3(Training). Naval Mine Identifier (Multilayer Perceptron) NN Tensorflow .ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import tensorflow as tf\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.utils import shuffle\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(166, 60)\n",
      "(166, 2)\n",
      "(42, 60)\n",
      "(42, 2)\n",
      "(208, 2)\n"
     ]
    }
   ],
   "source": [
    "def read_dataset():\n",
    "    df = pd.read_csv(\"Naval Mine/dataset_40_sonar.csv\")\n",
    "    X = df[df.columns[0:60]].values\n",
    "    Y = df[df.columns[60]]\n",
    "    encode = LabelEncoder()\n",
    "    encode.fit(Y)\n",
    "    Y = encode.transform(Y)\n",
    "    Y = one_hot_encode(Y)\n",
    "    return (X,Y)\n",
    "\n",
    "def one_hot_encode(labels):\n",
    "    n_labels=len(labels)\n",
    "    n_unique_labels=len(np.unique(labels))\n",
    "    one_hot_encode=np.zeros((n_labels,n_unique_labels))\n",
    "    one_hot_encode[np.arange(n_labels), labels]=1\n",
    "    return one_hot_encode\n",
    "\n",
    "X, Y=read_dataset()\n",
    "X, Y=shuffle(X, Y,random_state=1)\n",
    "train_x,test_x,train_y,test_y = train_test_split(X, Y,test_size=0.20,random_state=415)\n",
    "\n",
    "print(train_x.shape)\n",
    "print(train_y.shape)\n",
    "print(test_x.shape)\n",
    "print(test_y.shape)\n",
    "print(Y.shape)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "n_dim=  60\n",
      "WARNING:tensorflow:From <ipython-input-3-bcd5b9f9e911>:62: softmax_cross_entropy_with_logits (from tensorflow.python.ops.nn_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "\n",
      "Future major versions of TensorFlow will allow gradients to flow\n",
      "into the labels input on backprop by default.\n",
      "\n",
      "See @{tf.nn.softmax_cross_entropy_with_logits_v2}.\n",
      "\n",
      "WARNING:tensorflow:From <ipython-input-3-bcd5b9f9e911>:76: arg_max (from tensorflow.python.ops.gen_math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use `argmax` instead\n",
      "epoch=  0  cost= 119.1739  mse= 11719.692987837077  Train_accuracy= 0.45180723\n",
      "epoch=  1  cost= 3.5320482  mse= 23.401057859313294  Train_accuracy= 0.5481928\n",
      "epoch=  2  cost= 5.8227086  mse= 31.33302269454754  Train_accuracy= 0.45180723\n",
      "epoch=  3  cost= 0.7560767  mse= 0.4882731914296032  Train_accuracy= 0.5481928\n",
      "epoch=  4  cost= 0.69489497  mse= 0.32365135050813404  Train_accuracy= 0.560241\n",
      "epoch=  5  cost= 0.6828029  mse= 0.31576335334490463  Train_accuracy= 0.58433735\n",
      "epoch=  6  cost= 0.67804736  mse= 0.31030004420737795  Train_accuracy= 0.5903614\n",
      "epoch=  7  cost= 0.6747134  mse= 0.31131719251447393  Train_accuracy= 0.57831323\n",
      "epoch=  8  cost= 0.67192054  mse= 0.3172458785902289  Train_accuracy= 0.560241\n",
      "epoch=  9  cost= 0.6698908  mse= 0.3192726857309402  Train_accuracy= 0.58433735\n",
      "epoch=  10  cost= 0.6680153  mse= 0.3220571198922252  Train_accuracy= 0.58433735\n",
      "epoch=  11  cost= 0.66613066  mse= 0.3238889275495139  Train_accuracy= 0.60240966\n",
      "epoch=  12  cost= 0.66418445  mse= 0.3271342873091262  Train_accuracy= 0.60240966\n",
      "epoch=  13  cost= 0.6621548  mse= 0.32925429166864023  Train_accuracy= 0.6084337\n",
      "epoch=  14  cost= 0.6598099  mse= 0.3341424281803621  Train_accuracy= 0.6204819\n",
      "epoch=  15  cost= 0.6573939  mse= 0.3370070958069481  Train_accuracy= 0.6385542\n",
      "epoch=  16  cost= 0.6548795  mse= 0.3389420597860437  Train_accuracy= 0.6566265\n",
      "epoch=  17  cost= 0.65145797  mse= 0.339469805456137  Train_accuracy= 0.6686747\n",
      "epoch=  18  cost= 0.6470872  mse= 0.3510695133702051  Train_accuracy= 0.6626506\n",
      "epoch=  19  cost= 0.6439803  mse= 0.3496385071204474  Train_accuracy= 0.6686747\n",
      "epoch=  20  cost= 0.64104605  mse= 0.35478880753060116  Train_accuracy= 0.6686747\n",
      "epoch=  21  cost= 0.63742167  mse= 0.36331087392093003  Train_accuracy= 0.6686747\n",
      "epoch=  22  cost= 0.6337747  mse= 0.370922231091115  Train_accuracy= 0.686747\n",
      "epoch=  23  cost= 0.6297488  mse= 0.3799476806641919  Train_accuracy= 0.6927711\n",
      "epoch=  24  cost= 0.6254061  mse= 0.39163666010200776  Train_accuracy= 0.686747\n",
      "epoch=  25  cost= 0.6223252  mse= 0.4175722522990988  Train_accuracy= 0.6807229\n",
      "epoch=  26  cost= 0.62026525  mse= 0.3907788746596302  Train_accuracy= 0.71686745\n",
      "epoch=  27  cost= 0.6220174  mse= 0.47046601429862533  Train_accuracy= 0.6506024\n",
      "epoch=  28  cost= 0.6316551  mse= 0.4109025572272598  Train_accuracy= 0.67469877\n",
      "epoch=  29  cost= 0.65855896  mse= 0.6451599675328185  Train_accuracy= 0.57228917\n",
      "epoch=  30  cost= 0.68834215  mse= 0.4285185679743887  Train_accuracy= 0.5180723\n",
      "epoch=  31  cost= 0.6124286  mse= 0.46711146925458263  Train_accuracy= 0.686747\n",
      "epoch=  32  cost= 0.606451  mse= 0.42838532812359853  Train_accuracy= 0.7409639\n",
      "epoch=  33  cost= 0.607051  mse= 0.49829256072056455  Train_accuracy= 0.6807229\n",
      "epoch=  34  cost= 0.6161381  mse= 0.4441106301244682  Train_accuracy= 0.70481926\n",
      "epoch=  35  cost= 0.63944286  mse= 0.6435173533948805  Train_accuracy= 0.59638554\n",
      "epoch=  36  cost= 0.6763077  mse= 0.5047938168091284  Train_accuracy= 0.52409637\n",
      "epoch=  37  cost= 0.60409594  mse= 0.49669540785415206  Train_accuracy= 0.6807229\n",
      "epoch=  38  cost= 0.60368127  mse= 0.4148375139943162  Train_accuracy= 0.72289157\n",
      "epoch=  39  cost= 0.6124704  mse= 0.5810736689649131  Train_accuracy= 0.64457834\n",
      "epoch=  40  cost= 0.64404225  mse= 0.49434400268141504  Train_accuracy= 0.6084337\n",
      "epoch=  41  cost= 0.6338494  mse= 0.6641961417472645  Train_accuracy= 0.60240966\n",
      "epoch=  42  cost= 0.66301006  mse= 0.46322227913657715  Train_accuracy= 0.52409637\n",
      "epoch=  43  cost= 0.58991474  mse= 0.5407968377273026  Train_accuracy= 0.6987952\n",
      "epoch=  44  cost= 0.58679503  mse= 0.48148300599821897  Train_accuracy= 0.73493975\n",
      "epoch=  45  cost= 0.59390897  mse= 0.6287863131133943  Train_accuracy= 0.67469877\n",
      "epoch=  46  cost= 0.630754  mse= 0.5779141994538376  Train_accuracy= 0.62650603\n",
      "epoch=  47  cost= 0.6698397  mse= 0.8937353890576635  Train_accuracy= 0.58433735\n",
      "epoch=  48  cost= 0.70320493  mse= 0.4888388932899972  Train_accuracy= 0.4759036\n",
      "epoch=  49  cost= 0.5876095  mse= 0.44927534343430864  Train_accuracy= 0.75301206\n",
      "epoch=  50  cost= 0.57157636  mse= 0.5225643829007499  Train_accuracy= 0.71686745\n",
      "epoch=  51  cost= 0.5658969  mse= 0.4970972757363154  Train_accuracy= 0.7409639\n",
      "epoch=  52  cost= 0.5649728  mse= 0.6435562242123001  Train_accuracy= 0.6927711\n",
      "epoch=  53  cost= 0.5911963  mse= 0.5416468763148454  Train_accuracy= 0.6927711\n",
      "epoch=  54  cost= 0.6859366  mse= 1.1123756495412183  Train_accuracy= 0.56626505\n",
      "epoch=  55  cost= 0.76908123  mse= 0.5464748030641977  Train_accuracy= 0.45180723\n",
      "epoch=  56  cost= 0.65397644  mse= 0.30808647257200883  Train_accuracy= 0.6204819\n",
      "epoch=  57  cost= 0.5902284  mse= 0.3620138342496351  Train_accuracy= 0.73493975\n",
      "epoch=  58  cost= 0.57168126  mse= 0.4163971782356411  Train_accuracy= 0.73493975\n",
      "epoch=  59  cost= 0.5617304  mse= 0.43451289793622344  Train_accuracy= 0.74698794\n",
      "epoch=  60  cost= 0.5546364  mse= 0.560158496460908  Train_accuracy= 0.7409639\n",
      "epoch=  61  cost= 0.55043703  mse= 0.5892280438131956  Train_accuracy= 0.79518074\n",
      "epoch=  62  cost= 0.561263  mse= 0.7969331234317168  Train_accuracy= 0.70481926\n",
      "epoch=  63  cost= 0.62671906  mse= 0.6936033042257259  Train_accuracy= 0.6084337\n",
      "epoch=  64  cost= 0.6828082  mse= 1.1051053948375393  Train_accuracy= 0.58433735\n",
      "epoch=  65  cost= 0.7442306  mse= 0.49555452244212006  Train_accuracy= 0.45180723\n",
      "epoch=  66  cost= 0.64726835  mse= 0.3317196313898877  Train_accuracy= 0.6204819\n",
      "epoch=  67  cost= 0.5801047  mse= 0.36410712775306114  Train_accuracy= 0.72289157\n",
      "epoch=  68  cost= 0.552893  mse= 0.4501789807122659  Train_accuracy= 0.73493975\n",
      "epoch=  69  cost= 0.53909814  mse= 0.44378100149687927  Train_accuracy= 0.74698794\n",
      "epoch=  70  cost= 0.5331201  mse= 0.6046577651329653  Train_accuracy= 0.72289157\n",
      "epoch=  71  cost= 0.56038564  mse= 0.5331409895294966  Train_accuracy= 0.72289157\n",
      "epoch=  72  cost= 0.65799296  mse= 1.0973547780527741  Train_accuracy= 0.59638554\n",
      "epoch=  73  cost= 0.7662795  mse= 0.5822056576471253  Train_accuracy= 0.4698795\n",
      "epoch=  74  cost= 0.63135064  mse= 0.3082583692567588  Train_accuracy= 0.70481926\n",
      "epoch=  75  cost= 0.5660318  mse= 0.3625216085251844  Train_accuracy= 0.7409639\n",
      "epoch=  76  cost= 0.5437362  mse= 0.37884704862874646  Train_accuracy= 0.76506025\n",
      "epoch=  77  cost= 0.52733773  mse= 0.4516594423516963  Train_accuracy= 0.76506025\n",
      "epoch=  78  cost= 0.5120729  mse= 0.5820886174241051  Train_accuracy= 0.7710843\n",
      "epoch=  79  cost= 0.50066096  mse= 0.6450178681142854  Train_accuracy= 0.7710843\n",
      "epoch=  80  cost= 0.4907107  mse= 0.7349653761433628  Train_accuracy= 0.76506025\n",
      "epoch=  81  cost= 0.4907746  mse= 0.7815007710846715  Train_accuracy= 0.8012048\n",
      "epoch=  82  cost= 0.55568874  mse= 1.239202626522677  Train_accuracy= 0.6927711\n",
      "epoch=  83  cost= 0.8785146  mse= 1.3365003759682443  Train_accuracy= 0.45180723\n",
      "epoch=  84  cost= 0.5633733  mse= 0.47828696356662803  Train_accuracy= 0.6987952\n",
      "epoch=  85  cost= 0.5347318  mse= 0.36572231714964804  Train_accuracy= 0.76506025\n",
      "epoch=  86  cost= 0.5211154  mse= 0.4076613130488616  Train_accuracy= 0.76506025\n",
      "epoch=  87  cost= 0.5157293  mse= 0.3234573062083947  Train_accuracy= 0.78313255\n",
      "epoch=  88  cost= 0.5174466  mse= 0.5062494294143454  Train_accuracy= 0.76506025\n",
      "epoch=  89  cost= 0.5670271  mse= 0.30069255571045284  Train_accuracy= 0.72289157\n",
      "epoch=  90  cost= 0.65349394  mse= 1.1063737556059476  Train_accuracy= 0.61445785\n",
      "epoch=  91  cost= 0.7775316  mse= 0.4549783559377609  Train_accuracy= 0.46385542\n",
      "epoch=  92  cost= 0.6887935  mse= 0.3707254021407437  Train_accuracy= 0.5120482\n",
      "epoch=  93  cost= 0.65757906  mse= 0.3219432195882829  Train_accuracy= 0.6084337\n",
      "epoch=  94  cost= 0.6334422  mse= 0.3231141345596953  Train_accuracy= 0.64457834\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch=  95  cost= 0.61519206  mse= 0.3408916391464711  Train_accuracy= 0.6626506\n",
      "epoch=  96  cost= 0.58535504  mse= 0.4139226448573218  Train_accuracy= 0.686747\n",
      "epoch=  97  cost= 0.5602775  mse= 0.6224820843808617  Train_accuracy= 0.6807229\n",
      "epoch=  98  cost= 0.5462489  mse= 0.48714424943222123  Train_accuracy= 0.75301206\n",
      "epoch=  99  cost= 0.5365363  mse= 0.6224652521015389  Train_accuracy= 0.7108434\n",
      "epoch=  100  cost= 0.53343654  mse= 0.45757483813269184  Train_accuracy= 0.7710843\n",
      "epoch=  101  cost= 0.55024284  mse= 0.7711735824761813  Train_accuracy= 0.7108434\n",
      "epoch=  102  cost= 0.6096052  mse= 0.5770922803853835  Train_accuracy= 0.6325301\n",
      "epoch=  103  cost= 0.5970668  mse= 0.8798593759741009  Train_accuracy= 0.6807229\n",
      "epoch=  104  cost= 0.64399725  mse= 0.49802899196457867  Train_accuracy= 0.57228917\n",
      "epoch=  105  cost= 0.5115494  mse= 0.4187368277370638  Train_accuracy= 0.76506025\n",
      "epoch=  106  cost= 0.49042204  mse= 0.471492244342724  Train_accuracy= 0.7590361\n",
      "epoch=  107  cost= 0.4819179  mse= 0.5112428619468206  Train_accuracy= 0.7710843\n",
      "epoch=  108  cost= 0.4971556  mse= 0.7933277345238939  Train_accuracy= 0.74698794\n",
      "epoch=  109  cost= 0.64411217  mse= 0.9049075731215197  Train_accuracy= 0.59638554\n",
      "epoch=  110  cost= 0.6478973  mse= 1.4967706749428007  Train_accuracy= 0.62650603\n",
      "epoch=  111  cost= 0.69480085  mse= 0.4042549196054929  Train_accuracy= 0.5060241\n",
      "epoch=  112  cost= 0.5996069  mse= 0.3388596368798832  Train_accuracy= 0.6566265\n",
      "epoch=  113  cost= 0.5097167  mse= 0.39498046654809843  Train_accuracy= 0.77710843\n",
      "epoch=  114  cost= 0.49884984  mse= 0.5923217093707349  Train_accuracy= 0.7289157\n",
      "epoch=  115  cost= 0.5209012  mse= 0.4989671102958391  Train_accuracy= 0.7891566\n",
      "epoch=  116  cost= 0.56543434  mse= 0.8264510054188223  Train_accuracy= 0.6927711\n",
      "epoch=  117  cost= 0.6448938  mse= 0.6532894949948096  Train_accuracy= 0.57831323\n",
      "epoch=  118  cost= 0.5021279  mse= 0.3865354422332008  Train_accuracy= 0.76506025\n",
      "epoch=  119  cost= 0.47706035  mse= 0.37004649445276855  Train_accuracy= 0.79518074\n",
      "epoch=  120  cost= 0.45648047  mse= 0.40827241083810656  Train_accuracy= 0.8012048\n",
      "epoch=  121  cost= 0.44750586  mse= 0.5019063873308465  Train_accuracy= 0.79518074\n",
      "epoch=  122  cost= 0.46708018  mse= 0.5566713417892447  Train_accuracy= 0.82530123\n",
      "epoch=  123  cost= 0.5963485  mse= 1.1129801584513548  Train_accuracy= 0.67469877\n",
      "epoch=  124  cost= 0.8255048  mse= 1.0149698252254762  Train_accuracy= 0.5060241\n",
      "epoch=  125  cost= 0.62034315  mse= 0.35327683720204023  Train_accuracy= 0.6626506\n",
      "epoch=  126  cost= 0.52788997  mse= 0.32177025858142566  Train_accuracy= 0.73493975\n",
      "epoch=  127  cost= 0.50281984  mse= 0.5688217773648829  Train_accuracy= 0.7590361\n",
      "epoch=  128  cost= 0.5018401  mse= 0.32028253833996984  Train_accuracy= 0.8373494\n",
      "epoch=  129  cost= 0.46431613  mse= 0.6537036563044817  Train_accuracy= 0.78313255\n",
      "epoch=  130  cost= 0.49471688  mse= 0.48656115976503117  Train_accuracy= 0.7891566\n",
      "epoch=  131  cost= 0.47116852  mse= 1.1168439886878359  Train_accuracy= 0.76506025\n",
      "epoch=  132  cost= 0.5813128  mse= 0.7615206523777325  Train_accuracy= 0.6506024\n",
      "epoch=  133  cost= 0.49172133  mse= 0.9161841822049429  Train_accuracy= 0.74698794\n",
      "epoch=  134  cost= 0.56580096  mse= 0.6433800040142961  Train_accuracy= 0.67469877\n",
