB\n",
      "2023-04-13 13:08:19.979990: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 274945024 totalling 262.21MiB\n",
      "2023-04-13 13:08:19.980017: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 285925376 totalling 272.68MiB\n",
      "2023-04-13 13:08:19.980023: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 288380928 totalling 275.02MiB\n",
      "2023-04-13 13:08:19.980033: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 294279936 totalling 280.65MiB\n",
      "2023-04-13 13:08:19.980059: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 323224320 totalling 308.25MiB\n",
      "2023-04-13 13:08:19.980065: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 5 Chunks of size 439005952 totalling 2.04GiB\n",
      "2023-04-13 13:08:19.980075: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 442469376 totalling 421.97MiB\n",
      "2023-04-13 13:08:19.980080: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 494002176 totalling 471.12MiB\n",
      "2023-04-13 13:08:19.980110: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 609224192 totalling 581.00MiB\n",
      "2023-04-13 13:08:19.980121: I tensorflow/core/common_runtime/bfc_allocator.cc:1078] Sum Total of in-use chunks: 11.54GiB\n",
      "2023-04-13 13:08:19.980126: I tensorflow/core/common_runtime/bfc_allocator.cc:1080] total_region_allocated_bytes_: 14458814464 memory_limit_: 14458814464 available bytes: 0 curr_region_allocation_bytes_: 28917628928\n",
      "2023-04-13 13:08:19.980165: I tensorflow/core/common_runtime/bfc_allocator.cc:1086] Stats: \n",
      "Limit:                     14458814464\n",
      "InUse:                     12391026688\n",
      "MaxInUse:                  13848977920\n",
      "NumAllocs:                      292452\n",
      "MaxAllocSize:                639212032\n",
      "Reserved:                            0\n",
      "PeakReserved:                        0\n",
      "LargestFreeBlock:                    0\n",
      "\n",
      "2023-04-13 13:08:19.980256: W tensorflow/core/common_runtime/bfc_allocator.cc:474] *******************__***x**_******__***********************************_****************************\n",
      "2023-04-13 13:08:19.980329: W tensorflow/core/framework/op_kernel.cc:1745] OP_REQUIRES failed at numeric_op.h:82 : RESOURCE_EXHAUSTED: OOM when allocating tensor with shape[120606,910] and type float on /job:localhost/replica:0/task:0/device:GPU:0 by allocator GPU_0_bfc\n",
      " 27%|██▋       | 81/300 [01:12<03:14,  1.12it/s]\n"
     ]
    },
    {
     "ename": "ResourceExhaustedError",
     "evalue": "OOM when allocating tensor with shape[120606,910] and type float on /job:localhost/replica:0/task:0/device:GPU:0 by allocator GPU_0_bfc [Op:ReluGrad]",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mResourceExhaustedError\u001b[0m                    Traceback (most recent call last)",
      "\u001b[0;32m/tmp/ipykernel_634/4174207990.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mLBNdnn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mastype\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfloat32\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlabels\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mastype\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfloat32\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minputs_val\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mastype\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfloat32\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlabels_val\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mastype\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfloat32\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m/tmp/ipykernel_634/4146401263.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, X, Y, X_val, Y_val, show_loss)\u001b[0m\n\u001b[1;32m     59\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     60\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0miEpoch\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mtqdm\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtqdm\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnEpochs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 61\u001b[0;31m                 \u001b[0ml\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0ml_val\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_train\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mY\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mX_val\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mY_val\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     62\u001b[0m                 \u001b[0mlosses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m \u001b[0;34m(\u001b[0m \u001b[0ml\u001b[0m \u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     63\u001b[0m                 \u001b[0mlosses_val\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m \u001b[0;34m(\u001b[0m \u001b[0ml_val\u001b[0m \u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python3.8/dist-packages/tensorflow/python/util/traceback_utils.py\u001b[0m in \u001b[0;36merror_handler\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m    151\u001b[0m     \u001b[0;32mexcept\u001b[0m \u001b[0mException\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    152\u001b[0m       \u001b[0mfiltered_tb\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0m_process_traceback_frames\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0me\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__traceback__\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 153\u001b[0;31m       \u001b[0;32mraise\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mwith_traceback\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfiltered_tb\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    154\u001b[0m     \u001b[0;32mfinally\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    155\u001b[0m       \u001b[0;32mdel\u001b[0m \u001b[0mfiltered_tb\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/tmp/ipykernel_634/4146401263.py\u001b[0m in \u001b[0;36m_train\u001b[0;34m(self, X, Y, X_val, Y_val)\u001b[0m\n\u001b[1;32m    119\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    120\u001b[0m             \u001b[0;31m# Compute the gradient of the overall loss with respect to the classifier weights\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 121\u001b[0;31m             \u001b[0mgradients\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mgt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgradient\u001b[0m \u001b[0;34m(\u001b[0m \u001b[0mloss\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mweights\u001b[0m \u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    122\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    123\u001b[0m         \u001b[0;31m# Apply the gradients\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mResourceExhaustedError\u001b[0m: OOM when allocating tensor with shape[120606,910] and type float on /job:localhost/replica:0/task:0/device:GPU:0 by allocator GPU_0_bfc [Op:ReluGrad]"
     ]
    }
   ],
   "source": [
    "LBNdnn.fit(inputs.astype(np.float32), labels.astype(np.float32), inputs_val.astype(np.float32), labels_val.astype(np.float32))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "c27991a8-0610-4b73-b46c-04da509536d6",
   "metadata": {},
   "outputs": [],
   "source": [
    "# df[df['isMIX']==1]\n",
    "# seen = set()\n",
    "# uniq = [x for x in (df[df['isTT']==1]).index.values.tolist() if x not in seen and not seen.add(x)]\n",
    "# print(len(df[df['isMIX']==1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "50953bd8-c6e2-4675-bf3c-1ebaef4563f5",
   "metadata": {},
   "outputs": [],
   "source": [
    "# v = df['LHEPart_pt_'+str(2)][6637]*np.array([math.cosh(df['LHEPart_eta_'+str(2)][6637]),\n",
    "#                                                                                         math.cos(df['LHEPart_phi_'+str(2)][6637]),\n",
    "#                                                                                         math.sin(df['LHEPart_phi_'+str(2)][6637]),\n",
    "#                                                                                         math.sinh(df['LHEPart_eta_'+str(2)][6637])])\n",
    "# print(df.loc[[6637]])\n",
    "# print(df['LHEPart_eta_'+str(2)][6637]*np.array([math.cosh(df['LHEPart_eta_'+str(2)][6637]),math.cos(df['LHEPart_phi_'+str(2)][6637]),math.sin(df['LHEPart_phi_'+str(2)][6637]),math.sinh(df['LHEPart_eta_'+str(2)][6637])]))\n",
    "# v = df['LHEPart_eta_'+str(2)][29524]*np.array([math.cosh(df['LHEPart_eta_'+str(2)][29524]),math.cos(df['LHEPart_phi_'+str(2)][29524]),math.sin(df['LHEPart_phi_'+str(2)][29524]),math.sinh(df['LHEPart_eta_'+str(2)][29524])])\n",
    "# print(v)\n",
    "# v = df['LHEPart_pt_'+str(2)][6637]*np.array([math.cosh(df['LHEPart_eta_'+str(2)][29524]),math.cos(df['LHEPart_phi_'+str(2)][29524]),math.sin(df['LHEPart_phi_'+str(2)][29524]),math.sinh(df['LHEPart_eta_'+str(2)][29524])])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "id": "14f375de-3917-4351-ba4e-12f4452f275b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Index\n",
      "4000\n",
      "16000\n"
     ]
    }
   ],
   "source": [
    "inputs = np.ones((16000,6,4))\n",
    "inputs[event,particle,...]\n",
    "df.index.name = 'Index'\n",
    "print(df.index.name)\n",
    "df = df.rename_axis('Row',axis=1)\n",
    "seen = set()\n",
    "uniq = [x for x in (df[df['isBKG']>0]).index.values.tolist() if x not in seen and not seen.add(x)]\n",
    "print(len(uniq))\n",
    "print(len(df.index.values.tolist()))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b70103c6-75c0-43ba-9ab4-c1f9cf0779bb",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}