      "epoch=  135  cost= 0.44413382  mse= 0.8441730223775628  Train_accuracy= 0.7710843\n",
      "epoch=  136  cost= 0.4714692  mse= 0.6303608998144893  Train_accuracy= 0.79518074\n",
      "epoch=  137  cost= 0.4801205  mse= 1.1530057415257953  Train_accuracy= 0.7590361\n",
      "epoch=  138  cost= 0.63348144  mse= 1.0706872606691338  Train_accuracy= 0.57831323\n",
      "epoch=  139  cost= 0.4674536  mse= 0.7124531825985856  Train_accuracy= 0.7590361\n",
      "epoch=  140  cost= 0.47598496  mse= 0.41030548508625564  Train_accuracy= 0.8373494\n",
      "epoch=  141  cost= 0.4252813  mse= 0.8859221710659497  Train_accuracy= 0.7891566\n",
      "epoch=  142  cost= 0.45510727  mse= 0.7073052769355702  Train_accuracy= 0.813253\n",
      "epoch=  143  cost= 0.48081613  mse= 1.2398963674150574  Train_accuracy= 0.75301206\n",
      "epoch=  144  cost= 0.65969616  mse= 1.0845044612841466  Train_accuracy= 0.56626505\n",
      "epoch=  145  cost= 0.45283034  mse= 0.5599725122190735  Train_accuracy= 0.78313255\n",
      "epoch=  146  cost= 0.42764002  mse= 0.46579303909061004  Train_accuracy= 0.813253\n",
      "epoch=  147  cost= 0.38907087  mse= 0.7479199810675319  Train_accuracy= 0.8313253\n",
      "epoch=  148  cost= 0.3837194  mse= 0.6924712180013636  Train_accuracy= 0.8795181\n",
      "epoch=  149  cost= 0.40091255  mse= 1.141104224122578  Train_accuracy= 0.7891566\n",
      "epoch=  150  cost= 0.5511545  mse= 1.13258788849239  Train_accuracy= 0.67469877\n",
      "epoch=  151  cost= 0.671324  mse= 1.8555989075905277  Train_accuracy= 0.6325301\n",
      "epoch=  152  cost= 0.7372649  mse= 0.7956248780001958  Train_accuracy= 0.48795182\n",
      "epoch=  153  cost= 0.6318011  mse= 0.5140099558923023  Train_accuracy= 0.53614455\n",
      "epoch=  154  cost= 0.60798556  mse= 0.44007687418819125  Train_accuracy= 0.6325301\n",
      "epoch=  155  cost= 0.5655821  mse= 0.4955615988035133  Train_accuracy= 0.6927711\n",
      "epoch=  156  cost= 0.5463344  mse= 0.43675932235454085  Train_accuracy= 0.7289157\n",
      "epoch=  157  cost= 0.5119918  mse= 0.4860784016395694  Train_accuracy= 0.73493975\n",
      "epoch=  158  cost= 0.4734435  mse= 0.5204790545569831  Train_accuracy= 0.76506025\n",
      "epoch=  159  cost= 0.44389197  mse= 0.6474644169288268  Train_accuracy= 0.79518074\n",
      "epoch=  160  cost= 0.41351378  mse= 0.7201095901631882  Train_accuracy= 0.82530123\n",
      "epoch=  161  cost= 0.3917093  mse= 0.7523092275614484  Train_accuracy= 0.82530123\n",
      "epoch=  162  cost= 0.3855301  mse= 0.8285391315751572  Train_accuracy= 0.82530123\n",
      "epoch=  163  cost= 0.39832687  mse= 0.7723547831794073  Train_accuracy= 0.85542166\n",
      "epoch=  164  cost= 0.4934723  mse= 1.2456234828709671  Train_accuracy= 0.74698794\n",
      "epoch=  165  cost= 0.72752464  mse= 1.3027130110399332  Train_accuracy= 0.5903614\n",
      "epoch=  166  cost= 0.53481007  mse= 0.8309552939190011  Train_accuracy= 0.71686745\n",
      "epoch=  167  cost= 0.50591666  mse= 0.36674331045933906  Train_accuracy= 0.8012048\n",
      "epoch=  168  cost= 0.47829795  mse= 0.48484922374293404  Train_accuracy= 0.77710843\n",
      "epoch=  169  cost= 0.45282534  mse= 0.32763877331823665  Train_accuracy= 0.82530123\n",
      "epoch=  170  cost= 0.43168336  mse= 0.552679641266099  Train_accuracy= 0.7891566\n",
      "epoch=  171  cost= 0.4179641  mse= 0.5000767966732452  Train_accuracy= 0.8795181\n",
      "epoch=  172  cost= 0.41051686  mse= 0.8019010593922402  Train_accuracy= 0.7891566\n",
      "epoch=  173  cost= 0.46618527  mse= 0.7529115328878365  Train_accuracy= 0.79518074\n",
      "epoch=  174  cost= 0.4604259  mse= 1.0599241478667965  Train_accuracy= 0.78313255\n",
      "epoch=  175  cost= 0.5876456  mse= 0.9232111031817467  Train_accuracy= 0.61445785\n",
      "epoch=  176  cost= 0.40373015  mse= 0.5721560006461802  Train_accuracy= 0.8313253\n",
      "epoch=  177  cost= 0.37792802  mse= 0.5155609479003698  Train_accuracy= 0.8614458\n",
      "epoch=  178  cost= 0.34477738  mse= 0.7523657548319465  Train_accuracy= 0.8433735\n",
      "epoch=  179  cost= 0.33796382  mse= 0.7554969744435451  Train_accuracy= 0.8915663\n",
      "epoch=  180  cost= 0.3210827  mse= 0.9500918441126793  Train_accuracy= 0.8433735\n",
      "epoch=  181  cost= 0.3158948  mse= 0.9548161692748998  Train_accuracy= 0.8795181\n",
      "epoch=  182  cost= 0.30571678  mse= 1.1199554650514931  Train_accuracy= 0.8493976\n",
      "epoch=  183  cost= 0.31160444  mse= 1.0600397771506753  Train_accuracy= 0.90361446\n",
      "epoch=  184  cost= 0.33338985  mse= 1.4637554946023468  Train_accuracy= 0.8313253\n",
      "epoch=  185  cost= 0.6036562  mse= 1.724624320505661  Train_accuracy= 0.6686747\n",
      "epoch=  186  cost= 1.6027987  mse= 6.653395013907605  Train_accuracy= 0.55421686\n",
      "epoch=  187  cost= 0.71906656  mse= 0.5105431680698803  Train_accuracy= 0.5421687\n",
      "epoch=  188  cost= 0.71189797  mse= 1.001930225792672  Train_accuracy= 0.55421686\n",
      "epoch=  189  cost= 0.7334535  mse= 0.6927741092113445  Train_accuracy= 0.45783132\n",
      "epoch=  190  cost= 0.62483203  mse= 0.3492195116790795  Train_accuracy= 0.6385542\n",
      "epoch=  191  cost= 0.5863179  mse= 0.4621155410450191  Train_accuracy= 0.6626506\n",
      "epoch=  192  cost= 0.571396  mse= 0.5505515066541985  Train_accuracy= 0.6686747\n",
      "epoch=  193  cost= 0.55855054  mse= 0.5668574970061981  Train_accuracy= 0.6807229\n",
      "epoch=  194  cost= 0.5438763  mse= 0.6192952823484748  Train_accuracy= 0.686747\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch=  195  cost= 0.5307848  mse= 0.674527456246417  Train_accuracy= 0.6987952\n",
      "epoch=  196  cost= 0.52442616  mse= 0.664979610599548  Train_accuracy= 0.6927711\n",
      "epoch=  197  cost= 0.51883733  mse= 0.6727780085129285  Train_accuracy= 0.6987952\n",
      "epoch=  198  cost= 0.51292646  mse= 0.6802873303525104  Train_accuracy= 0.72289157\n",
      "epoch=  199  cost= 0.507025  mse= 0.6827061703743619  Train_accuracy= 0.72289157\n",
      "epoch=  200  cost= 0.5002668  mse= 0.7067523426433074  Train_accuracy= 0.71686745\n",
      "epoch=  201  cost= 0.49359226  mse= 0.7219941638317693  Train_accuracy= 0.7289157\n",
      "epoch=  202  cost= 0.48831582  mse= 0.8032498513234041  Train_accuracy= 0.7289157\n",
      "epoch=  203  cost= 0.490192  mse= 0.768256839075599  Train_accuracy= 0.76506025\n",
      "epoch=  204  cost= 0.50022835  mse= 1.0870517828633282  Train_accuracy= 0.7289157\n",
      "epoch=  205  cost= 0.53958035  mse= 0.8004955559158491  Train_accuracy= 0.6686747\n",
      "epoch=  206  cost= 0.506015  mse= 1.0238580734825677  Train_accuracy= 0.71686745\n",
      "epoch=  207  cost= 0.46408266  mse= 1.1236864247526188  Train_accuracy= 0.79518074\n",
      "epoch=  208  cost= 0.47849867  mse= 1.531282955804176  Train_accuracy= 0.78313255\n",
      "epoch=  209  cost= 0.53440446  mse= 1.1122291604084706  Train_accuracy= 0.72289157\n",
      "epoch=  210  cost= 0.5441113  mse= 1.9261200867567758  Train_accuracy= 0.6987952\n",
      "epoch=  211  cost= 0.46198267  mse= 1.2340398735363867  Train_accuracy= 0.79518074\n",
      "epoch=  212  cost= 0.5153891  mse= 2.2468794665163148  Train_accuracy= 0.7710843\n",
      "epoch=  213  cost= 0.5495966  mse= 1.0680730004548555  Train_accuracy= 0.6807229\n",
      "epoch=  214  cost= 0.5093654  mse= 1.6366302332444564  Train_accuracy= 0.74698794\n",
      "epoch=  215  cost= 0.47521728  mse= 1.5384288486708404  Train_accuracy= 0.7710843\n",
      "epoch=  216  cost= 0.49563605  mse= 2.5379785798199364  Train_accuracy= 0.74698794\n",
      "epoch=  217  cost= 0.47233465  mse= 1.9005551143711268  Train_accuracy= 0.79518074\n",
      "epoch=  218  cost= 0.57759327  mse= 3.608955553070681  Train_accuracy= 0.7108434\n",
      "epoch=  219  cost= 0.5865018  mse= 1.210773797650515  Train_accuracy= 0.6325301\n",
      "epoch=  220  cost= 0.5607856  mse= 1.5642804135131902  Train_accuracy= 0.70481926\n",
      "epoch=  221  cost= 0.45481426  mse= 1.2298492239745187  Train_accuracy= 0.76506025\n",
      "epoch=  222  cost= 0.42676595  mse= 1.8925235769883568  Train_accuracy= 0.813253\n",
      "epoch=  223  cost= 0.3948163  mse= 2.0295091399225256  Train_accuracy= 0.8072289\n",
      "epoch=  224  cost= 0.40422982  mse= 2.699495608647735  Train_accuracy= 0.8192771\n",
      "epoch=  225  cost= 0.44977143  mse= 1.661017885966095  Train_accuracy= 0.7891566\n",
      "epoch=  226  cost= 0.5154038  mse= 4.261963158529564  Train_accuracy= 0.7590361\n",
      "epoch=  227  cost= 0.65707695  mse= 1.8271477168172783  Train_accuracy= 0.6325301\n",
      "epoch=  228  cost= 0.7066862  mse= 2.2071463200211094  Train_accuracy= 0.58433735\n",
      "epoch=  229  cost= 0.6132644  mse= 0.45632560354564333  Train_accuracy= 0.6506024\n",
      "epoch=  230  cost= 0.5859842  mse= 0.9402520935285453  Train_accuracy= 0.64457834\n",
      "epoch=  231  cost= 0.59693325  mse= 0.3837897457989573  Train_accuracy= 0.7108434\n",
      "epoch=  232  cost= 0.52108455  mse= 0.7530355740448129  Train_accuracy= 0.77710843\n",
      "epoch=  233  cost= 0.4547041  mse= 0.8548056322058735  Train_accuracy= 0.8313253\n",
      "epoch=  234  cost= 0.43595904  mse= 0.8517017767327512  Train_accuracy= 0.813253\n",
      "epoch=  235  cost= 0.45754325  mse= 2.1666359633831527  Train_accuracy= 0.77710843\n",
      "epoch=  236  cost= 0.5577125  mse= 1.3465477662966834  Train_accuracy= 0.70481926\n",
      "epoch=  237  cost= 0.50752693  mse= 1.367682520553757  Train_accuracy= 0.7710843\n",
      "epoch=  238  cost= 0.38936844  mse= 1.1227899504565457  Train_accuracy= 0.85542166\n",
      "epoch=  239  cost= 0.3776345  mse= 1.690670430335849  Train_accuracy= 0.8373494\n",
      "epoch=  240  cost= 0.40041366  mse= 1.3469926758443707  Train_accuracy= 0.8433735\n",
      "epoch=  241  cost= 0.40876365  mse= 2.4611336161355206  Train_accuracy= 0.8072289\n",
      "epoch=  242  cost= 0.5367181  mse= 1.5327775939685375  Train_accuracy= 0.72289157\n",
      "epoch=  243  cost= 0.58905333  mse= 2.3213119018360993  Train_accuracy= 0.6626506\n",
      "epoch=  244  cost= 0.3968611  mse= 0.8896578201912407  Train_accuracy= 0.87349397\n",
      "epoch=  245  cost= 0.3711328  mse= 1.35823026071349  Train_accuracy= 0.8373494\n",
      "epoch=  246  cost= 0.3814408  mse= 1.349780596434719  Train_accuracy= 0.8373494\n",
      "epoch=  247  cost= 0.4074603  mse= 2.4256503646769803  Train_accuracy= 0.8072289\n",
      "epoch=  248  cost= 0.5061671  mse= 1.4743490868814693  Train_accuracy= 0.72289157\n",
      "epoch=  249  cost= 0.5106668  mse= 1.8087227948286164  Train_accuracy= 0.71686745\n",
      "epoch=  250  cost= 0.38995215  mse= 1.2697513954810584  Train_accuracy= 0.8373494\n",
      "epoch=  251  cost= 0.3885748  mse= 1.5534288645609682  Train_accuracy= 0.82530123\n",
      "epoch=  252  cost= 0.36232737  mse= 1.1428758531273486  Train_accuracy= 0.8674699\n",
      "epoch=  253  cost= 0.37528595  mse= 2.2426069839068754  Train_accuracy= 0.813253\n",
      "epoch=  254  cost= 0.41333434  mse= 1.364078035421328  Train_accuracy= 0.7891566\n",
      "epoch=  255  cost= 0.36575168  mse= 1.7727382650025414  Train_accuracy= 0.8493976\n",
      "epoch=  256  cost= 0.33345118  mse= 1.5786676740289856  Train_accuracy= 0.8915663\n",
      "epoch=  257  cost= 0.4134829  mse= 2.507167864733294  Train_accuracy= 0.8192771\n",
      "epoch=  258  cost= 0.4455558  mse= 1.4416542370864993  Train_accuracy= 0.7710843\n",
      "epoch=  259  cost= 0.4511109  mse= 1.9478593982443924  Train_accuracy= 0.7891566\n",
      "epoch=  260  cost= 0.40942433  mse= 1.3160621955381475  Train_accuracy= 0.82530123\n",
      "epoch=  261  cost= 0.41396475  mse= 1.901954186517243  Train_accuracy= 0.8072289\n",
      "epoch=  262  cost= 0.3596933  mse= 1.1565996582725564  Train_accuracy= 0.87349397\n",
      "epoch=  263  cost= 0.32876697  mse= 1.5478335159871304  Train_accuracy= 0.8373494\n",
      "epoch=  264  cost= 0.3492745  mse= 1.5360749958063857  Train_accuracy= 0.8674699\n",
      "epoch=  265  cost= 0.38632396  mse= 2.574800149523681  Train_accuracy= 0.8192771\n",
      "epoch=  266  cost= 0.49022827  mse= 1.887877798901123  Train_accuracy= 0.73493975\n",
      "epoch=  267  cost= 0.62288415  mse= 2.0241549309815006  Train_accuracy= 0.6506024\n",
      "epoch=  268  cost= 0.4920216  mse= 0.6943425617438735  Train_accuracy= 0.77710843\n",
      "epoch=  269  cost= 0.3599307  mse= 1.2994055815030774  Train_accuracy= 0.82530123\n",
      "epoch=  270  cost= 0.39217842  mse= 1.2705684370668822  Train_accuracy= 0.7891566\n",
      "epoch=  271  cost= 0.3718421  mse= 2.318716341416955  Train_accuracy= 0.8192771\n",
      "epoch=  272  cost= 0.41975087  mse= 1.6839709119019415  Train_accuracy= 0.7710843\n",
      "epoch=  273  cost= 0.39247334  mse= 2.0171816759347387  Train_accuracy= 0.813253\n",
      "epoch=  274  cost= 0.38007233  mse= 1.4370324501482212  Train_accuracy= 0.82530123\n",
      "epoch=  275  cost= 0.3404492  mse= 1.6913893550533763  Train_accuracy= 0.82530123\n",
      "epoch=  276  cost= 0.33033967  mse= 1.793584653717667  Train_accuracy= 0.8915663\n",
      "epoch=  277  cost= 0.37002754  mse= 2.6115211471131077  Train_accuracy= 0.813253\n",
      "epoch=  278  cost= 0.3879297  mse= 1.8304564152715246  Train_accuracy= 0.813253\n",
      "epoch=  279  cost= 0.36995107  mse= 1.835746610663046  Train_accuracy= 0.82530123\n",
      "epoch=  280  cost= 0.34832582  mse= 1.560179633244683  Train_accuracy= 0.87349397\n",
      "epoch=  281  cost= 0.37814417  mse= 2.492461068817216  Train_accuracy= 0.8192771\n",
      "epoch=  282  cost= 0.39293686  mse= 2.0210316399863135  Train_accuracy= 0.8192771\n",
      "epoch=  283  cost= 0.522191  mse= 2.1012515438045534  Train_accuracy= 0.7409639\n",
      "epoch=  284  cost= 0.41402385  mse= 1.0738467934457467  Train_accuracy= 0.8072289\n",
      "epoch=  285  cost= 0.35770503  mse= 1.9677071433699462  Train_accuracy= 0.813253\n",
      "epoch=  286  cost= 0.38211313  mse= 1.6670402404042552  Train_accuracy= 0.7710843\n",
      "epoch=  287  cost= 0.3098824  mse= 1.598152242936418  Train_accuracy= 0.88554215\n",
      "epoch=  288  cost= 0.31988692  mse= 2.2326182004493336  Train_accuracy= 0.88554215\n",
      "epoch=  289  cost= 0.42815825  mse= 2.1970569448355084  Train_accuracy= 0.8373494\n",
      "epoch=  290  cost= 0.35695896  mse= 2.0308049674412296  Train_accuracy= 0.8493976\n",
      "epoch=  291  cost= 0.38977808  mse= 2.208843364943902  Train_accuracy= 0.8072289\n",
      "epoch=  292  cost= 0.3575575  mse= 1.6429458196532423  Train_accuracy= 0.8313253\n",
      "epoch=  293  cost= 0.28456125  mse= 1.6358823904417796  Train_accuracy= 0.87349397\n",
      "epoch=  294  cost= 0.26248223  mse= 2.338509815718185  Train_accuracy= 0.90361446\n",
      "epoch=  295  cost= 0.3346507  mse= 2.5128781898193826  Train_accuracy= 0.8614458\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch=  296  cost= 0.3176507  mse= 2.52680118358562  Train_accuracy= 0.87349397\n",
      "epoch=  297  cost= 0.43661848  mse= 2.820151401282845  Train_accuracy= 0.7891566\n",
      "epoch=  298  cost= 0.42859012  mse= 2.142667547857441  Train_accuracy= 0.76506025\n",
      "epoch=  299  cost= 0.44711345  mse= 1.104436480233129  Train_accuracy= 0.7710843\n",
      "epoch=  300  cost= 0.26949  mse= 1.7843627140277853  Train_accuracy= 0.9096386\n",
      "epoch=  301  cost= 0.2337831  mse= 2.059375652090042  Train_accuracy= 0.92168677\n",
      "epoch=  302  cost= 0.21419157  mse= 2.7131102253297876  Train_accuracy= 0.92771083\n",
      "epoch=  303  cost= 0.20922826  mse= 2.662237725341598  Train_accuracy= 0.93373495\n",
      "epoch=  304  cost= 0.2041309  mse= 2.8731445048502815  Train_accuracy= 0.93373495\n",
      "epoch=  305  cost= 0.20366694  mse= 2.6477711043037355  Train_accuracy= 0.92771083\n",
      "epoch=  306  cost= 0.21820381  mse= 3.0617220246709516  Train_accuracy= 0.91566265\n",
      "epoch=  307  cost= 0.316812  mse= 3.128953764897446  Train_accuracy= 0.8433735\n",
      "epoch=  308  cost= 0.8262561  mse= 5.824332310016296  Train_accuracy= 0.64457834\n",
      "epoch=  309  cost= 0.7900932  mse= 2.2093036383123295  Train_accuracy= 0.4698795\n",
      "epoch=  310  cost= 0.47349498  mse= 1.3322372426625981  Train_accuracy= 0.79518074\n",
      "epoch=  311  cost= 0.44281074  mse= 1.7909260388576247  Train_accuracy= 0.813253\n",
      "epoch=  312  cost= 0.4993153  mse= 1.1853525508181453  Train_accuracy= 0.75301206\n",
      "epoch=  313  cost= 0.48460034  mse= 2.3259879452804264  Train_accuracy= 0.7409639\n",
      "epoch=  314  cost= 0.6650518  mse= 1.2690688681419555  Train_accuracy= 0.57831323\n",
      "epoch=  315  cost= 0.3913156  mse= 0.6804489812825655  Train_accuracy= 0.8313253\n",
      "epoch=  316  cost= 0.3260379  mse= 1.2603974491529313  Train_accuracy= 0.85542166\n",
      "epoch=  317  cost= 0.29403362  mse= 1.7460012786283157  Train_accuracy= 0.85542166\n",
      "epoch=  318  cost= 0.2772777  mse= 2.2624722677181914  Train_accuracy= 0.88554215\n",
      "epoch=  319  cost= 0.26729658  mse= 2.6567634689020787  Train_accuracy= 0.88554215\n",
      "epoch=  320  cost= 0.259445  mse= 2.9652751481788804  Train_accuracy= 0.8795181\n",
      "epoch=  321  cost= 0.25222284  mse= 3.04736960990622  Train_accuracy= 0.8795181\n",
      "epoch=  322  cost= 0.25002918  mse= 3.376184229190562  Train_accuracy= 0.8795181\n",
      "epoch=  323  cost= 0.26007572  mse= 3.2111697627847153  Train_accuracy= 0.8614458\n",
      "epoch=  324  cost= 0.30060482  mse= 4.016436993043335  Train_accuracy= 0.8493976\n",
      "epoch=  325  cost= 0.36461085  mse= 3.000358236459429  Train_accuracy= 0.8373494\n",
      "epoch=  326  cost= 0.5298313  mse= 4.8085985653292065  Train_accuracy= 0.70481926\n",
      "epoch=  327  cost= 0.8618858  mse= 3.0767689316592617  Train_accuracy= 0.59638554\n",
      "epoch=  328  cost= 0.4603925  mse= 0.7044265188674349  Train_accuracy= 0.85542166\n",
      "epoch=  329  cost= 0.40534127  mse= 0.9723499706875349  Train_accuracy= 0.8915663\n",
      "epoch=  330  cost= 0.37590733  mse= 1.409908266691301  Train_accuracy= 0.87349397\n",
      "epoch=  331  cost= 0.35293248  mse= 1.1968785845593932  Train_accuracy= 0.8915663\n",
      "epoch=  332  cost= 0.329005  mse= 1.4956510645061223  Train_accuracy= 0.9096386\n",
      "epoch=  333  cost= 0.31238595  mse= 1.245231308966107  Train_accuracy= 0.90361446\n",
      "epoch=  334  cost= 0.30767354  mse= 1.98372269458489  Train_accuracy= 0.88554215\n",
      "epoch=  335  cost= 0.3745737  mse= 1.047583939122527  Train_accuracy= 0.8433735\n",
      "epoch=  336  cost= 0.32048258  mse= 2.8405615969286044  Train_accuracy= 0.8433735\n",
      "epoch=  337  cost= 0.42962623  mse= 1.467502079301751  Train_accuracy= 0.77710843\n",
      "epoch=  338  cost= 0.32070878  mse= 2.5321010946992937  Train_accuracy= 0.8795181\n",
      "epoch=  339  cost= 0.27670678  mse= 1.6094840515679072  Train_accuracy= 0.8674699\n",
      "epoch=  340  cost= 0.25054225  mse= 3.177847970397093  Train_accuracy= 0.9096386\n",
      "epoch=  341  cost= 0.37014464  mse= 1.6759214997441683  Train_accuracy= 0.82530123\n",
      "epoch=  342  cost= 0.28528985  mse= 3.5114206898490385  Train_accuracy= 0.87349397\n",
      "epoch=  343  cost= 0.32805887  mse= 2.427139083629016  Train_accuracy= 0.8433735\n",
      "epoch=  344  cost= 0.4714685  mse= 4.342487567032709  Train_accuracy= 0.7409639\n",
      "epoch=  345  cost= 0.8930664  mse= 2.705063417170336  Train_accuracy= 0.58433735\n",
      "epoch=  346  cost= 0.5288991  mse= 0.4356972481998331  Train_accuracy= 0.7710843\n",
      "epoch=  347  cost= 0.40246344  mse= 1.0310375036955346  Train_accuracy= 0.85542166\n",
      "epoch=  348  cost= 0.38190207  mse= 1.6683953053365665  Train_accuracy= 0.85542166\n",
      "epoch=  349  cost= 0.35199592  mse= 2.6015659280674783  Train_accuracy= 0.82530123\n",
      "epoch=  350  cost= 0.38648292  mse= 3.2303389816949046  Train_accuracy= 0.8072289\n",
      "epoch=  351  cost= 0.35256028  mse= 1.383521566047599  Train_accuracy= 0.8373494\n",
      "epoch=  352  cost= 0.24870689  mse= 3.2073921556060725  Train_accuracy= 0.90361446\n",
      "epoch=  353  cost= 0.23856913  mse= 2.4570380262722713  Train_accuracy= 0.87349397\n",
      "epoch=  354  cost= 0.22983856  mse= 4.317077706050126  Train_accuracy= 0.92168677\n",
      "epoch=  355  cost= 0.252031  mse= 3.185495078737029  Train_accuracy= 0.8674699\n",
      "epoch=  356  cost= 0.25073135  mse= 4.9348254810768415  Train_accuracy= 0.89759034\n",
      "epoch=  357  cost= 0.41026357  mse= 3.4807804333855965  Train_accuracy= 0.7891566\n",
      "epoch=  358  cost= 0.38099968  mse= 4.540214799435908  Train_accuracy= 0.8072289\n",
      "epoch=  359  cost= 0.43553534  mse= 2.153779831397467  Train_accuracy= 0.75301206\n",
      "epoch=  360  cost= 0.24477881  mse= 2.0724607175879965  Train_accuracy= 0.939759\n",
      "epoch=  361  cost= 0.19134888  mse= 3.1564227422452014  Train_accuracy= 0.939759\n",
      "epoch=  362  cost= 0.17728251  mse= 3.163417196368092  Train_accuracy= 0.92771083\n",
      "epoch=  363  cost= 0.18305522  mse= 4.394499412415926  Train_accuracy= 0.92771083\n",
      "epoch=  364  cost= 0.24553825  mse= 3.2219366473109985  Train_accuracy= 0.88554215\n",
      "epoch=  365  cost= 0.25135496  mse= 5.1605265779612  Train_accuracy= 0.8915663\n",
      "epoch=  366  cost= 0.4730924  mse= 4.122694748875554  Train_accuracy= 0.7710843\n",
      "epoch=  367  cost= 0.38351846  mse= 3.5086140845261675  Train_accuracy= 0.8192771\n",
      "epoch=  368  cost= 0.32713062  mse= 1.4585484515348308  Train_accuracy= 0.85542166\n",
      "epoch=  369  cost= 0.24173783  mse= 3.3141416929588736  Train_accuracy= 0.89759034\n",
      "epoch=  370  cost= 0.31992465  mse= 2.0689377146939596  Train_accuracy= 0.8493976\n",
      "epoch=  371  cost= 0.19895631  mse= 3.5207108788926047  Train_accuracy= 0.9518072\n",
      "epoch=  372  cost= 0.18365966  mse= 3.063434591700048  Train_accuracy= 0.91566265\n",
      "epoch=  373  cost= 0.17608166  mse= 4.359223352784726  Train_accuracy= 0.94578314\n",
      "epoch=  374  cost= 0.23268266  mse= 3.401112129178671  Train_accuracy= 0.8915663\n",
      "epoch=  375  cost= 0.23160836  mse= 5.215741198538514  Train_accuracy= 0.8915663\n",
      "epoch=  376  cost= 0.41008502  mse= 3.7618157742528253  Train_accuracy= 0.77710843\n",
      "epoch=  377  cost= 0.3383613  mse= 4.279768921956277  Train_accuracy= 0.8373494\n",
      "epoch=  378  cost= 0.3115584  mse= 2.4580772723380013  Train_accuracy= 0.8373494\n",
      "epoch=  379  cost= 0.24655202  mse= 3.610327683458854  Train_accuracy= 0.92168677\n",
      "epoch=  380  cost= 0.31320468  mse= 2.6082524737077515  Train_accuracy= 0.8373494\n",
      "epoch=  381  cost= 0.19584134  mse= 3.8930505712065013  Train_accuracy= 0.9518072\n",
      "epoch=  382  cost= 0.17322055  mse= 3.6231496986245846  Train_accuracy= 0.92168677\n",
      "epoch=  383  cost= 0.1729143  mse= 4.810657541611038  Train_accuracy= 0.939759\n",
      "epoch=  384  cost= 0.245686  mse= 3.7042153127958892  Train_accuracy= 0.88554215\n",
      "epoch=  385  cost= 0.22716112  mse= 5.317278828269599  Train_accuracy= 0.89759034\n",
      "epoch=  386  cost= 0.36561242  mse= 3.8892293052888283  Train_accuracy= 0.8012048\n",
      "epoch=  387  cost= 0.32764015  mse= 4.043544112951493  Train_accuracy= 0.8313253\n",
      "epoch=  388  cost= 0.2518143  mse= 2.471867755431705  Train_accuracy= 0.8674699\n",
      "epoch=  389  cost= 0.19750133  mse= 4.253455464214245  Train_accuracy= 0.92771083\n",
      "epoch=  390  cost= 0.28750148  mse= 2.66237738962841  Train_accuracy= 0.8614458\n",
      "epoch=  391  cost= 0.19073862  mse= 4.7556935214357186  Train_accuracy= 0.939759\n",
      "epoch=  392  cost= 0.17737567  mse= 3.6518736256602873  Train_accuracy= 0.9096386\n",
      "epoch=  393  cost= 0.15322544  mse= 4.736128019553563  Train_accuracy= 0.9518072\n",
      "epoch=  394  cost= 0.17434128  mse= 4.3948832826271635  Train_accuracy= 0.91566265\n",
      "epoch=  395  cost= 0.1853506  mse= 5.503463197449393  Train_accuracy= 0.92168677\n",
      "epoch=  396  cost= 0.35538313  mse= 4.558179682483019  Train_accuracy= 0.8192771\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch=  397  cost= 0.38274124  mse= 5.027379415586792  Train_accuracy= 0.8072289\n",
      "epoch=  398  cost= 0.30210844  mse= 2.962297532707033  Train_accuracy= 0.82530123\n",
      "epoch=  399  cost= 0.29032153  mse= 4.812611802460448  Train_accuracy= 0.8433735\n",
      "epoch=  400  cost= 0.6153259  mse= 2.98583354650536  Train_accuracy= 0.64457834\n",
      "epoch=  401  cost= 0.2717613  mse= 2.978132533417867  Train_accuracy= 0.92168677\n",
      "epoch=  402  cost= 0.20428771  mse= 3.48543048846626  Train_accuracy= 0.9518072\n",
      "epoch=  403  cost= 0.16529277  mse= 4.154172950937772  Train_accuracy= 0.94578314\n",
      "epoch=  404  cost= 0.14394385  mse= 4.771631968265153  Train_accuracy= 0.9518072\n",
      "epoch=  405  cost= 0.13324845  mse= 5.288022549363687  Train_accuracy= 0.9518072\n",
      "epoch=  406  cost= 0.12380099  mse= 5.751619253498251  Train_accuracy= 0.9518072\n",
      "epoch=  407  cost= 0.118015036  mse= 5.811359725288751  Train_accuracy= 0.9518072\n",
      "epoch=  408  cost= 0.11394636  mse= 5.932179665347914  Train_accuracy= 0.9578313\n",
      "epoch=  409  cost= 0.11115547  mse= 5.980050709166098  Train_accuracy= 0.9578313\n",
      "epoch=  410  cost= 0.10876728  mse= 6.058330330360834  Train_accuracy= 0.9578313\n",
      "epoch=  411  cost= 0.10622444  mse= 6.156592796944629  Train_accuracy= 0.96385545\n",
      "epoch=  412  cost= 0.10403773  mse= 6.254367718140721  Train_accuracy= 0.96385545\n",
      "epoch=  413  cost= 0.10198447  mse= 6.31803205296937  Train_accuracy= 0.96385545\n",
      "epoch=  414  cost= 0.0999113  mse= 6.480455360357856  Train_accuracy= 0.97590363\n",
      "epoch=  415  cost= 0.09785936  mse= 6.522752671449153  Train_accuracy= 0.96385545\n",
      "epoch=  416  cost= 0.1014697  mse= 6.767179869771948  Train_accuracy= 0.9819277\n",
      "epoch=  417  cost= 0.12373536  mse= 6.384084065458735  Train_accuracy= 0.93373495\n",
      "epoch=  418  cost= 0.21024548  mse= 7.866834930337295  Train_accuracy= 0.87349397\n",
      "epoch=  419  cost= 0.8946716  mse= 8.922958889662292  Train_accuracy= 0.6686747\n",
      "epoch=  420  cost= 0.617861  mse= 4.608163119347564  Train_accuracy= 0.6927711\n",
      "epoch=  421  cost= 0.46344572  mse= 1.2203747492066643  Train_accuracy= 0.74698794\n",
      "epoch=  422  cost= 0.3372406  mse= 2.6720413885687817  Train_accuracy= 0.85542166\n",
      "epoch=  423  cost= 0.31329656  mse= 1.7479162456518864  Train_accuracy= 0.8373494\n",
      "epoch=  424  cost= 0.30167708  mse= 3.514767893798014  Train_accuracy= 0.8493976\n",
      "epoch=  425  cost= 0.37768716  mse= 1.7643969019033379  Train_accuracy= 0.8072289\n",
      "epoch=  426  cost= 0.26074404  mse= 2.7827692506440846  Train_accuracy= 0.89759034\n",
      "epoch=  427  cost= 0.23570348  mse= 2.5843384673610217  Train_accuracy= 0.8614458\n",
      "epoch=  428  cost= 0.25824794  mse= 4.6128214972471335  Train_accuracy= 0.87349397\n",
      "epoch=  429  cost= 0.40649313  mse= 2.5019558687337122  Train_accuracy= 0.813253\n",
      "epoch=  430  cost= 0.27393365  mse= 3.507054387765759  Train_accuracy= 0.87349397\n",
      "epoch=  431  cost= 0.22173479  mse= 2.8320128888440843  Train_accuracy= 0.91566265\n",
      "epoch=  432  cost= 0.22384927  mse= 5.170018218167409  Train_accuracy= 0.91566265\n",
      "epoch=  433  cost= 0.36287284  mse= 2.9151504795154684  Train_accuracy= 0.8192771\n",
      "epoch=  434  cost= 0.2481239  mse= 4.620183464967784  Train_accuracy= 0.90361446\n",
      "epoch=  435  cost= 0.26690662  mse= 2.592386421776488  Train_accuracy= 0.8795181\n",
      "epoch=  436  cost= 0.16171342  mse= 4.437310729195835  Train_accuracy= 0.9698795\n",
      "epoch=  437  cost= 0.15248333  mse= 3.9445963714649666  Train_accuracy= 0.92771083\n",
      "epoch=  438  cost= 0.15904024  mse= 6.082772427262461  Train_accuracy= 0.939759\n",
      "epoch=  439  cost= 0.32219455  mse= 3.9731737114499177  Train_accuracy= 0.8433735\n",
      "epoch=  440  cost= 0.23394687  mse= 6.0865551047149395  Train_accuracy= 0.91566265\n",
      "epoch=  441  cost= 0.31711653  mse= 2.8827538580676073  Train_accuracy= 0.8433735\n",
      "epoch=  442  cost= 0.18263805  mse= 3.9863787479711172  Train_accuracy= 0.93373495\n",
      "epoch=  443  cost= 0.13203298  mse= 4.143570186293726  Train_accuracy= 0.96385545\n",
      "epoch=  444  cost= 0.11385557  mse= 5.545944511069999  Train_accuracy= 0.9578313\n",
      "epoch=  445  cost= 0.10920769  mse= 5.042871975922664  Train_accuracy= 0.96385545\n",
      "epoch=  446  cost= 0.11233063  mse= 6.393293149202238  Train_accuracy= 0.96385545\n",
      "epoch=  447  cost= 0.14510845  mse= 4.734234892997198  Train_accuracy= 0.94578314\n",
      "epoch=  448  cost= 0.16936123  mse= 7.40374140508404  Train_accuracy= 0.92168677\n",
      "epoch=  449  cost= 0.47456494  mse= 4.6488440235446244  Train_accuracy= 0.78313255\n",
      "epoch=  450  cost= 0.29384995  mse= 6.098720478389039  Train_accuracy= 0.8614458\n",
      "epoch=  451  cost= 0.21305059  mse= 3.128311136278535  Train_accuracy= 0.8915663\n",
      "epoch=  452  cost= 0.14950913  mse= 4.441653622591125  Train_accuracy= 0.96385545\n",
      "epoch=  453  cost= 0.111627415  mse= 4.772126908939483  Train_accuracy= 0.9578313\n",
      "epoch=  454  cost= 0.10415538  mse= 5.885341356914565  Train_accuracy= 0.9819277\n",
      "epoch=  455  cost= 0.118735716  mse= 4.806529007164787  Train_accuracy= 0.94578314\n",
      "epoch=  456  cost= 0.15250383  mse= 7.252906599211951  Train_accuracy= 0.94578314\n",
      "epoch=  457  cost= 0.38755062  mse= 4.346113206044253  Train_accuracy= 0.813253\n",
      "epoch=  458  cost= 0.20343542  mse= 5.005771474408901  Train_accuracy= 0.92168677\n",
      "epoch=  459  cost= 0.16412377  mse= 4.508474192674292  Train_accuracy= 0.92771083\n",
      "epoch=  460  cost= 0.17359951  mse= 6.526377799858316  Train_accuracy= 0.93373495\n",
      "epoch=  461  cost= 0.28769588  mse= 4.131234517288001  Train_accuracy= 0.8493976\n",
      "epoch=  462  cost= 0.15229607  mse= 5.129614342969539  Train_accuracy= 0.94578314\n",
      "epoch=  463  cost= 0.12532705  mse= 4.897300134874497  Train_accuracy= 0.94578314\n",
      "epoch=  464  cost= 0.12720472  mse= 6.467806009941897  Train_accuracy= 0.9518072\n",
      "epoch=  465  cost= 0.21285325  mse= 4.736267292848434  Train_accuracy= 0.8915663\n",
      "epoch=  466  cost= 0.12269819  mse= 5.964409685885919  Train_accuracy= 0.9698795\n",
      "epoch=  467  cost= 0.13197987  mse= 4.906988003525707  Train_accuracy= 0.93373495\n",
      "epoch=  468  cost= 0.10523067  mse= 6.49219088633696  Train_accuracy= 0.9819277\n",
      "epoch=  469  cost= 0.14254342  mse= 5.303473610822605  Train_accuracy= 0.92168677\n",
      "epoch=  470  cost= 0.12559952  mse= 7.105322966290533  Train_accuracy= 0.9518072\n",
      "epoch=  471  cost= 0.24073906  mse= 5.387073524126972  Train_accuracy= 0.87349397\n",
      "epoch=  472  cost= 0.1661735  mse= 6.770945979738051  Train_accuracy= 0.9096386\n",
      "epoch=  473  cost= 0.24801055  mse= 5.176365717288014  Train_accuracy= 0.8795181\n",
      "epoch=  474  cost= 0.15739846  mse= 5.443041742922994  Train_accuracy= 0.939759\n",
      "epoch=  475  cost= 0.11259002  mse= 4.8973297170705745  Train_accuracy= 0.96385545\n",
      "epoch=  476  cost= 0.08707309  mse= 6.343629468194284  Train_accuracy= 0.97590363\n",
      "epoch=  477  cost= 0.094363526  mse= 5.161381309199599  Train_accuracy= 0.9698795\n",
      "epoch=  478  cost= 0.09298747  mse= 7.256693653949983  Train_accuracy= 0.96385545\n",
      "epoch=  479  cost= 0.11821673  mse= 5.224603953476035  Train_accuracy= 0.96385545\n",
      "epoch=  480  cost= 0.10098584  mse= 7.771608569865973  Train_accuracy= 0.96385545\n",
      "epoch=  481  cost= 0.14582442  mse= 5.14021576287757  Train_accuracy= 0.939759\n",
      "epoch=  482  cost= 0.12540564  mse= 8.413489747859217  Train_accuracy= 0.9578313\n",
      "epoch=  483  cost= 0.24737448  mse= 4.759165669612969  Train_accuracy= 0.8674699\n",
      "epoch=  484  cost= 0.13777499  mse= 6.957716913048035  Train_accuracy= 0.9698795\n",
      "epoch=  485  cost= 0.08340846  mse= 6.327702385145882  Train_accuracy= 0.9698795\n",
      "epoch=  486  cost= 0.07579114  mse= 7.012422127427358  Train_accuracy= 0.9819277\n",
      "epoch=  487  cost= 0.072997876  mse= 6.412391444857856  Train_accuracy= 0.97590363\n",
      "epoch=  488  cost= 0.06547746  mse= 7.082617389456963  Train_accuracy= 0.9879518\n",
      "epoch=  489  cost= 0.06350875  mse= 6.747584674719117  Train_accuracy= 0.9819277\n",
      "epoch=  490  cost= 0.05759403  mse= 7.252344922292169  Train_accuracy= 0.9879518\n",
      "epoch=  491  cost= 0.055970598  mse= 7.036675736663066  Train_accuracy= 0.9879518\n",
      "epoch=  492  cost= 0.054751616  mse= 7.614488404248083  Train_accuracy= 0.9879518\n",
      "epoch=  493  cost= 0.055708684  mse= 7.159204791758552  Train_accuracy= 0.9879518\n",
      "epoch=  494  cost= 0.054856353  mse= 7.857371377433868  Train_accuracy= 0.9879518\n",
      "epoch=  495  cost= 0.058559023  mse= 7.251006822009282  Train_accuracy= 0.9879518\n",
      "epoch=  496  cost= 0.052021313  mse= 7.855999159468185  Train_accuracy= 0.9879518\n",
      "epoch=  497  cost= 0.05338084  mse= 7.432411437222366  Train_accuracy= 0.9879518\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch=  498  cost= 0.053033873  mse= 8.257597258413796  Train_accuracy= 0.9879518\n",
      "epoch=  499  cost= 0.063735224  mse= 7.45958145122699  Train_accuracy= 0.9879518\n",
      "epoch=  500  cost= 0.056596756  mse= 8.339489982080295  Train_accuracy= 0.9879518\n",
      "epoch=  501  cost= 0.06656984  mse= 7.470068658676726  Train_accuracy= 0.9879518\n",
      "epoch=  502  cost= 0.05905148  mse= 8.394869275333116  Train_accuracy= 0.9879518\n",
      "epoch=  503  cost= 0.08222232  mse= 7.336978601677139  Train_accuracy= 0.9879518\n",
      "epoch=  504  cost= 0.09849796  mse= 9.243985303947191  Train_accuracy= 0.9698795\n",
      "epoch=  505  cost= 0.31783983  mse= 7.344038817406248  Train_accuracy= 0.8433735\n",
      "epoch=  506  cost= 0.50031924  mse= 12.349030938515078  Train_accuracy= 0.7891566\n",
      "epoch=  507  cost= 1.2795311  mse= 10.110840068143224  Train_accuracy= 0.560241\n",
      "epoch=  508  cost= 0.9776006  mse= 2.0105467712189085  Train_accuracy= 0.53614455\n",
      "epoch=  509  cost= 1.609752  mse= 11.186920928661621  Train_accuracy= 0.48795182\n",
      "epoch=  510  cost= 0.6214639  mse= 0.45045569720342454  Train_accuracy= 0.62650603\n",
      "epoch=  511  cost= 0.5943844  mse= 0.40357717538832344  Train_accuracy= 0.70481926\n",
      "epoch=  512  cost= 0.57063264  mse= 0.3682672655667996  Train_accuracy= 0.7289157\n",
      "epoch=  513  cost= 0.5496953  mse= 0.35290441344007284  Train_accuracy= 0.76506025\n",
      "epoch=  514  cost= 0.53104454  mse= 0.36133320490220444  Train_accuracy= 0.76506025\n",
      "epoch=  515  cost= 0.50950956  mse= 0.39610550970241254  Train_accuracy= 0.78313255\n",
      "epoch=  516  cost= 0.48670995  mse= 0.4321913147691667  Train_accuracy= 0.7891566\n",
      "epoch=  517  cost= 0.46136594  mse= 0.5181870125475785  Train_accuracy= 0.8072289\n",
      "epoch=  518  cost= 0.4374443  mse= 0.6309569638555396  Train_accuracy= 0.8433735\n",
      "epoch=  519  cost= 0.4085683  mse= 0.8413354292062311  Train_accuracy= 0.8313253\n",
      "epoch=  520  cost= 0.4320099  mse= 1.6296959150500907  Train_accuracy= 0.8192771\n",
      "epoch=  521  cost= 0.61507577  mse= 1.3702874767724562  Train_accuracy= 0.67469877\n",
      "epoch=  522  cost= 0.6585058  mse= 2.006345092668097  Train_accuracy= 0.5481928\n",
      "epoch=  523  cost= 0.48216307  mse= 0.39500723311479746  Train_accuracy= 0.71686745\n",
      "epoch=  524  cost= 0.42532924  mse= 0.45786312897629433  Train_accuracy= 0.82530123\n",
      "epoch=  525  cost= 0.36848393  mse= 0.6644220035433408  Train_accuracy= 0.8614458\n",
      "epoch=  526  cost= 0.30566868  mse= 1.018234307103579  Train_accuracy= 0.89759034\n",
      "epoch=  527  cost= 0.26502374  mse= 1.6535784504900786  Train_accuracy= 0.89759034\n",
      "epoch=  528  cost= 0.24989197  mse= 1.5597048519134344  Train_accuracy= 0.89759034\n",
      "epoch=  529  cost= 0.25028625  mse= 2.3249658248090372  Train_accuracy= 0.92168677\n",
      "epoch=  530  cost= 0.34319884  mse= 1.8505079682390875  Train_accuracy= 0.8192771\n",
      "epoch=  531  cost= 0.60723805  mse= 5.531420250437032  Train_accuracy= 0.6686747\n",
      "epoch=  532  cost= 1.0051863  mse= 4.274835272087257  Train_accuracy= 0.6506024\n",
      "epoch=  533  cost= 0.77588737  mse= 4.429723936917657  Train_accuracy= 0.61445785\n",
      "epoch=  534  cost= 0.5449926  mse= 0.40337366247189266  Train_accuracy= 0.6566265\n",
      "epoch=  535  cost= 0.43830186  mse= 0.4335180309861327  Train_accuracy= 0.7891566\n",
      "epoch=  536  cost= 0.3806503  mse= 0.5811331849497728  Train_accuracy= 0.82530123\n",
      "epoch=  537  cost= 0.33977655  mse= 0.7501575469183451  Train_accuracy= 0.8433735\n",
      "epoch=  538  cost= 0.29587027  mse= 1.0160837566585663  Train_accuracy= 0.90361446\n",
      "epoch=  539  cost= 0.24829991  mse= 1.43096489622105  Train_accuracy= 0.92168677\n",
      "epoch=  540  cost= 0.2124471  mse= 1.9292990724119647  Train_accuracy= 0.91566265\n",
      "epoch=  541  cost= 0.18966752  mse= 2.372912858103502  Train_accuracy= 0.93373495\n",
      "epoch=  542  cost= 0.17747094  mse= 2.4570189376385785  Train_accuracy= 0.93373495\n",
      "epoch=  543  cost= 0.16982397  mse= 2.8994247732810132  Train_accuracy= 0.939759\n",
      "epoch=  544  cost= 0.17742409  mse= 2.4811710769423705  Train_accuracy= 0.92771083\n",
      "epoch=  545  cost= 0.22652782  mse= 4.016495929046139  Train_accuracy= 0.89759034\n",
      "epoch=  546  cost= 0.4812509  mse= 3.1357380777045503  Train_accuracy= 0.78313255\n",
      "epoch=  547  cost= 0.7755727  mse= 7.576974114682974  Train_accuracy= 0.686747\n",
      "epoch=  548  cost= 1.1928592  mse= 4.518530030559293  Train_accuracy= 0.55421686\n",
      "epoch=  549  cost= 0.48907328  mse= 0.34219632496085817  Train_accuracy= 0.8674699\n",
      "epoch=  550  cost= 0.39947775  mse= 0.680305253585459  Train_accuracy= 0.8373494\n",
      "epoch=  551  cost= 0.3490114  mse= 0.7693484286074378  Train_accuracy= 0.8795181\n",
      "epoch=  552  cost= 0.32012334  mse= 1.0426415394021729  Train_accuracy= 0.85542166\n",
      "epoch=  553  cost= 0.29776895  mse= 1.286958391659984  Train_accuracy= 0.8915663\n",
      "epoch=  554  cost= 0.2774779  mse= 1.612341043639727  Train_accuracy= 0.90361446\n",
      "epoch=  555  cost= 0.25257647  mse= 2.0427033915755506  Train_accuracy= 0.91566265\n",
      "epoch=  556  cost= 0.2321758  mse= 2.5433008418082017  Train_accuracy= 0.91566265\n",
      "epoch=  557  cost= 0.22191606  mse= 2.55619549530652  Train_accuracy= 0.93373495\n",
      "epoch=  558  cost= 0.21022457  mse= 2.882935087278223  Train_accuracy= 0.92771083\n",
      "epoch=  559  cost= 0.20043598  mse= 2.969106276063581  Train_accuracy= 0.94578314\n",
      "epoch=  560  cost= 0.19378632  mse= 2.989682345624413  Train_accuracy= 0.94578314\n",
      "epoch=  561  cost= 0.18714662  mse= 3.1069815066974336  Train_accuracy= 0.94578314\n",
      "epoch=  562  cost= 0.17706949  mse= 3.444670691039132  Train_accuracy= 0.9518072\n",
      "epoch=  563  cost= 0.16988671  mse= 3.5772446486529295  Train_accuracy= 0.94578314\n",
      "epoch=  564  cost= 0.16441329  mse= 3.7300794839678404  Train_accuracy= 0.9518072\n",
      "epoch=  565  cost= 0.16769944  mse= 4.164152337337971  Train_accuracy= 0.94578314\n",
      "epoch=  566  cost= 0.1943099  mse= 3.541463598860623  Train_accuracy= 0.91566265\n",
      "epoch=  567  cost= 0.24904405  mse= 5.269207113671463  Train_accuracy= 0.8795181\n",
      "epoch=  568  cost= 0.5207229  mse= 4.76649728143209  Train_accuracy= 0.7710843\n",
      "epoch=  569  cost= 0.6826456  mse= 6.3013228440429385  Train_accuracy= 0.67469877\n",
      "epoch=  570  cost= 0.6448888  mse= 3.150276923302388  Train_accuracy= 0.62650603\n",
      "epoch=  571  cost= 0.40697685  mse= 0.5541025909242249  Train_accuracy= 0.88554215\n",
      "epoch=  572  cost= 0.32778814  mse= 0.877503678398246  Train_accuracy= 0.90361446\n",
      "epoch=  573  cost= 0.28513122  mse= 1.2633858659619484  Train_accuracy= 0.93373495\n",
      "epoch=  574  cost= 0.2501198  mse= 1.6374317095466293  Train_accuracy= 0.939759\n",
      "epoch=  575  cost= 0.21022703  mse= 2.3268810238422324  Train_accuracy= 0.94578314\n",
      "epoch=  576  cost= 0.18998069  mse= 2.9614020635591323  Train_accuracy= 0.94578314\n",
      "epoch=  577  cost= 0.17406558  mse= 3.2559579834967307  Train_accuracy= 0.9518072\n",
      "epoch=  578  cost= 0.15960504  mse= 3.5085684656730276  Train_accuracy= 0.9518072\n",
      "epoch=  579  cost= 0.14733863  mse= 3.6083641802937363  Train_accuracy= 0.9518072\n",
      "epoch=  580  cost= 0.13713855  mse= 3.8226409264189805  Train_accuracy= 0.9578313\n",
      "epoch=  581  cost= 0.12886979  mse= 4.033307504147087  Train_accuracy= 0.96385545\n",
      "epoch=  582  cost= 0.122896515  mse= 4.249496268160912  Train_accuracy= 0.9698795\n",
      "epoch=  583  cost= 0.1187341  mse= 4.273860038489665  Train_accuracy= 0.9698795\n",
      "epoch=  584  cost= 0.11526589  mse= 4.599590165768413  Train_accuracy= 0.96385545\n",
      "epoch=  585  cost= 0.11474069  mse= 4.3774982966738465  Train_accuracy= 0.9698795\n",
      "epoch=  586  cost= 0.120984696  mse= 5.198216587052247  Train_accuracy= 0.9518072\n",
      "epoch=  587  cost= 0.15836103  mse= 4.431451804874407  Train_accuracy= 0.939759\n",
      "epoch=  588  cost= 0.18397218  mse= 5.998191066426967  Train_accuracy= 0.91566265\n",
      "epoch=  589  cost= 0.38840246  mse= 5.224549870036814  Train_accuracy= 0.82530123\n",
      "epoch=  590  cost= 0.47133908  mse= 6.959148473606839  Train_accuracy= 0.7710843\n",
      "epoch=  591  cost= 0.88073456  mse= 6.7476359855251  Train_accuracy= 0.61445785\n",
      "epoch=  592  cost= 0.45120046  mse= 1.1650123161286732  Train_accuracy= 0.7289157\n",
      "epoch=  593  cost= 0.318486  mse= 1.209713828285564  Train_accuracy= 0.90361446\n",
      "epoch=  594  cost= 0.22270331  mse= 1.4394511133756795  Train_accuracy= 0.939759\n",
      "epoch=  595  cost= 0.17934647  mse= 1.9333483029713245  Train_accuracy= 0.939759\n",
      "epoch=  596  cost= 0.14839603  mse= 2.6217141946153153  Train_accuracy= 0.9578313\n",
      "epoch=  597  cost= 0.13313101  mse= 3.0665911509782466  Train_accuracy= 0.9578313\n",
      "epoch=  598  cost= 0.12292748  mse= 3.5268739314549924  Train_accuracy= 0.9578313\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch=  599  cost= 0.11264559  mse= 3.940941664892383  Train_accuracy= 0.9518072\n",
      "epoch=  600  cost= 0.10376206  mse= 4.432753884678354  Train_accuracy= 0.9578313\n",
      "epoch=  601  cost= 0.0991035  mse= 4.5833461323969145  Train_accuracy= 0.96385545\n",
      "epoch=  602  cost= 0.095054425  mse= 4.915541959836657  Train_accuracy= 0.96385545\n",
      "epoch=  603  cost= 0.091980234  mse= 5.010288712866705  Train_accuracy= 0.9698795\n",
      "epoch=  604  cost= 0.08847706  mse= 5.243669026867587  Train_accuracy= 0.9698795\n",
      "epoch=  605  cost= 0.08607365  mse= 5.181291770760192  Train_accuracy= 0.9698795\n",
      "epoch=  606  cost= 0.08270517  mse= 5.485157611890227  Train_accuracy= 0.97590363\n",
      "epoch=  607  cost= 0.0810294  mse= 5.446578157934032  Train_accuracy= 0.97590363\n",
      "epoch=  608  cost= 0.08039109  mse= 5.871311964149743  Train_accuracy= 0.97590363\n",
      "epoch=  609  cost= 0.08235428  mse= 5.4969372729228585  Train_accuracy= 0.9698795\n",
      "epoch=  610  cost= 0.087826505  mse= 6.320279894420927  Train_accuracy= 0.97590363\n",
      "epoch=  611  cost= 0.1117605  mse= 5.3503255045943146  Train_accuracy= 0.9578313\n",
      "epoch=  612  cost= 0.15686387  mse= 7.643014585582843  Train_accuracy= 0.92771083\n",
      "epoch=  613  cost= 0.40513417  mse= 6.202193083989869  Train_accuracy= 0.85542166\n",
      "epoch=  614  cost= 0.4895418  mse= 8.689072182722665  Train_accuracy= 0.7710843\n",
      "epoch=  615  cost= 0.97746074  mse= 8.45312688558298  Train_accuracy= 0.6204819\n",
      "epoch=  616  cost= 0.5020021  mse= 1.9355625966219776  Train_accuracy= 0.6987952\n",
      "epoch=  617  cost= 0.25018978  mse= 2.2291000607909295  Train_accuracy= 0.94578314\n",
      "epoch=  618  cost= 0.1813278  mse= 2.807785639826783  Train_accuracy= 0.9578313\n",
      "epoch=  619  cost= 0.14278337  mse= 3.6490555990236366  Train_accuracy= 0.9698795\n",
      "epoch=  620  cost= 0.121900775  mse= 4.340173082098058  Train_accuracy= 0.97590363\n",
      "epoch=  621  cost= 0.10926166  mse= 4.621292771332232  Train_accuracy= 0.97590363\n",
      "epoch=  622  cost= 0.10163591  mse= 5.14698160748721  Train_accuracy= 0.97590363\n",
      "epoch=  623  cost= 0.096081674  mse= 5.2187316259594665  Train_accuracy= 0.9698795\n",
      "epoch=  624  cost= 0.091894194  mse= 5.580782999417832  Train_accuracy= 0.97590363\n",
      "epoch=  625  cost= 0.088273965  mse= 5.526190741713578  Train_accuracy= 0.9698795\n",
      "epoch=  626  cost= 0.0852536  mse= 5.896600316778906  Train_accuracy= 0.97590363\n",
      "epoch=  627  cost= 0.08217903  mse= 5.794076734743548  Train_accuracy= 0.9698795\n",
      "epoch=  628  cost= 0.080772586  mse= 6.263175772818753  Train_accuracy= 0.9819277\n",
      "epoch=  629  cost= 0.08209996  mse= 5.864495765244024  Train_accuracy= 0.9698795\n",
      "epoch=  630  cost= 0.082862005  mse= 6.811993614221605  Train_accuracy= 0.9879518\n",
      "epoch=  631  cost= 0.09250401  mse= 5.839972956130665  Train_accuracy= 0.96385545\n",
      "epoch=  632  cost= 0.093206435  mse= 7.476532458892815  Train_accuracy= 0.9819277\n",
      "epoch=  633  cost= 0.13258116  mse= 5.569649406632319  Train_accuracy= 0.93373495\n",
      "epoch=  634  cost= 0.1441379  mse= 8.149835942975264  Train_accuracy= 0.94578314\n",
      "epoch=  635  cost= 0.30001265  mse= 5.490806278266829  Train_accuracy= 0.8493976\n",
      "epoch=  636  cost= 0.2287051  mse= 8.033737756291279  Train_accuracy= 0.89759034\n",
      "epoch=  637  cost= 0.38977608  mse= 4.770902785726326  Train_accuracy= 0.8012048\n",
      "epoch=  638  cost= 0.15612756  mse= 3.998820111600782  Train_accuracy= 0.96385545\n",
      "epoch=  639  cost= 0.10266423  mse= 4.103941937385953  Train_accuracy= 0.97590363\n",
      "epoch=  640  cost= 0.08410102  mse= 5.396616900562902  Train_accuracy= 0.9879518\n",
      "epoch=  641  cost= 0.07515379  mse= 5.8278098425946405  Train_accuracy= 0.9879518\n",
      "epoch=  642  cost= 0.0702281  mse= 6.3475951314261  Train_accuracy= 0.9879518\n",
      "epoch=  643  cost= 0.06716299  mse= 6.606939637551958  Train_accuracy= 0.9879518\n",
      "epoch=  644  cost= 0.06512377  mse= 6.874689025144971  Train_accuracy= 0.9879518\n",
      "epoch=  645  cost= 0.06342418  mse= 6.799599807697405  Train_accuracy= 0.9879518\n",
      "epoch=  646  cost= 0.06144832  mse= 7.239382273717111  Train_accuracy= 0.9939759\n",
      "epoch=  647  cost= 0.059974905  mse= 7.12467227655443  Train_accuracy= 0.9879518\n",
      "epoch=  648  cost= 0.058714133  mse= 7.471775369669568  Train_accuracy= 0.9939759\n",
      "epoch=  649  cost= 0.057433974  mse= 7.346791012486041  Train_accuracy= 0.9879518\n",
      "epoch=  650  cost= 0.056432616  mse= 7.723256842346881  Train_accuracy= 0.9939759\n",
      "epoch=  651  cost= 0.05497568  mse= 7.517619347099231  Train_accuracy= 0.9879518\n",
      "epoch=  652  cost= 0.05328844  mse= 7.944591155016277  Train_accuracy= 0.9939759\n",
      "epoch=  653  cost= 0.051825453  mse= 7.873361777016338  Train_accuracy= 0.9939759\n",
      "epoch=  654  cost= 0.050914586  mse= 8.136678813321554  Train_accuracy= 0.9939759\n",
      "epoch=  655  cost= 0.049833328  mse= 8.015529237449819  Train_accuracy= 0.9939759\n",
      "epoch=  656  cost= 0.04936973  mse= 8.430663334307077  Train_accuracy= 0.9939759\n",
      "epoch=  657  cost= 0.04936451  mse= 8.07122390010324  Train_accuracy= 0.9879518\n",
      "epoch=  658  cost= 0.04898875  mse= 8.716461258766932  Train_accuracy= 0.9939759\n",
      "epoch=  659  cost= 0.0508283  mse= 7.944279353656265  Train_accuracy= 0.9879518\n",
      "epoch=  660  cost= 0.049700696  mse= 8.962817622230315  Train_accuracy= 0.9939759\n",
      "epoch=  661  cost= 0.054416496  mse= 7.842999123077112  Train_accuracy= 0.97590363\n",
      "epoch=  662  cost= 0.055500675  mse= 9.447348828281905  Train_accuracy= 0.9939759\n",
      "epoch=  663  cost= 0.0733211  mse= 7.574829035792301  Train_accuracy= 0.97590363\n",
      "epoch=  664  cost= 0.10545125  mse= 10.993997737103898  Train_accuracy= 0.9578313\n",
      "epoch=  665  cost= 0.40311015  mse= 7.650904207708989  Train_accuracy= 0.8373494\n",
      "epoch=  666  cost= 0.6477774  mse= 14.545441983007967  Train_accuracy= 0.74698794\n",
      "epoch=  667  cost= 1.8373535  mse= 12.560160676563685  Train_accuracy= 0.5481928\n",
      "epoch=  668  cost= 0.8085929  mse= 0.7478569088412598  Train_accuracy= 0.52409637\n",
      "epoch=  669  cost= 0.5509742  mse= 0.5404085938996501  Train_accuracy= 0.67469877\n",
      "epoch=  670  cost= 0.47281072  mse= 0.7475215056019432  Train_accuracy= 0.85542166\n",
      "epoch=  671  cost= 0.3982434  mse= 1.0075253001659608  Train_accuracy= 0.8795181\n",
      "epoch=  672  cost= 0.3298063  mse= 1.4275672430455455  Train_accuracy= 0.90361446\n",
      "epoch=  673  cost= 0.28176644  mse= 1.662085762381628  Train_accuracy= 0.89759034\n",
      "epoch=  674  cost= 0.2548998  mse= 2.359928369352576  Train_accuracy= 0.93373495\n",
      "epoch=  675  cost= 0.31820354  mse= 2.101196002251018  Train_accuracy= 0.8674699\n",
      "epoch=  676  cost= 0.45203882  mse= 5.0607996758804665  Train_accuracy= 0.75301206\n",
      "epoch=  677  cost= 0.92315364  mse= 4.187658372277757  Train_accuracy= 0.64457834\n",
      "epoch=  678  cost= 0.442822  mse= 1.151331279141798  Train_accuracy= 0.8072289\n",
      "epoch=  679  cost= 0.30532357  mse= 1.1836041168161275  Train_accuracy= 0.8795181\n",
      "epoch=  680  cost= 0.2606517  mse= 1.2360497163166837  Train_accuracy= 0.92771083\n",
      "epoch=  681  cost= 0.23298636  mse= 1.5537338898537985  Train_accuracy= 0.91566265\n",
      "epoch=  682  cost= 0.20434982  mse= 1.916163541726728  Train_accuracy= 0.92771083\n",
      "epoch=  683  cost= 0.17785424  mse= 2.4248875998485735  Train_accuracy= 0.92771083\n",
      "epoch=  684  cost= 0.15807933  mse= 2.695018539893357  Train_accuracy= 0.939759\n",
      "epoch=  685  cost= 0.13335492  mse= 3.181737790777247  Train_accuracy= 0.9578313\n",
      "epoch=  686  cost= 0.112467505  mse= 3.7479400906108444  Train_accuracy= 0.9698795\n",
      "epoch=  687  cost= 0.09919226  mse= 4.110925183259401  Train_accuracy= 0.9698795\n",
      "epoch=  688  cost= 0.08915862  mse= 4.394029211296621  Train_accuracy= 0.9819277\n",
      "epoch=  689  cost= 0.08019612  mse= 4.780380969789149  Train_accuracy= 0.9819277\n",
      "epoch=  690  cost= 0.073877  mse= 4.965064905483274  Train_accuracy= 0.9879518\n",
      "epoch=  691  cost= 0.06969049  mse= 5.229348818272047  Train_accuracy= 0.9879518\n",
      "epoch=  692  cost= 0.066543914  mse= 5.402820453618929  Train_accuracy= 0.9879518\n",
      "epoch=  693  cost= 0.063560694  mse= 5.62695521559476  Train_accuracy= 0.9879518\n",
      "epoch=  694  cost= 0.061259177  mse= 5.720546836978744  Train_accuracy= 0.9879518\n",
      "epoch=  695  cost= 0.059389666  mse= 5.908476761187783  Train_accuracy= 0.9879518\n",
      "epoch=  696  cost= 0.05770819  mse= 5.987421967235617  Train_accuracy= 0.9879518\n",
      "epoch=  697  cost= 0.056113616  mse= 6.127960106022529  Train_accuracy= 0.9879518\n",
      "epoch=  698  cost= 0.054486435  mse= 6.205910099012163  Train_accuracy= 0.9879518\n",
      "epoch=  699  cost= 0.053087424  mse= 6.375583256025558  Train_accuracy= 0.9879518\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch=  700  cost= 0.051826563  mse= 6.411182424997027  Train_accuracy= 0.9879518\n",
      "epoch=  701  cost= 0.050529223  mse= 6.595552791389488  Train_accuracy= 0.9939759\n",
      "epoch=  702  cost= 0.04907021  mse= 6.637278035490574  Train_accuracy= 0.9879518\n",
      "epoch=  703  cost= 0.047805987  mse= 6.815985419257517  Train_accuracy= 0.9939759\n",
      "epoch=  704  cost= 0.04638715  mse= 6.895538807591482  Train_accuracy= 0.9939759\n",
      "epoch=  705  cost= 0.045280267  mse= 7.057662471053303  Train_accuracy= 0.9939759\n",
      "epoch=  706  cost= 0.044174172  mse= 7.064387266522933  Train_accuracy= 0.9939759\n",
      "epoch=  707  cost= 0.043178327  mse= 7.331164921296183  Train_accuracy= 0.9939759\n",
      "epoch=  708  cost= 0.04261663  mse= 7.191191003139489  Train_accuracy= 0.9939759\n",
      "epoch=  709  cost= 0.042347055  mse= 7.626151844744141  Train_accuracy= 0.9939759\n",
      "epoch=  710  cost= 0.04253167  mse= 7.226436042076017  Train_accuracy= 0.9939759\n",
      "epoch=  711  cost= 0.043664694  mse= 8.033588884303661  Train_accuracy= 0.9939759\n",
      "epoch=  712  cost= 0.049643166  mse= 7.053753086893992  Train_accuracy= 0.9819277\n",
      "epoch=  713  cost= 0.056973815  mse= 8.848137970032296  Train_accuracy= 0.9879518\n",
      "epoch=  714  cost= 0.09485843  mse= 6.6101872679396285  Train_accuracy= 0.96385545\n",
      "epoch=  715  cost= 0.18540874  mse= 11.352742026050342  Train_accuracy= 0.91566265\n",
      "epoch=  716  cost= 0.79803824  mse= 8.771499580408726  Train_accuracy= 0.74698794\n",
      "epoch=  717  cost= 1.1938577  mse= 15.583913136758955  Train_accuracy= 0.57228917\n",
      "epoch=  718  cost= 1.6448046  mse= 10.974338684204351  Train_accuracy= 0.55421686\n",
      "epoch=  719  cost= 0.64926827  mse= 0.5944798249438326  Train_accuracy= 0.59638554\n",
      "epoch=  720  cost= 0.4768969  mse= 0.43889690390854763  Train_accuracy= 0.78313255\n",
      "epoch=  721  cost= 0.4095761  mse= 0.5133476419476488  Train_accuracy= 0.82530123\n",
      "epoch=  722  cost= 0.36972108  mse= 0.6446292284610383  Train_accuracy= 0.8373494\n",
      "epoch=  723  cost= 0.323361  mse= 0.9045348835719361  Train_accuracy= 0.87349397\n",
      "epoch=  724  cost= 0.28049108  mse= 1.2315680475381383  Train_accuracy= 0.9096386\n",
      "epoch=  725  cost= 0.22609562  mse= 1.822168218499264  Train_accuracy= 0.939759\n",
      "epoch=  726  cost= 0.1884345  mse= 2.564770823058145  Train_accuracy= 0.9578313\n",
      "epoch=  727  cost= 0.1635629  mse= 3.081396140376504  Train_accuracy= 0.9578313\n",
      "epoch=  728  cost= 0.15463889  mse= 3.979420686252841  Train_accuracy= 0.9698795\n",
      "epoch=  729  cost= 0.1750746  mse= 3.2035544964265745  Train_accuracy= 0.92168677\n",
      "epoch=  730  cost= 0.36543724  mse= 7.47694759726112  Train_accuracy= 0.8012048\n",
      "epoch=  731  cost= 1.4682012  mse= 8.212182232266633  Train_accuracy= 0.60240966\n",
      "epoch=  732  cost= 0.71125454  mse= 3.852231984022685  Train_accuracy= 0.4698795\n",
      "epoch=  733  cost= 0.52426755  mse= 0.9313571589197744  Train_accuracy= 0.57831323\n",
      "epoch=  734  cost= 0.43810037  mse= 1.2921472061202284  Train_accuracy= 0.70481926\n",
      "epoch=  735  cost= 0.37844428  mse= 1.5127729599124122  Train_accuracy= 0.90361446\n",
      "epoch=  736  cost= 0.31759578  mse= 1.6179357463710493  Train_accuracy= 0.94578314\n",
      "epoch=  737  cost= 0.25493112  mse= 2.011916845562439  Train_accuracy= 0.9698795\n",
      "epoch=  738  cost= 0.20225632  mse= 2.2597147784470026  Train_accuracy= 0.94578314\n",
      "epoch=  739  cost= 0.16651778  mse= 3.2548223590057344  Train_accuracy= 0.96385545\n",
      "epoch=  740  cost= 0.14803667  mse= 3.1336492490488492  Train_accuracy= 0.96385545\n",
      "epoch=  741  cost= 0.13685125  mse= 4.483719388781042  Train_accuracy= 0.9819277\n",
      "epoch=  742  cost= 0.1348646  mse= 3.547771857750811  Train_accuracy= 0.9518072\n",
      "epoch=  743  cost= 0.13247389  mse= 5.465048128958832  Train_accuracy= 0.9819277\n",
      "epoch=  744  cost= 0.15022582  mse= 3.5347097535076606  Train_accuracy= 0.92771083\n",
      "epoch=  745  cost= 0.16322775  mse= 6.875914648788905  Train_accuracy= 0.9518072\n",
      "epoch=  746  cost= 0.28490263  mse= 3.0754041415913043  Train_accuracy= 0.85542166\n",
      "epoch=  747  cost= 0.21223447  mse= 6.694904482273004  Train_accuracy= 0.91566265\n",
      "epoch=  748  cost= 0.3038182  mse= 2.7156647269533485  Train_accuracy= 0.8493976\n",
      "epoch=  749  cost= 0.14405838  mse= 4.694553777086609  Train_accuracy= 0.9819277\n",
      "epoch=  750  cost= 0.13168854  mse= 3.2563176746970077  Train_accuracy= 0.96385545\n",
      "epoch=  751  cost= 0.09867594  mse= 5.143389859460375  Train_accuracy= 0.9879518\n",
      "epoch=  752  cost= 0.087732054  mse= 4.6742679873740345  Train_accuracy= 0.9879518\n",
      "epoch=  753  cost= 0.0827003  mse= 5.871378498027106  Train_accuracy= 0.9939759\n",
      "epoch=  754  cost= 0.079737805  mse= 5.237465957294506  Train_accuracy= 0.9879518\n",
      "epoch=  755  cost= 0.078835204  mse= 6.628590401502096  Train_accuracy= 0.9879518\n",
      "epoch=  756  cost= 0.08231702  mse= 5.33365851279838  Train_accuracy= 0.9819277\n",
      "epoch=  757  cost= 0.08486129  mse= 7.47204019441402  Train_accuracy= 0.9819277\n",
      "epoch=  758  cost= 0.10404059  mse= 5.127591247979164  Train_accuracy= 0.9698795\n",
      "epoch=  759  cost= 0.11540142  mse= 8.654759841948469  Train_accuracy= 0.96385545\n",
      "epoch=  760  cost= 0.22746003  mse= 4.496200544029121  Train_accuracy= 0.8795181\n",
      "epoch=  761  cost= 0.25014552  mse= 9.598886508401332  Train_accuracy= 0.88554215\n",
      "epoch=  762  cost= 0.5560463  mse= 4.469664831215463  Train_accuracy= 0.76506025\n",
      "epoch=  763  cost= 0.17854707  mse= 4.238088029678146  Train_accuracy= 0.97590363\n",
      "epoch=  764  cost= 0.10455443  mse= 3.9156864083881584  Train_accuracy= 0.9879518\n",
      "epoch=  765  cost= 0.07763646  mse= 5.107472501959599  Train_accuracy= 1.0\n",
      "epoch=  766  cost= 0.064555414  mse= 5.327730838415104  Train_accuracy= 1.0\n",
      "epoch=  767  cost= 0.057725992  mse= 6.143586622957774  Train_accuracy= 1.0\n",
      "epoch=  768  cost= 0.053210013  mse= 6.271899495827465  Train_accuracy= 1.0\n",
      "epoch=  769  cost= 0.050019287  mse= 6.812388769619183  Train_accuracy= 1.0\n",
      "epoch=  770  cost= 0.047432136  mse= 7.011885697330781  Train_accuracy= 1.0\n",
      "epoch=  771  cost= 0.04556992  mse= 7.396484327141216  Train_accuracy= 1.0\n",
      "epoch=  772  cost= 0.04383642  mse= 7.43938039140126  Train_accuracy= 1.0\n",
      "epoch=  773  cost= 0.04240941  mse= 7.809013375754765  Train_accuracy= 1.0\n",
      "epoch=  774  cost= 0.04115025  mse= 7.847389527755435  Train_accuracy= 1.0\n",
      "epoch=  775  cost= 0.040015113  mse= 8.102106422065836  Train_accuracy= 1.0\n",
      "epoch=  776  cost= 0.038864657  mse= 8.285874358400179  Train_accuracy= 1.0\n",
      "epoch=  777  cost= 0.037835058  mse= 8.422872832217816  Train_accuracy= 1.0\n",
      "epoch=  778  cost= 0.036904994  mse= 8.518137356960516  Train_accuracy= 1.0\n",
      "epoch=  779  cost= 0.036038682  mse= 8.655614016498529  Train_accuracy= 1.0\n",
      "epoch=  780  cost= 0.03523344  mse= 8.742479850626905  Train_accuracy= 1.0\n",
      "epoch=  781  cost= 0.034492444  mse= 8.87059396760267  Train_accuracy= 1.0\n",
      "epoch=  782  cost= 0.033813853  mse= 8.906201193043195  Train_accuracy= 1.0\n",
      "epoch=  783  cost= 0.03318901  mse= 9.003247482325554  Train_accuracy= 1.0\n",
      "epoch=  784  cost= 0.032628346  mse= 8.946564528478469  Train_accuracy= 1.0\n",
      "epoch=  785  cost= 0.032028344  mse= 9.112739109154553  Train_accuracy= 1.0\n",
      "epoch=  786  cost= 0.03145719  mse= 9.127991230408872  Train_accuracy= 1.0\n",
      "epoch=  787  cost= 0.03091473  mse= 9.274382226718867  Train_accuracy= 1.0\n",
      "epoch=  788  cost= 0.030385615  mse= 9.27705007925946  Train_accuracy= 1.0\n",
      "epoch=  789  cost= 0.029893113  mse= 9.384886462491854  Train_accuracy= 1.0\n",
      "epoch=  790  cost= 0.029401926  mse= 9.40133912436366  Train_accuracy= 1.0\n",
      "epoch=  791  cost= 0.028989576  mse= 9.502829737426095  Train_accuracy= 1.0\n",
      "epoch=  792  cost= 0.028593747  mse= 9.394402915484582  Train_accuracy= 1.0\n",
      "epoch=  793  cost= 0.028076166  mse= 9.603683209366327  Train_accuracy= 1.0\n",
      "epoch=  794  cost= 0.027608834  mse= 9.582533940695921  Train_accuracy= 1.0\n",
      "epoch=  795  cost= 0.027183348  mse= 9.721784339423241  Train_accuracy= 1.0\n",
      "epoch=  796  cost= 0.02673954  mse= 9.723826339794448  Train_accuracy= 1.0\n",
      "epoch=  797  cost= 0.0263192  mse= 9.87498488840729  Train_accuracy= 1.0\n",
      "epoch=  798  cost= 0.025939211  mse= 9.853423240749251  Train_accuracy= 1.0\n",
      "epoch=  799  cost= 0.02557842  mse= 10.028181049223539  Train_accuracy= 1.0\n",
      "epoch=  800  cost= 0.025240347  mse= 9.936715098745083  Train_accuracy= 1.0\n",
      "epoch=  801  cost= 0.024882942  mse= 10.168795036517993  Train_accuracy= 1.0\n",
      "epoch=  802  cost= 0.024648698  mse= 9.934501304674834  Train_accuracy= 1.0\n",
      "epoch=  803  cost= 0.024282075  mse= 10.269030316934609  Train_accuracy= 1.0\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch=  804  cost= 0.023930352  mse= 10.07968699312817  Train_accuracy= 1.0\n",
      "epoch=  805  cost= 0.023527153  mse= 10.363205664767323  Train_accuracy= 1.0\n",
      "epoch=  806  cost= 0.023184597  mse= 10.305324934249004  Train_accuracy= 1.0\n",
      "epoch=  807  cost= 0.022876944  mse= 10.440182847023792  Train_accuracy= 1.0\n",
      "epoch=  808  cost= 0.02260175  mse= 10.394838918050631  Train_accuracy= 1.0\n",
      "epoch=  809  cost= 0.02230964  mse= 10.552412627799821  Train_accuracy= 1.0\n",
      "epoch=  810  cost= 0.022004843  mse= 10.524955240990746  Train_accuracy= 1.0\n",
      "epoch=  811  cost= 0.021731935  mse= 10.63863374616759  Train_accuracy= 1.0\n",
      "epoch=  812  cost= 0.021475714  mse= 10.614849407813512  Train_accuracy= 1.0\n",
      "epoch=  813  cost= 0.02119415  mse= 10.716860292782496  Train_accuracy= 1.0\n",
      "epoch=  814  cost= 0.020937009  mse= 10.768633136747894  Train_accuracy= 1.0\n",
      "epoch=  815  cost= 0.020702427  mse= 10.778477595573827  Train_accuracy= 1.0\n",
      "epoch=  816  cost= 0.02044476  mse= 10.867206592354366  Train_accuracy= 1.0\n",
      "epoch=  817  cost= 0.020213732  mse= 10.908271129533118  Train_accuracy= 1.0\n",
      "epoch=  818  cost= 0.01999566  mse= 10.988052748578049  Train_accuracy= 1.0\n",
      "epoch=  819  cost= 0.019774986  mse= 10.950431240652685  Train_accuracy= 1.0\n",
      "epoch=  820  cost= 0.019574234  mse= 11.110691018355437  Train_accuracy= 1.0\n",
      "epoch=  821  cost= 0.019354682  mse= 11.008630022196959  Train_accuracy= 1.0\n",
      "epoch=  822  cost= 0.019138351  mse= 11.201913801601728  Train_accuracy= 1.0\n",
      "epoch=  823  cost= 0.018876528  mse= 11.076167108900355  Train_accuracy= 1.0\n",
      "epoch=  824  cost= 0.018654041  mse= 11.275821032320572  Train_accuracy= 1.0\n",
      "epoch=  825  cost= 0.01836719  mse= 11.204134202521796  Train_accuracy= 1.0\n",
      "epoch=  826  cost= 0.018147096  mse= 11.338603983045584  Train_accuracy= 1.0\n",
      "epoch=  827  cost= 0.017921774  mse= 11.295133113078979  Train_accuracy= 1.0\n",
      "epoch=  828  cost= 0.017682312  mse= 11.418595295751134  Train_accuracy= 1.0\n",
      "epoch=  829  cost= 0.017504688  mse= 11.389543411693577  Train_accuracy= 1.0\n",
      "epoch=  830  cost= 0.017344598  mse= 11.545666704141231  Train_accuracy= 1.0\n",
      "epoch=  831  cost= 0.01711942  mse= 11.423083479095746  Train_accuracy= 1.0\n",
      "epoch=  832  cost= 0.016868254  mse= 11.61760197821205  Train_accuracy= 1.0\n",
      "epoch=  833  cost= 0.01665104  mse= 11.638272839624197  Train_accuracy= 1.0\n",
      "epoch=  834  cost= 0.016468994  mse= 11.70099419675686  Train_accuracy= 1.0\n",
      "epoch=  835  cost= 0.016257819  mse= 11.674157978958613  Train_accuracy= 1.0\n",
      "epoch=  836  cost= 0.016074402  mse= 11.804802238914917  Train_accuracy= 1.0\n",
      "epoch=  837  cost= 0.015897417  mse= 11.753343677768056  Train_accuracy= 1.0\n",
      "epoch=  838  cost= 0.015718771  mse= 11.898006646949494  Train_accuracy= 1.0\n",
      "epoch=  839  cost= 0.01556362  mse= 11.81860866421801  Train_accuracy= 1.0\n",
      "epoch=  840  cost= 0.015442114  mse= 11.986639441253724  Train_accuracy= 1.0\n",
      "epoch=  841  cost= 0.015238606  mse= 11.87189820689522  Train_accuracy= 1.0\n",
      "epoch=  842  cost= 0.015033187  mse= 12.047070815533829  Train_accuracy= 1.0\n",
      "epoch=  843  cost= 0.014887796  mse= 12.043495813803547  Train_accuracy= 1.0\n",
      "epoch=  844  cost= 0.01477549  mse= 12.149880623902654  Train_accuracy= 1.0\n",
      "epoch=  845  cost= 0.014621358  mse= 12.031822508203039  Train_accuracy= 1.0\n",
      "epoch=  846  cost= 0.014439686  mse= 12.229623934518392  Train_accuracy= 1.0\n",
      "epoch=  847  cost= 0.014286905  mse= 12.153904669295061  Train_accuracy= 1.0\n",
      "epoch=  848  cost= 0.014158447  mse= 12.301588710632249  Train_accuracy= 1.0\n",
      "epoch=  849  cost= 0.014022313  mse= 12.217124186660076  Train_accuracy= 1.0\n",
      "epoch=  850  cost= 0.013915493  mse= 12.38767506109214  Train_accuracy= 1.0\n",
      "epoch=  851  cost= 0.013730634  mse= 12.342115953634407  Train_accuracy= 1.0\n",
      "epoch=  852  cost= 0.01358291  mse= 12.448040784179458  Train_accuracy= 1.0\n",
      "epoch=  853  cost= 0.01346937  mse= 12.427137146395703  Train_accuracy= 1.0\n",
      "epoch=  854  cost= 0.013337239  mse= 12.558344043582094  Train_accuracy= 1.0\n",
      "epoch=  855  cost= 0.013215857  mse= 12.506966031072213  Train_accuracy= 1.0\n",
      "epoch=  856  cost= 0.0131023405  mse= 12.639286399836479  Train_accuracy= 1.0\n",
      "epoch=  857  cost= 0.0130046  mse= 12.545768234916924  Train_accuracy= 1.0\n",
      "epoch=  858  cost= 0.012904316  mse= 12.71632228142794  Train_accuracy= 1.0\n",
      "epoch=  859  cost= 0.012754001  mse= 12.613527242966537  Train_accuracy= 1.0\n",
      "epoch=  860  cost= 0.012632388  mse= 12.772871259517046  Train_accuracy= 1.0\n",
      "epoch=  861  cost= 0.012529949  mse= 12.70897772716888  Train_accuracy= 1.0\n",
      "epoch=  862  cost= 0.012433308  mse= 12.85965829881113  Train_accuracy= 1.0\n",
      "epoch=  863  cost= 0.012291853  mse= 12.782188576840362  Train_accuracy= 1.0\n",
      "epoch=  864  cost= 0.012171142  mse= 12.880004121139104  Train_accuracy= 1.0\n",
      "epoch=  865  cost= 0.012077668  mse= 12.901408260375344  Train_accuracy= 1.0\n",
      "epoch=  866  cost= 0.011983606  mse= 12.966325491735114  Train_accuracy= 1.0\n",
      "epoch=  867  cost= 0.011867462  mse= 12.970643402997123  Train_accuracy= 1.0\n",
      "epoch=  868  cost= 0.01175817  mse= 13.014580532763173  Train_accuracy= 1.0\n",
      "epoch=  869  cost= 0.011663916  mse= 13.034785799640558  Train_accuracy= 1.0\n",
      "epoch=  870  cost= 0.01156334  mse= 13.09364285935064  Train_accuracy= 1.0\n",
      "epoch=  871  cost= 0.011469577  mse= 13.121048070611701  Train_accuracy= 1.0\n",
      "epoch=  872  cost= 0.011406008  mse= 13.166239245789582  Train_accuracy= 1.0\n",
      "epoch=  873  cost= 0.011311172  mse= 13.210098715397816  Train_accuracy= 1.0\n",
      "epoch=  874  cost= 0.01120616  mse= 13.19417363047106  Train_accuracy= 1.0\n",
      "epoch=  875  cost= 0.011115769  mse= 13.253420282396617  Train_accuracy= 1.0\n",
      "epoch=  876  cost= 0.011053755  mse= 13.303038892528326  Train_accuracy= 1.0\n",
      "epoch=  877  cost= 0.0109662935  mse= 13.362330646192385  Train_accuracy= 1.0\n",
      "epoch=  878  cost= 0.010874468  mse= 13.31366730959907  Train_accuracy= 1.0\n",
      "epoch=  879  cost= 0.010788049  mse= 13.396558254591994  Train_accuracy= 1.0\n",
      "epoch=  880  cost= 0.010708014  mse= 13.396469928664146  Train_accuracy= 1.0\n",
      "epoch=  881  cost= 0.010630124  mse= 13.467567633817675  Train_accuracy= 1.0\n",
      "epoch=  882  cost= 0.01056498  mse= 13.479582384209328  Train_accuracy= 1.0\n",
      "epoch=  883  cost= 0.010480951  mse= 13.516027173422112  Train_accuracy= 1.0\n",
      "epoch=  884  cost= 0.010405137  mse= 13.530326757928693  Train_accuracy= 1.0\n",
      "epoch=  885  cost= 0.010332646  mse= 13.595841364606713  Train_accuracy= 1.0\n",
      "epoch=  886  cost= 0.010275672  mse= 13.594306678643456  Train_accuracy= 1.0\n",
      "epoch=  887  cost= 0.010220699  mse= 13.685695384029053  Train_accuracy= 1.0\n",
      "epoch=  888  cost= 0.010126251  mse= 13.62857208672419  Train_accuracy= 1.0\n",
      "epoch=  889  cost= 0.010055846  mse= 13.728411588707544  Train_accuracy= 1.0\n",
      "epoch=  890  cost= 0.009996056  mse= 13.71195101388912  Train_accuracy= 1.0\n",
      "epoch=  891  cost= 0.009943799  mse= 13.812464638254877  Train_accuracy= 1.0\n",
      "epoch=  892  cost= 0.009853361  mse= 13.763414100402295  Train_accuracy= 1.0\n",
      "epoch=  893  cost= 0.009790023  mse= 13.838267802672865  Train_accuracy= 1.0\n",
      "epoch=  894  cost= 0.0097198635  mse= 13.84142707603755  Train_accuracy= 1.0\n",
      "epoch=  895  cost= 0.00965437  mse= 13.883872610175759  Train_accuracy= 1.0\n",
      "epoch=  896  cost= 0.00959977  mse= 13.903594485233421  Train_accuracy= 1.0\n",
      "epoch=  897  cost= 0.009546549  mse= 13.961412805951042  Train_accuracy= 1.0\n",
      "epoch=  898  cost= 0.009473719  mse= 13.944355305138341  Train_accuracy= 1.0\n",
      "epoch=  899  cost= 0.009410963  mse= 13.988303621484258  Train_accuracy= 1.0\n",
      "epoch=  900  cost= 0.0093493685  mse= 14.025583938133227  Train_accuracy= 1.0\n",
      "epoch=  901  cost= 0.009298292  mse= 14.04073470810202  Train_accuracy= 1.0\n",
      "epoch=  902  cost= 0.009245943  mse= 14.080019071495009  Train_accuracy= 1.0\n",
      "epoch=  903  cost= 0.009180956  mse= 14.083343165581297  Train_accuracy= 1.0\n",
      "epoch=  904  cost= 0.009125916  mse= 14.149828954477961  Train_accuracy= 1.0\n",
      "epoch=  905  cost= 0.009064314  mse= 14.156524759933605  Train_accuracy= 1.0\n",
      "epoch=  906  cost= 0.009006044  mse= 14.200167851460503  Train_accuracy= 1.0\n",
      "epoch=  907  cost= 0.008951098  mse= 14.231936097812897  Train_accuracy= 1.0\n",
      "epoch=  908  cost= 0.0089032445  mse= 14.258478244173844  Train_accuracy= 1.0\n",
      "epoch=  909  cost= 0.00885324  mse= 14.295702929058006  Train_accuracy= 1.0\n",
      "epoch=  910  cost= 0.008791178  mse= 14.30855308179097  Train_accuracy= 1.0\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch=  911  cost= 0.008737452  mse= 14.351065604703617  Train_accuracy= 1.0\n",
      "epoch=  912  cost= 0.008683607  mse= 14.38141512963105  Train_accuracy= 1.0\n",
      "epoch=  913  cost= 0.00863133  mse= 14.417030773011518  Train_accuracy= 1.0\n",
      "epoch=  914  cost= 0.00858814  mse= 14.450422093939737  Train_accuracy= 1.0\n",
      "epoch=  915  cost= 0.008540666  mse= 14.505014766816736  Train_accuracy= 1.0\n",
      "epoch=  916  cost= 0.008482938  mse= 14.499875799617842  Train_accuracy= 1.0\n",
      "epoch=  917  cost= 0.008428114  mse= 14.53220496245223  Train_accuracy= 1.0\n",
      "epoch=  918  cost= 0.008377095  mse= 14.576454873003357  Train_accuracy= 1.0\n",
      "epoch=  919  cost= 0.008338795  mse= 14.596386284704035  Train_accuracy= 1.0\n",
      "epoch=  920  cost= 0.008291296  mse= 14.641913960188456  Train_accuracy= 1.0\n",
      "epoch=  921  cost= 0.00823426  mse= 14.648916262383565  Train_accuracy= 1.0\n",
      "epoch=  922  cost= 0.008183822  mse= 14.691393923977365  Train_accuracy= 1.0\n",
      "epoch=  923  cost= 0.008137476  mse= 14.724174254581374  Train_accuracy= 1.0\n",
      "epoch=  924  cost= 0.008099223  mse= 14.747488596813483  Train_accuracy= 1.0\n",
      "epoch=  925  cost= 0.008052977  mse= 14.798645631364385  Train_accuracy= 1.0\n",
      "epoch=  926  cost= 0.008001586  mse= 14.80480864179549  Train_accuracy= 1.0\n",
      "epoch=  927  cost= 0.007954216  mse= 14.83630352646253  Train_accuracy= 1.0\n",
      "epoch=  928  cost= 0.007907507  mse= 14.87336283703821  Train_accuracy= 1.0\n",
      "epoch=  929  cost= 0.007863334  mse= 14.909416504691736  Train_accuracy= 1.0\n",
      "epoch=  930  cost= 0.007827406  mse= 14.939397367272528  Train_accuracy= 1.0\n",
      "epoch=  931  cost= 0.007788739  mse= 14.983392738396695  Train_accuracy= 1.0\n",
      "epoch=  932  cost= 0.007736148  mse= 14.962624661978206  Train_accuracy= 1.0\n",
      "epoch=  933  cost= 0.007692145  mse= 15.00809284560415  Train_accuracy= 1.0\n",
      "epoch=  934  cost= 0.0076486757  mse= 15.042945430028729  Train_accuracy= 1.0\n",
      "epoch=  935  cost= 0.0076088984  mse= 15.076132739922649  Train_accuracy= 1.0\n",
      "epoch=  936  cost= 0.007574566  mse= 15.131270364347111  Train_accuracy= 1.0\n",
      "epoch=  937  cost= 0.007526149  mse= 15.11139950984395  Train_accuracy= 1.0\n",
      "epoch=  938  cost= 0.007483389  mse= 15.15665180174869  Train_accuracy= 1.0\n",
      "epoch=  939  cost= 0.0074418625  mse= 15.197585921221693  Train_accuracy= 1.0\n",
      "epoch=  940  cost= 0.007403089  mse= 15.231011383926319  Train_accuracy= 1.0\n",
      "epoch=  941  cost= 0.0073718335  mse= 15.274477112684284  Train_accuracy= 1.0\n",
      "epoch=  942  cost= 0.0073246835  mse= 15.255905669001342  Train_accuracy= 1.0\n",
      "epoch=  943  cost= 0.0072836597  mse= 15.31374371134846  Train_accuracy= 1.0\n",
      "epoch=  944  cost= 0.0072460356  mse= 15.353835651899814  Train_accuracy= 1.0\n",
      "epoch=  945  cost= 0.00720686  mse= 15.361203188025021  Train_accuracy= 1.0\n",
      "epoch=  946  cost= 0.0071706334  mse= 15.398706305150691  Train_accuracy= 1.0\n",
      "epoch=  947  cost= 0.007136021  mse= 15.43018476045452  Train_accuracy= 1.0\n",
      "epoch=  948  cost= 0.0070937607  mse= 15.442847105970925  Train_accuracy= 1.0\n",
      "epoch=  949  cost= 0.007054485  mse= 15.455836939312059  Train_accuracy= 1.0\n",
      "epoch=  950  cost= 0.0070155426  mse= 15.513403349147548  Train_accuracy= 1.0\n",
      "epoch=  951  cost= 0.006983329  mse= 15.524632607631261  Train_accuracy= 1.0\n",
      "epoch=  952  cost= 0.0069587086  mse= 15.58704912099259  Train_accuracy= 1.0\n",
      "epoch=  953  cost= 0.0069119823  mse= 15.54932164380604  Train_accuracy= 1.0\n",
      "epoch=  954  cost= 0.0068753525  mse= 15.612029693675037  Train_accuracy= 1.0\n",
      "epoch=  955  cost= 0.0068396623  mse= 15.634852317565272  Train_accuracy= 1.0\n",
      "epoch=  956  cost= 0.006805486  mse= 15.639959229910867  Train_accuracy= 1.0\n",
      "epoch=  957  cost= 0.0067748637  mse= 15.684768146911582  Train_accuracy= 1.0\n",
      "epoch=  958  cost= 0.006743027  mse= 15.701423138726048  Train_accuracy= 1.0\n",
      "epoch=  959  cost= 0.0067086336  mse= 15.723596653663842  Train_accuracy= 1.0\n",
      "epoch=  960  cost= 0.0066742753  mse= 15.719672302895633  Train_accuracy= 1.0\n",
      "epoch=  961  cost= 0.0066415574  mse= 15.791731216248374  Train_accuracy= 1.0\n",
      "epoch=  962  cost= 0.0066119405  mse= 15.772481483883897  Train_accuracy= 1.0\n",
      "epoch=  963  cost= 0.0065821637  mse= 15.834503167897505  Train_accuracy= 1.0\n",
      "epoch=  964  cost= 0.00654754  mse= 15.8217805788067  Train_accuracy= 1.0\n",
      "epoch=  965  cost= 0.0065146224  mse= 15.866390855847438  Train_accuracy= 1.0\n",
      "epoch=  966  cost= 0.0064837467  mse= 15.86472519990766  Train_accuracy= 1.0\n",
      "epoch=  967  cost= 0.0064515895  mse= 15.923392895640282  Train_accuracy= 1.0\n",
      "epoch=  968  cost= 0.00642481  mse= 15.906302152782489  Train_accuracy= 1.0\n",
      "epoch=  969  cost= 0.006395907  mse= 15.972056494543148  Train_accuracy= 1.0\n",
      "epoch=  970  cost= 0.006363728  mse= 15.95266132142113  Train_accuracy= 1.0\n",
      "epoch=  971  cost= 0.006333807  mse= 16.011200373152043  Train_accuracy= 1.0\n",
      "epoch=  972  cost= 0.0063024955  mse= 15.995561972370641  Train_accuracy= 1.0\n",
      "epoch=  973  cost= 0.0062766145  mse= 16.04815147444432  Train_accuracy= 1.0\n",
      "epoch=  974  cost= 0.006244406  mse= 16.040612774493475  Train_accuracy= 1.0\n",
      "epoch=  975  cost= 0.0062136105  mse= 16.09387542048052  Train_accuracy= 1.0\n",
      "epoch=  976  cost= 0.006185669  mse= 16.111208984949663  Train_accuracy= 1.0\n",
      "epoch=  977  cost= 0.006159768  mse= 16.149070759321862  Train_accuracy= 1.0\n",
      "epoch=  978  cost= 0.0061320188  mse= 16.1016819535508  Train_accuracy= 1.0\n",
      "epoch=  979  cost= 0.0061045066  mse= 16.170160281247963  Train_accuracy= 1.0\n",
      "epoch=  980  cost= 0.0060738986  mse= 16.17688450594061  Train_accuracy= 1.0\n",
      "epoch=  981  cost= 0.0060500437  mse= 16.219678768548803  Train_accuracy= 1.0\n",
      "epoch=  982  cost= 0.006019576  mse= 16.226708104801254  Train_accuracy= 1.0\n",
      "epoch=  983  cost= 0.005991967  mse= 16.24533058862349  Train_accuracy= 1.0\n",
      "epoch=  984  cost= 0.005962656  mse= 16.274359115342097  Train_accuracy= 1.0\n",
      "epoch=  985  cost= 0.005941042  mse= 16.30441421656306  Train_accuracy= 1.0\n",
      "epoch=  986  cost= 0.0059210085  mse= 16.256996619266186  Train_accuracy= 1.0\n",
      "epoch=  987  cost= 0.0058938786  mse= 16.341387685352743  Train_accuracy= 1.0\n",
      "epoch=  988  cost= 0.0058631627  mse= 16.33363963921518  Train_accuracy= 1.0\n",
      "epoch=  989  cost= 0.0058335275  mse= 16.363816813041872  Train_accuracy= 1.0\n",
      "epoch=  990  cost= 0.005807939  mse= 16.396522232631924  Train_accuracy= 1.0\n",
      "epoch=  991  cost= 0.0057872906  mse= 16.419259987512977  Train_accuracy= 1.0\n",
      "epoch=  992  cost= 0.00576078  mse= 16.43113993629358  Train_accuracy= 1.0\n",
      "epoch=  993  cost= 0.0057344562  mse= 16.44928280874565  Train_accuracy= 1.0\n",
      "epoch=  994  cost= 0.0057093627  mse= 16.467588171226073  Train_accuracy= 1.0\n",
      "epoch=  995  cost= 0.0056842216  mse= 16.49422463267622  Train_accuracy= 1.0\n",
      "epoch=  996  cost= 0.0056646424  mse= 16.51922444869016  Train_accuracy= 1.0\n",
      "epoch=  997  cost= 0.005638813  mse= 16.54067995342743  Train_accuracy= 1.0\n",
      "epoch=  998  cost= 0.005618603  mse= 16.500273605628934  Train_accuracy= 1.0\n",
      "epoch=  999  cost= 0.0055917217  mse= 16.56451111439289  Train_accuracy= 1.0\n"
     ]
    }
   ],
   "source": [
    "# Important parameter and variables to work with tensors\n",
    "learning_rate=0.3\n",
    "training_epochs=1000\n",
    "cost_history = np.empty(shape=[1], dtype=float)\n",
    "n_dim = X.shape[1]\n",
    "print(\"n_dim= \",n_dim)\n",
    "n_class = 2\n",
    "model_path = \"Naval Mine/NMI\"\n",
    "\n",
    "# Define the number of layers and number of neurons for each layer\n",
    "\n",
    "n_hidden_1 = 60\n",
    "n_hidden_2 = 60\n",
    "n_hidden_3 = 60\n",
    "n_hidden_4 = 60\n",
    "\n",
    "x = tf.placeholder(tf.float32,[None,n_dim])\n",
    "y_ = tf.placeholder(tf.float32,[None,n_class])\n",
    "w = tf.Variable(tf.zeros([n_dim,n_class]))\n",
    "b = tf.Variable(tf.zeros(n_class))\n",
    "\n",
    "\n",
    "weigths = {\n",
    "    'h1': tf.Variable(tf.truncated_normal([n_dim, n_hidden_1])),\n",
    "    'h2': tf.Variable(tf.truncated_normal([n_hidden_1, n_hidden_2])),\n",
    "    'h3': tf.Variable(tf.truncated_normal([n_hidden_2, n_hidden_3])),\n",
    "    'h4': tf.Variable(tf.truncated_normal([n_hidden_3, n_hidden_4])),\n",
    "    'out': tf.Variable(tf.truncated_normal([n_hidden_4, n_class]))\n",
    "}\n",
    "biases = {\n",
    "    'b1': tf.Variable(tf.truncated_normal([n_hidden_1])),\n",
    "    'b2': tf.Variable(tf.truncated_normal([n_hidden_2])),\n",
    "    'b3': tf.Variable(tf.truncated_normal([n_hidden_3])),\n",
    "    'b4': tf.Variable(tf.truncated_normal([n_hidden_4])),\n",
    "    'out': tf.Variable(tf.truncated_normal([n_class]))\n",
    "}\n",
    "\n",
    "#define our model\n",
    "def multilayer_perceptron(x,weigths,biases):\n",
    "    \n",
    "    layer_1 = tf.add(tf.matmul(x, weigths['h1']), biases['b1'])\n",
    "    layer_1 = tf.nn.sigmoid(layer_1)\n",
    "    \n",
    "    layer_2 = tf.add(tf.matmul(layer_1, weigths['h2']), biases['b2'])\n",
    "    layer_2 = tf.nn.sigmoid(layer_2)\n",
    "    \n",
    "    layer_3 = tf.add(tf.matmul(layer_2, weigths['h3']), biases['b3'])\n",
    "    layer_3 = tf.nn.sigmoid(layer_3)\n",
    "    \n",
    "    layer_4 = tf.add(tf.matmul(layer_3, weigths['h4']), biases['b4'])\n",
    "    layer_4 = tf.nn.relu(layer_4)\n",
    "    \n",
    "    out_layer = tf.matmul(layer_4, weigths['out']) + biases['out']\n",
    "    return out_layer\n",
    "\n",
    "init = tf.global_variables_initializer()\n",
    "\n",
    "saver = tf.train.Saver()\n",
    "\n",
    "y = multilayer_perceptron(x,weigths,biases)\n",
    "\n",
    "cost_function = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=y,labels=y_))\n",
    "training_step = tf.train.GradientDescentOptimizer(learning_rate).minimize(cost_function)\n",
    "\n",
    "sess = tf.Session()\n",
    "sess.run(init)\n",
    "\n",
    "mse_history = []\n",
    "accuracy_history = []\n",
    "\n",
    "for epoch in range(training_epochs):\n",
    "    sess.run(training_step, feed_dict = {x:train_x, y_:train_y})\n",
    "    cost = sess.run(cost_function, feed_dict = {x:train_x, y_:train_y})\n",
    "    cost_history = np.append(cost_history, cost)\n",
    "    \n",
    "    correct_prediction = tf.equal(tf.arg_max(y,1), tf.arg_max(y_,1))\n",
    "    accuracy = tf.reduce_mean(tf.cast(correct_prediction,tf.float32))\n",
    "    accuracy = sess.run(accuracy, feed_dict = {x: train_x, y_:train_y})\n",
    "    accuracy_history.append(accuracy)\n",
    "    \n",
    "    pred_y = sess.run(y, feed_dict={x:test_x})\n",
    "    mse = tf.reduce_mean(tf.square(pred_y - test_y))\n",
    "    mse_ = sess.run(mse)\n",
    "    mse_history.append(mse_)\n",
    "    \n",
    "    \n",
    "    print('epoch= ',epoch,' cost=',cost,' mse=',mse_,' Train_accuracy=',accuracy)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model Saved in file :  Naval Mine/NMI\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAD8CAYAAACMwORRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJztvXm8JUV9Nv58+yx3m32F2VlmgGFzYNgRFERxCRhjDERjNEYiP5cETQwaY5RfePWNSUxM0IjGGFEhBoyAoLiBGpVl2JkZBoZhhrnMwNxZ7szc7Wxd7x/d1V1dXd1dffY+t57PZ+ae011dXX26++mnn/rWt4gxBgMDAwOD3oLV6QYYGBgYGDQfhtwNDAwMehCG3A0MDAx6EIbcDQwMDHoQhtwNDAwMehCG3A0MDAx6EIbcDQwMDHoQhtwNDAwMehCG3A0MDAx6EPlO7XjBggVs1apVndq9gYGBQSbx8MMP72WMLUwq1zFyX7VqFTZs2NCp3RsYGBhkEkS0Q6ecsWUMDAwMehCG3A0MDAx6EIbcDQwMDHoQhtwNDAwMehCG3A0MDAx6EInkTkRfI6I9RPRUxHoioi8Q0VYieoKITmt+Mw0MDAwM0kBHuX8dwKUx618PYLX77yoAX2q8WQYGBgYGjSAxzp0x9gsiWhVT5HIA32DOfH33E9EcIjqSMba7SW00MDDIEA6Ml/HN+3egUrM73ZSuxcUnLMapy+e0dB/NGMS0FMBO4fuwuyxE7kR0FRx1jxUrVjRh1wYGBt2GHzz1Ev7hx88AAIg63JguxaJZ/Zkgd9XpU866zRi7EcCNALB+/XozM7eBQROwb6yEXz23D6VKDQPFnKeYzz1mARbP6teq47mRMTz6wij6C45Te+aqeVikuS0AbNp1CEcvHEJ/IYcDE2UAwJa/vRR9+VzKozFoFppB7sMAlgvflwHY1YR6DQwMNHDOZ36GcoQFsv2zb9Sq4+J/+Hng++WvWIJ/vmKd1rYHJyp4wxd+ictOXYIvXLkOoxNlDBZzhtg7jGaEQt4B4J1u1MzZAA4av90gS5iq1LBz/wS27jns/Rs+MNHpZilRsxkmylUAAGMM20bGIok9CbtGJ7F1z2E89eJBb9ltV5+Lk5bOwvN7x7Xr2ToyBgC48wlH041OVDB7oFBXmwyah0TlTkQ3A3gVgAVENAzgbwAUAIAx9m8A7gbwBgBbAUwAeHerGmtg0Aoc/9c/VC6/7epzcfrKuW1uTTw+eusTuO2RYWz/7BvxzQdewF9/LxihPLMvj8OlamI9m3Ydwhu+8MvQ8tNWzMHzI+MYL9fwq617cd6xCxLr+vLPnwMAMAYcnqpgdNKQezdAJ1rmyoT1DMD7m9YiA4MO4S8vPR7L5g7gpYNTuP7uzXjp4FSnmxTCbY8MAwCqNRvD+ydQzFn4+7edislyFbMHCrjo+MX43D1P4yu/fD62nk27DwEAPn3ZiZg3VESpauPkpbNBRBgv15wyuw5pkbvYebZ97wQOTlQwZ9CQe6fRsZS/BgZpcdU3NuDQVAW3XHVOYtm9YyWs/9uf4D/efQZefdwiAMCFn7sXF6xeiP//zSd55X729Mve5yvOWI65Q0Xs3D+B6+/ejPFysgJuJ/aPl73Pn75zE8pVG3MGC7js1CWBcm9etzSR3D/vRrNceeYKFPNqd3bBzGLk9u/6jwdx35YRfODVx+LHm/zf8Es/34rRyTKOXjAj8XgMWguTfsAgM/jRppdx/7b9WmWfHHZ85P/41XZv2Y59E7jp/mAq7Bt/sc37PHfIIbMZfY7mmdCwN9qJza7aBoCb7t+B0cmyUiH3F5I7MkvVGo5aMKQk9luuOhsAEGfl37dlBADwr/duDSw/NFnFqFHuXQFD7gaZwM796To4a7ZjFuQUgbpPDI/igzc/imrN9iI6zj56nrd+sM9Zxu2JduPvfvg0Vl17Fz5086MAgB8+9RJO+pt78PavPhAoNzpRwZyBsLrOCcHl5WqYoRljODRZxetOPEK5/6MXDgFwOprT4KLjF2F0smw89y6BIXeDTOCzP3w6Vfkac8ndCrP7R77zOO58fBee3TOGpXMHAAB/9zuneuuLOQt5i7yolHbji/c5HZR3PO5En9z15G6MKd4iDk5WMFuhkFfOH/TePiYVD6jJSg3lmh2prrnyT0vuswcKGJ2ooFy1td4eDFoLQ+4GmcfDOw7gu25HI4ftKnfLVbFiaN/Mfof4dh+cxO2PvoijFgxhxfxBbz0RYbCYw3ipNcq9UrPxse8+iTf9yy8xLpH2dzbsDJXfsU8dlnhwsoI5CoVMRPjEG08AAGW/waMvjAKAclsA6HffZkqC6r9/2z7vYfNfD72g3I6Tu9MGZRGDNsJ0qBpkHr/zpV8DAN5y2jJvmazcf+/Lv/HWzR10rIzteycirZe+Qg6lamvIffPuQ7j5QYcgb7p/B9534THeuv/89fZA2WrNjozaifO2B3m/gZLcDwAATl42W7ltIUewKKjcr7jxfgDAZacuwV/e9qRyu4Uz+7w3DFIOXDdoJ4xyN+h6VGo27noieVzcyOESrrtzE3729Mt40h2YY7nkLipkcmXlDW5n4JtfsTRUl0VO3HYrsH2f338wINgXh6cq2LjrEN517ipBedcwOlnBn1x4NLZ/9o349GUneuUnKzXMGVRHtAwV3X4DxdvH4VIVfXkLJy5RkzsRob+QS23LrJo/JNSRalODFsAod4Oux2+e26dV7v3fegQPbt+Pr/3KDwPknYuW4L0fnnKsg31uaKHClgeBYLeI3XcIFlGfEK3CH2AnHDnTi1TZP152Qh7djlM57pyTuIzBIlfuCnKfqnrWVBT68lbAluHgdpeMC9csxBGz/Vw0hts7D6PcDboe3E+/+PhFWDizL7Lctr1joWXclhHJZqpqY/Esvx5Lwe6tVu5csXPyfX7vOG5+aCdyFuGtpy/HkBuxs3t0EgA8++XYRTNw2oo5OP6ImQDUHcYAvO1VtszYVNXrcI1CPmeh6hK5SOgqwgeAj156HPJCW4xy7zwMuRt0NbaNjOF7jzkdeUfO6fdCHFU4ZmF44EzBjYUUtypVap7vDvidriKICDG7agg79o1j7ZJZAHzyfcsXf4XHd46iZjPkLPLI/0VO7kLnZ96y/FzpESwqPzxEjJWqmNkfH6qYtwhVdx8bdhzwlkdZNRZR4EFDht07DkPuBm0HY8yLv2aMBTouazbzSKVctfG/W/cCAL7yzvXIW5a3ToU1i2d6n49a4Pi/hZxziYuKc6pSC8Rhq8SvZTltaxQT5Wog1rxas/HICwdw7MIZKOTII98DbpQJBx9ctOdwCQACIY/5HHmqOopCOdGqrKXDUxUN5U6o1pxt94+XvOVTEZ3MREatdxsMuRu0HV+87zms+cQPcHCygm8+8AKO+8QPsfugo1Av+Lt7ccqnfwQAWPOJH+CTt28E4OQXz1sUq9xFVckVcaVmY8/hqQDBbt83EYgyUSr3JnjuD+84gLWfvAdrPvEDb9knvvcUbOYMFBos5j1y5w+Yk5c6nZx8cNWeQ06kjDhYKWcRKu7xqNouLleTe7Lnnrd8W0ZU/1MV9cM1rNxjqzdoAwy5G7Qd3EMfOTyFu9w0sc+POJ2ML45OYqJcC5D41999BmYPFpATFKsKIgnxKJFKjeHxnQdDZUWyVBGRRREzzqTAr9y3DsAfKbpx1yH05S38/lkrMFTMYbzkKHubAa8/6Qjc9J4zAYSVu/gwKuQslF1VHUWiHrkruPjwVBUzEsmdUHU3Hg+Qe4RyR/BBY0IhOw8TLWPQdnCF95e3PYlBN9pDzkkuKu1XuYm/kpS7OGCHx1v/z6MvKuPEZw34l75K/VoNeu4HJyv4Rzc5F+A8tGo2w5MvHsQfnL0SM/sLGOxzlDsfRXrGqnleaCOPohnhtoxgI23adQh7x5zlURTKD0ml3MdKVcxMsGVylm/LiDl2ojpUiYJ0bpR752GUu0HbkbOcy+7hHQc8T5wTCQcn91cftzCwXdVmkV74RMSIUtUDQYwvVwackJoYdbHL7QjlGJuq4hG3Y/I1axcDcMIYx8tV76HEI1wAX7nzh5SY4OulQ/7DKtKWcQ9KPgTGmFaHakGIlhGVe8lV7n98/lHB/VEwcsdwe+dhyN2gLfjKL7bhQzc/CsZYIGSOf/7WAzu8yBAAKNUcErnohMWhslGK+sHt4YyReYuUTNMnkrsyFFLfl2GM4eP/86Q38hPwPX9OguWajR37x5G3COcdMx+AE4s+Uarh/m1OHP9A0VfTRakjOJIsI20Z56/8gJqsOJZXki2Ts3wLTFTuU257Llm7OFCeiAIPSaPcOw9D7gZtwfV3b8Ydj+/CoclqQOFx5X7vlhH8n7s3e8u5cu/L+Zco366qMpJd8AmeOWzGlPxXFOpV2zL6yv2lQ1P49gMv4E9uethbxj1/7pWXqza275vA8nmDyLv7HupzlPsOd8TqGav8WZ/63ONQZXUUEcWhfodqcPmhSYeokzpUCzk/FFKMkOGeez4X/J0tMp57t8GQu0EsqjUbn/nBZvxa6BxsBKOT5aByF3LyFoTlnNREO4JvF+e7y6lmbQY88HxY0ecswlvWOWkHoj13PXLnI0sHhdGiP3jKWcY99ErNxo5941gxz09QxqNlKjUbhRzhyNkD3jpfuTtkGhU3HmXLiJ47Ywxf/vlz2DtWwuikMyp3bkTaAg5RuZelMFLAHz8gtsMy0TJdBUPuBrHYtPsQvvzzbbjmO481pb7RiYpSuQO+Fw/4HawiufvKPZp0ZyV4yXJdQITnjmj7R8YT7sQgywXivvlBJ7ujqNxHDpcCI2OH+pxoGYfcg7ciP+5SJWzLvP4kPw97UrQMYwxPDB/EZ37wNP78vx/3sjZGZYTkKOT8MQUVoT+Ed/7mrTB15AyjdxUMuRvEgie5evlQKaGkHh4fHg2qdYHUHh8e9T5zUutTKfdaNOsmeckcQXJXK3fd/tS4vO885LJSs1Gq+pODAMBAgSt3Fk3u3HMXmvjOc1Z5nxNDIZlfx6HJCj77AycvvioPvIicEJkkRjJNRil3S/bcDdF3GiYU0iAWYpIrxljdN+3iWX14+VAJDz6/P0CmIkls3ePnhlEqdx5ZEyOph4p6l7TY0aqMc08xQpX764cmK6F182c45F6u2ShX7cDDinvupapCubvf+e8g/u4iiUYPYnL+2ox5JP3IC/7Dc+mcAdVmHpwUBypbxvms9NxNtExXwSh3g1jsEKa3q2faufFSFXsOTWHBDMeOYAh65i8emIzcDgh2fEZ57utWzPE+D0ZkSZShipARETdClTEWmPzjqV2OLSPGgM8fKuIdZ6/whvmXqw65iw+rwWIejDnpAIqSEiaiSFUelSxM3h5wlLvqOKJSBXOIYwoqEcr9M2852d8fKGDLGOHeeRjlbhALce7S0YlyYk4SGb/1L/+LbXvHsfZIJ1FWtWYHYtp/+vQe5XafdwcAFZSeezCCROT6Qs5CMWeFBkXJyFsUG9ERN0L11oeH8Re3PoGb33s2yjUbh6ecB5G4z/FyFUPFvEfmU1UbVZsFyJ3HtR+crASOkyNHhCpjIaKkAInGK3cmKPc0yOcIFZt77v5xlTxyt6Q3CDlaxqDTMMrdIBajExWPXEYnwrZDEra5Cperx2qNKUMZj5jVH/j+BJ9sQyAMlXIfL1VRE+ojciyVJCSp37iskA+7g5G27xsPvHlw+6JmM0xVbAwW896bx9hUeDASz7k+OlEJRBBx8LcLeY1MqiqIuWXqGYyVF0aoirYMfzvJWxR8sEi/u/HcOw9D7gaxGJ0sY6UbBaLylNOiYjPMSojUAPysjiJFyNEy+8fLOPFv7sFTLx7yyhycrGjFWCeTe7Tn7oVp5qxABy5f/jP3baS/YHmKXGUzcQtpdLIc8tyB6OgTnXhyK8aWKSreEmTkhWiZsvCm5ZF7zgq0wyIKtstwe8dhyN0gFqMTFW+CjCSrIw6cXypV20vNO3/I931lNb9SCCvk4OF3NY/cwxE8Lx2cilSzIpLC9uLi3EtCZ++U2w+xbsUc7/cZPuBYWZeedISv3F1yF0fGcrU+WbaVhMuPQ1bBOtkXxTh3+bTdctXZ6o3E7eHbUhVBufMHWCFHEJ9HoayQiXswaDUMuRtEYqpSQ6lqY55LwlFk98zLh3HG9T/B7oOTuPLG+z2/XARzqaJq297MPiIxySMxuVgUy3jKvcbrCrdHVJAqq0Ouy2lbGHEzMYkDrDhpH7twhhe+ye2rZXMHvWggj9wVI25L1ZpSuUfZMuJvkmjL2GHPfcFQ9GxW4j748ZdrtrdP/gDLW1bIYw88aIx07zi0yJ2ILiWiLUS0lYiuVaxfSUQ/JaIniOg+IlqmqscgW+Cv4NwbjhLuNz/4AkYOl3Dn47vwm2378M8/fTZUxlPuNeZ52WJ0SUWKXfceAAK1yZ67nGwMcAnR3SSfiyaYuHUA99wTbJm85cW4zx0qesR3cLKCWf155FxfupizlJ47J8dSxQ7FjQPCFIHSqmD4Y1IoZDi6KJdw7Hwf/IFcqfkhnBVBuZNky+QksjfoLBLJnYhyAG4A8HoAawFcSURrpWJ/D+AbjLFTAFwH4DPNbuh0wZ2P78K3HtjR6Wbgyz9/DvdtcbzjnKIjUwS3bfZIA53EwT18y6pte6QpRmHIlk9Noe45KXELR63cfWIpxPSsWjGhhoBTR1SHqui5j5drKOQIQ8U8aq5KHp0oB0INCzlSZncU1XCc5y776jretjgT038/vDOwLu6NRqzXFqw0PviqXLO9h1Zg0JKlZxcZtA86cW1nAtjKGNsGAER0C4DLAWwSyqwFcI37+V4A32tmI6cTPnjzowCAt5+1sqPt+Iw7khHwBxpFKVl+48tTsH3nIZ9UbEFt845K0YqRHxyqfcnKvaaIuhGJL06hBobPKw7Loujc8fxBxJijuos5yyPtctXG6GQlMLlGMW/h0JRj1YiJzUQyVHXwespYWiV73SqIce73bRkJHVsyKGDLeMq9ZnsPnbAtk9zRa9A+6NgySwGIj/5hd5mIxwH8jvv5twHMJKL5ckVEdBURbSCiDSMjI/Jqgy5FknKfdBW6zMc5gYV4SGSlZuNml/Tjwq9V5C5Gyzy2cxQ/3hSOkbcsn5RU+U/8cvHWgWWp21Cu2l4oZM0NM8zL5D5RCSQwK+QsHJhwEnaJI2jlaBMZ/OcLe+569ocVEfGjq9z5U69ctb0slTXbj7uPa79R7p2HDrmrTpN8xfw5gAuJ6FEAFwJ4EUAo4QZj7EbG2HrG2PqFCxfKqw26FJwko5Q7H7lakawVFYnsGy8nprEFfMtFJAkxWubNN/wK//bz50LbiYOPVD62XFcUokaobt7th13azInZz1nkjTAt12wcnKwEbJli3vI6WYf61OSuaqlny8R47nEkGhXxkzQ619lWsGVqDP3u21mQ3P3yoYFWiXswaDV0yH0YwHLh+zIAu8QCjLFdjLG3MMbWAfgrd1l44kqDWDw53Lqf7PBUBQ8pJrMoVWuBuT5V4CSpUu6MMfzgSSe9rZyeQBVuqDsQStWhqpMVMicQWlwsu7iKKXwZ0XMWIfYNMDfMkHvQvK7RiXIg62IxZ3mjWMX0CEmTW0SRsE5uGb7OZsDR7pgBDi3lDvLts5ofqsmYf06Mcu9u6JD7QwBWE9FRRFQEcAWAO8QCRLSAiHhdHwPwteY2c3rgt/71f1tW99XffAS/+2+/weGpILlef9dmvP2rD2DjLv/BIr/K56T4chE/eOolL3Mkjwjxt6v/Due7Cip3/pCJVv4OKbNAeY5V8/3Y+aQh+U60SBhizLdtO23JCeGXTq6YamAyjGDKAUG5BzogY6JlYjpU4yQy/y3kSa31ctP4eYBqNvM895qQDiE4IlXa3mj3jiOR3BljVQAfAHAPgM0AvsMY20hE1xHRZW6xVwHYQkTPAFgM4PoWtdegTjzpDueXwweffdnJxHjQVdQ7909gZCwY9SJ3qE5Vap49wSdwBpwQQBFJ4YZxUJGvHOceBb5eJNVizsKPP3xhbP0iovxqUbk7tgxDTkh3W6mFc8gUFKNSnX3E2zJWlC2jOViIpy2ekmwwHXLn23Krrb8g2DKK8iEyN9zecWhlgWKM3Q3gbmnZJ4XPtwK4tblNM2gmOFHFvS7XbIZX/t29oTwvfoeq8/36uzbjpvt34Pb3nxcgLjk9QWPKXeG5x9hDHIz5to086lO0iap2OCGXiKg4dzEe32aOfZTP+crdH8EpPFiEdkRNzK20ZdTBMilsGad9IeWu6ZnYjHkPM67cbSHts/gCJZ9qw+2dhxmhOs0QR4x83UuHpgLLOVHVXLLjedefePFggLgmJRKRbREeD5+mnapBTHGeu4iinHNcaI+ecg8vr0iee9VmDlm6VZeEGHgOsWM3KvxRZWP4g5iibZmkDtWaZMs4idU0o2WYb0P15cPKXfx5wp67ofdOw5B7G3D7Yy9i1bV3YY9EmgDw/Sd2YdW1d+Glg+F1zQS/EWVS8zoTKToaxhsQ4267yJ0q7q+/9xTGBA9/t3QMMn/qpgsm8h8kwfQD0d4/h7gmLkHWrAG/LarDjsoKWQnYMs5v5tgyTkP/7BZnOkKR0IvC7EvyqE7vs6Kpni0TsdxZF02iRE57xePQ6Uzl9TL4byo8FNJmfoNE28pEy3QfDLm3ATc/+AIAYOvIWGjdf7kx35t2tye4qCYxGfO4PXpaufDgIb+gTOiBfUnsqOvSiFPciZvoKHeRcERSlSNiTl85L54YEeG5VyXPvRb03Le8fBiAlIc+YjfBftFwIY/EQ6GQ6s+h7S3ChBzBpHkSeJ9DuSrZMhHKnT+0htw+BSPcOw9D7m0A56I4fzSpkzAtxktVvPLvfuaHP7rVf/He5/C+mx4Olb/yK/fjsZ2joeWAT6pc2Ys5YfjgHBkvjk6G3gTSEEt8h6pedkrRGolLaa5aFTWHatBzZyHlziF67lHnPfB7KIr40TLhtsVtJ5ab1AhPVYGHgvqeu2vLCJ676uF32sq54TYadASG3NsAfhPEknsds+XEYfPuQ9i5fxKfuXuz0wZ3+U3378APN77kt03Y5vM/CWdzBPz5Mr0Jk6u2Rzz7xx1yXzwr6Kf/zyPDCuWuSyz+0P9gBkT+kIneNmjLhPf3ubeegh9fc0GobhlRI1QrimgZZ+KKYLnglHMRUS9J0TKRce56toxFqFu5k5s4zI+W8ZW7mJQsenut3Ri0EIbc2wBOVHH3VbPJnRNy1Wb48aaXvcRVHE+/5I60FHcb0QRvEJOQE4bnYufkfrqr2DhqdviYdIklF8jrEu48VA06UqEvH55P9XfXL8dqN598HKKjZeQ4dwZLodxF+ys66kXtv3NwOyfUoRrILRN/DCHlru25O287ZalD1Yky4spdqyqDDsGQexvg2TIxN5au1aALbqVUagzv/caG0Porbrw/tCy6Q9VXbYDzqs5zvB9w4+MHCsHOUpsxr7xfTwpbRtGh6nm9ccpdWCdHy8j48CVr8LoTF+OyU5eE1nFykyHHuddc5S6Ts3js0VEvwv4UP01knLvirUAFi4CJijywTO+WJ+IdqopQSLeMemSv+lgN2g8zQXYbYGvYMtvdxFrNAvd8ox4aylQCkXXxDlXne7lqY8GMPgCHsc8d8CQOzgFc4mP12TKO381DIQV4yj0a4rqCYMuotlk0qx9f/oP10W1QLK9U/aU8pt6ZICRYLqjcdaJeFG3Q8NyTQiFlW0Y7WoZ3qHJyF2wZ7w1KFWUk/TXoHIxybwN8co8u84WfbW3qPvNe7vP48EZRfUUpd04mnLBK1Rpm9DmTPx9S5EwB/GHrgE8oOsrdmQRC/fAhxLCKAknJweLgJM5K9tz5ICZZqYrK3Qt6kX35pPQDESo42A8RdwzNt2WcayDYwa5qmxHunYch9zZAlQpl+95x7BqdbMn+Nu065I0WlTM1cnDiEO/PKM7kytS2nVzsz7w8hr6CPzk0UXBuUADYNjKOXzzjJCTjDxodXrn7Q6+EZYkdqmGV+szLY6EOXNVBNDJCVstzj1Pudli5h/eh/swRFS0TPK7oYyShQzWX4gHLtw3EufPcMnY8cfvK3bB7p2FsmTaAk4TIFa/6+/sAAK9cvUC5DRNCztLiDV/4pfc5KsRSRVxRetgihxRqjHmjU8tVGzP68tg/XkZeSHnLIUbkFCwLU7C1bJlCzpmbs6qwZfjnm+7fkVgPbzeHKmwvDo4tEV4e5bnL5+o0oYPZ71ANlkkaoRqlgtPYMnzUcH/emTVKv9/Dscbk3DI2Y8i7Oz1xyazI7Y1y7zyMcm8DOJFG2R7qbZqz72pEBkWVXRNFgM6UagTbZp4N87YzlnuZD4nIi85RgSt3nURixbwVsESC6jZ5+8CQ+AaUe3Scu5zylyFnWSHlfsqyOYG6ACgGI8WTtE/E0R2xcQ/MnPAGxM9PGlvGZsH5YgGeW8Ypc+yicNSR16GqtReDVsKQexvAeTQNuSflPtGFPPE0B/eEA5GQEbsk8omi5CrBwULOSydAiO+o4527uspdjHMPtCNx6+AxiPu7cE26yWEiPfeqn/GR9yvkrPhjiyK8pLFIUQOOdGdiEvsuPFtGV1K75eTEYU5uGb+O448IEjwFNzfoIAy5twG+ck+/TaOIjJbxrCK9DtUcuUmo3HlS+ws5b55QIigneOYoKFTjf/7RmcqyXLmrEoelJQxOZK86biG+9I7TU20bNRNTpWaj31Oxzu+Yt6wEe8StU+5QTVDuXrSMYh0n2yRbRv6cRrkDvnL3bBkpm+atV5+LX3701f52JNdg0CkYcm8DOEfIcd9xaJpyj6hH5dZEKnc4ROOkj/Vv9jkDRXc9xZKG36Hql1k6p19ZtpizkCNSTrOn00m3Yp4/IQdv0qr5Qx456cIZoep//8h3Hsf6v/2JM1m04D9HDWIK1OUpd9leCZNvcD3c7cJYNX8ocjt5e8B/s9LNsc/rLUm5ZWosmM99Rl8ey4XfnMMo987DkHsbwIlah68vWbvY2aZO5S4/QKKUu8qLj1TuFrcp4KWP7S9YnnIHgHOPCc2H7oETv0hExZyabIt5K9pbjyGM299/Hj508Wr837ee4hdvgGE4GJmlAAAgAElEQVRy7sOM47ZHhrF3rIRKzfaIjjFgslzDYCEXT7JRaQQCoZDqNkSt04HYJtU5iAMvVnLf1HgopG0n/a6G1bsFhtybiF88M4Lf/bdfe2T+qTs24oZ7twrRMvGEfd6x83HO0Q5JplH5IuSHQlQ13rRwGnUSyMs1Iir32S65iyNWVch5loC/LCodb07K05IULshx3BEz8eFL1gTSCjcSCpm3LGWnc7lqB0ZrjperGCjmEuLNnb/hqBfxm0q5qxW/WFdsWKL4ZuD+3Popfx3IHariNHvqfTp/TWqCzsOQexPxoVsexUPbD3jTzX3919vxuXu2eBe6yucWUcxZwqxHdSp3zbvq1ccvCi2L61Ald72n3PO5wAQOcZEpnKTEQUV9Arl//vdODZSPOvY4Woq1NergeDHSRMSBiUpgQM9EuYahvlyASD98yRpl2+Rm6EbL1K/chbrcSnQjiPg+D09VUcxbgU5knTh3Pdlg0EoYcm8i/EmSZfUctGXKEVZJMW95N189tszWPWP45588q1WWeQ8aYVnEDUlE3qAW3qHaV7CU5KHe3vkrEovo/f72umWB8pOBmYNI+VmGSqU3EgqZt0hpXY0cLnlD8acqNmo2w2AxH/gt1q2YE9jGf8gE2yO2WdXUXMRDQawr7jJR2TL66QeccgcmypgzUAh2/iYMnDLoDphBTE0Ev65lwcdjo1X50EXkLcu7iSLC02Pxzn9/ALs0Z3RSWQ7RFo6bSIr5g6L4YCOOOAuEl9ONzxaHzIuldKwP1X7rQZRyHxkrYfm8AQBOznzAmaAi7sGjFwqpsl4o8DewLrb1DgIP3zpGqALAgfEy5gwWtO0xDmPLdB6G3JuActXG/zw67H2XlTsnK778wW37vXW/fHavX5D8G1I3ra2INGmDXzwwiR9tfCkwQUeUXWQReTMTiTMkqchDub37fiiqvzjiFR9+utEyKgKMGhmqA0e5R3nuji0zxvPq9AWVe2RUjNSMpFDIuKSWOgSrmtJPPxSSK/cK5gwUE2Py5e0Mt3cehtybgBvu3Yp//qlvh8iWCrcZuBr/z99sV9ZD8G/aeiz3NOF+z+4Zw1XSjExR6pO3izH/oUMULB9ny3jKPSE6JAlpt2lMuVtgzJ2cwiL0FyyvM5nH7Y+VHXIfKOSgyoHjtyO5faoicUSs03GpevimyQoJAIenKlgxb1A7zbDpUO0eGM+9CXhZmvhazufCiVqc7AIATl46O1DO8bbVvn0cHnnhABhjgU7KesDzxsggInfCZEG5UzCqRSRuOakXPyZd5R7YtkOhdVw1c/U+VMwH1hEBpYofSRJU7kFEpu4VIoPUbx7clgm3j/8ucR3ojYRC8uMp1+yQBRer3Bt48zRoLgy5NwFyB2lUtAcnbF5ezthI8G8cXW7/yaaX8ZYv/hq3PLQz9UAdXTgqnSt3H1FEIS/nRCE+AHT7OtN6vUl16IJPasHPpfjbcpuKx4AXJfKL2r+KwLmSVm0eS+4ax6QaJKU7iIk/PCpV5pJ7YGXidgadhyH3JkDO3xLlfXvRMtUIcicx4kZv348PO575jn0TDSv3KPhkZqNUqfnRLxp2g/hd9JD1lbv4Wb3N5usu1aorDTjp8ogZ8YHNzxM/j7KyjfLWY8MdFccWt44j7jIR9+fn1Nedicn5W6nZoXz1OmfO2DKdh/Hcm4BKVU+5y9EysuIPeu56d8e/uJN8DBZzXohes8H99Tse3+V9d/5GePQRnnNSB6K6suh6OQaK6jeWle4Q/dWLZmjuzIc83qAaIHfHTil55C4NvIroUFU134n9t+tQ7sn2XUC5e4nDIosr6y9X7VDunNgRqp4tY9BpGOXeBMgKPCrNbki5V4O3AE+tC6S/OQaLOW1VJuNTv7UWbzltaeR6i+SoleDfcPngGk4GwQ7V9K/vqi3u/9jFkeUvWbsYt7//PPzeGctT70ueyUo8p85DmHxyz8fbMlFx7oDohYe34286qpo9+y76EAITafujhDWVu/u3XLNRzJO+587bZaR7x6F1ponoUiLaQkRbiehaxfoVRHQvET1KRE8Q0Rua39Tuha7n7k0w3UTlztGXt7RVmYyzjp6P5XPDyZ+8dpGcEiC+c04mqgefd0I/eVx4EsTkX8GskMGKF83swxGz1QnIOE5dPqeuB0lIuQvWm9JzF+4keXdx+/c993AZX7nH2DIpBzElzBnuIWDLSPnqddIbG3QeiaeaiHIAbgDwegBrAVxJRGulYp8A8B3G2DoAVwD4YrMb2s2Qo2OiPXepQ1UezETi63a6NlRqTNvHXiIRYiEXzOp4zWvW4HvvP09sVpBk3b9RIpAr9BOOnIV7/uwCb/nO/XrTCv73+87x9xVhZZ+xai5u/8B5aBV8z92ZWnBKGDVrKTx31e/jl4+2UKKm0gPiR9j6v0v0hUIKck/b12EzhD33ZFfGoAug8xw/E8BWxtg2xlgZwC0ALpfKMAB8zq3ZAHY1r4ntxz/95Bl89ZfbtMv/Ztu+wPdoz935y0nhsKRkCf5cnLqvtbPc2ZDKNVt7uP0xkgedt6wAua9fNTcwhZoVUu7+chX48rOOmofjhMkcorxxGYtn+Q+fQIeq8OXiExbjyNkDWvXVA25f2DZDuWYneu46o2dVZ9SLO1fZMjFe/YAXvZNsBwHp89SI11IoWkYDxpXpPHQ6VJcC2Cl8HwZwllTmUwB+REQfBDAE4DWqiojoKgBXAcCKFSvStrVt+Cc3P8sfv/LoxLKqlLpJ85aWI9IPEInxy1pN9W7actXWvgH5CEuOfE7yVCnc+RkgWcTbBVYE+cdF83zuradg6dx4sg6OuIwt2jBE5T5RqgXW8d9DVO42E0fVSh2qVvTbWE6R617eTsXf//i2V+Abv9mOdcvnhFfyuhWhkLq2iVhKniNWaxCT6VLtOHTIXXUm5TN3JYCvM8b+gYjOAXATEZ3EGAuwGGPsRgA3AsD69es7fvZve3gYK+cPYv2qeXXXsVuRy6VmM3zlF2HlzxjzlKAKhPj0Ay8fmsK37t+Bqs3wtvXLcdsjw56iLFdt7aHl/VJUjazMZKXufA/7I1F7U+WSScLvrld3ekY/QFrL7r7nbmO8rHrDokA6XLFTPey5O39Vb2M8U6bSlok5xCNm9+Ojlx4fewzib5ePsX8iNva3zVnSwz1mM/evUe6dhw65DwMQ77xlCNsu7wFwKQAwxn5DRP0AFgDY04xGtgof+e/HAQDbP/vGuuvYO1YKLTs0VcH1d28OLa/Z0RkhAX+wEKBOHPbx7z6Jnz7t/KRfvO+5wLpyzdZWZUWpVy1vBT13xvQUXhTBeiNSJXaq54aPakWrO+4Cyr0cVO6W0wnhdaiG4twlcAWtVO4xdkncdjpoxJYRixWs8Jtd5HYNttmgedDx3B8CsJqIjiKiIpwO0zukMi8AuBgAiOgEAP0ARprZ0G7EoakK7npid2j5xl0HleVtFq3aAXiTYgBq5R53v5SrtrZPL3vzBWn2I7nPIBTa6C2PqJ/bMjK51/GqHkUkrbZlOBlWaywU5cMHMfGfqSDbWlJdfJ0qAspX1NG2TL3z6ari3PXTD0jKXdHnooLpUO0eJCp3xliViD4A4B4AOQBfY4xtJKLrAGxgjN0B4CMAvkJE18DhoHexaRDo+qc3P4p7t4SfYTfc+5yitPNaHuW3A0FvW/XribMMyShVa9oTfMhJvgqWFXgwyInPomyGJMtEJuCY51okokZnts+WYTgwUQ7tW9x9eJBPsC7fh47ej1K5N0rugnRLa8uI7SlIMbY6KQZ6/ubPALSiXhljdzPG1jDGjmGMXe8u+6RL7GCMbWKMnccYO5Ux9grG2I9a2eh2Yce+cYzFxGZv3zeRqj6b+Z1w4s2zYIY/RV1c+oFZA3HkbseS+yN/fYm/D+msy/lG5Cn+eFw3B/+WpNz5Q+Sr71zv1NvE532rlbtIyDuk8ywONuNlg28pUodqzDmNj3N3/taT21+us/EOVbHe5A2ngbbrepgRqjG48HP34R1ffSByfVqyqtm+ch8QElHxEMFJIW+Lqu6hGOVeTiD3uEEochpYOU5fJtIk5Q6JSBpSoJG7iGYYcRBUvfDykjOGFw9MYqCQw6B7nuToIYt0QyHDxx+VMZLXy9tQD4KDmJy/2p676NenyQpp8rl3DUxumQSIk1nIkC0WPsFDMWcpvXWbMS82ur+Q8zrquDKzWbw/G2dFlKt27GQdovaOSg/AIT8kQjaDV0/UvhzIlkM988JGe+7Rv8WPrrmg7jlo/R07fxiAsVIVM/vz3qQrcjoGi6RBXlLTYkMhKfgbBdZ5D8X6DiHYocqjcjSVuxR2SsGVMdu5Hwy7dxxGudeJctUOhUF6YYkR5jITbJl+IeZbjO+O82fjbstyzY5VxiTmGUnwNOR6SJaqLqIIlnnrg/tTtW/uYCG2LVEtjTuE/kIu9i1HB2Lfx3i55tTH31gQjvuOy+fud5KH4Sv38AE17Lk3otwDnylE9jrbGXQWRrlLkP3mKMRFvUTWzRjKNUf9ifnBLz3xCC//Sly2v7gbs1SJJ/ekUDZx05Byh3SzR6jNX/zFq5HPEa7+1iPOPr1MhH7npIgfX3MB5g0VUQ9a3aFKggSdKFUxWMwFOoqDtky8TRQ36jgn1BnerjHlLo9ViNqPetsgmcc9vFQwg5g6j2lL7lEk/n8U8ekcJ3/qHnzkkjX43mO7cNLSWZHlIvfJ/CHrfS65H7toRugVH1C/wse9Updr8SNUA6/oCiISb0Z50o+keHaOFfNdr9ttvDy9HrefeJz96sUzkQTdtMLNhqjcJ8o1DBXzgb4Gedq5APlJjYvrULViFHUzPfdGOlRJeniZafaygWlL7nK4H8dX//d55XLGGA5PVfGpOzcBiPfio2ALoZB8lChPQsX3EZd+IO6+LFft2Fl2LCJ88z1nYf6MIr736IsAgOMWz8Rfvv64QLnBYg6vOWFRYNmCmX2BG9r2yDu6PeJ6L6yQMdz4B6fj+CP0H4zRtkyrlbvzlwGYKFcxd6gYyK/uE324PXLL+MM0rh8lLlqmXojbpw2FDGS5lLaKq8PMxNQ9mLaeu6rD7cXR6KyFDXfQwU1CJUXLhIf6O39ltfbDp17yJuZQoVy1Ey2l81cvwAlHzvKI5JhFQ7jo+MWBMm8/a0WIaOT4eltS5iFwwvJCJn1ye+2JR/gKXwORHaotvnJ52yfKNTw+fBBDxbx3bnj6AQChv8q6YvpR4lBvbn95e0Ds3NZV7mEfxs9LH70dz5XT6jcrg2RMW+Wuiix5cjisxmcPOB1+UUo/DWzme/Xc+iAphpx/kZv3vm8+HFt3uZas3Dn8zjV/mTjxNcfX3rUeu0bDuXO4vRLJ7Qiu9+O164iW6dAgJl79jn3jAPhYhLBv7R1rTD53nakTlccT49XrIOCbp00/oDpGIkB4u1Tho687DnmL8OZ10ZO/GLQH01e5KzI3cj/8eCFNLY9tbopyZwylStiWgaDQfLWmv79CjuqKcxdvUb6luOyi4xfjHWevDC2vJSh3ebFoyzQLrc4tw2vnb1oXrFkYSKsg2zFBW0by3DXusljPXb/Z0vb+57j0wUkgeduYSuYMFnHd5SeFMo8atB/TltxVU+GVpc5OwJ/8Oi6GXBdMyC3Tn/dtGRXxpuHBvnwO5VoSudffueaU9T/z/URtzZfLxFdX+oEoW6bVr/1u/SVxQg7yV8nzm8Z3Zic3VlWi0UMU2xiVhjlpW+czvHqa0S6D9mBakfvI4RIecCfWUBGhT7z+z8IfAsPSLEIz+9M7WjWhQ7XP89zVA2DSxDYXcgTG4pWxKiyuUURNDiKHSnrx2k14QHr7brlyd+oXc7aLD0W5QzVuEJPOQ1TpyjR4iGI6+KjwVZ32iFFCzWiXQXswrcj9rf/2a/zejfcDUCtxbpmIMwbxiTfe8IVfBsqK067pQswt43WoWlEdqvr15nNO4q845R6cco3vwy/PVL6MuL0qa2GS5+7+5YO03nnuysj2RaFTyp3vl+dpL+ZJ8J7DbyVxhKfTVuVkHXW8xYkQ57pNW1dw2sCgLWMiYrKBadWhyhNAMcYwIUzAUKnZ2D06FbJM+DoVKhGzLcVBTPnLPXfnFd8twPiSdMo9bxFKdjy5i/CGwwvLuMef5sZNikHn62f1F+rOmR/VnnZNxByYJ9Ujy2jrSYV63zIaV+7+gydtXcG4ffVfg+7GtFLuHJUaw0dvfcL7/snbN+KCz92LkcPOxBviTEVR5F4PmDCIqehaP060TNjfTPPoyOcIdoJyF+GRTYMuSWSHKnxSaRXaZLl7D+NCzvIHHEGwZdxyKjKEYl3k/hQ/VtxoZR2IoZu+ftCrK66Dt9WWmEFz0LPkzhgL3RT8mizXbDzygh/2+KONLwEADk5WYJFzI3PYrHlesW0zlKo1FPOCf+v9x9uY/oYu5CzYKZS7aoRqEmIi9SJXNIMCoppayLX20uXnoSxMgi2mOpb959homQ51qIpK23uea19WYh+CZMsYbs8EepbcX/v5X+CUTwXTyhdc6XX1Nx/GHCFhFffPD09VAsTLUak3obYE3qHaFzMnZT2euzPZhn64pi/c/fJL5zi++PJ58ZNUi4hW7tKOGkBUDbrzxda9Xy4EVNEyQoSTqpMxKs49DnG5ZeqVFmLb0tpYqn6gdllhBs1Bz3ruz+4ZCy3LWQTUgF8+uxcXH78IP316D9YsnoHnRpyBKocmqygKr98c1RT++mAxF5pzk4Mxx+Yp5q2AqhJj2+PSD0QhZxEY9OPIOTGKxS87dQnmDhbxytULtPebdK83g3+jCEXOQd9s8NpLgi0jKld5wpLYxGFace4qW8b9UG+HqvfXr1u3KtWbiNyXYtDd6FnlLqJctbH+b3+MSSHChU80bRF5ivfQVAV9hVzo4k1D7gOF6MEbPJ97UZizVPRvgfpDIevx3MVdEBEuWLMwppM0vDxJkbYyqqITyp1DTBmh+g2aFgqp19SYSsN160fLCJ8lm81QezbQs8pdxN6xEvaOlZXrREI8NFVxlLt09aaxZYr56OelzRhePjSFRbP6I1/j6wl/y+cs52Gg2Uw/WqaxvoQozmpmVEWnbBm+Z07uRYHcCfEx32k899DoT8W6RnPLyPmLdBC4PqX2GOGeDUwL5R6ngsV1hyar6FN47mmUe1x+l5oNbN83jlXzB/19UJDQ/Y4vf58HJyvx+7Qoledej6+vridJuTeOqF20Xbnn/f1ZFoUILw5xTfXOs8YbQFo0orRV4kMcFGXQ/ZgW5B5HYuK6Q26HqnzxpgmHLMQYrIwxHBivYN5QMfIGUXWibXzxYPw+c+k6VHN1KEJVe5P4tSnKPdJzb3G0jPu3VHWsvKIUncN3r/oNmt6hWnfiMP7BX6YfChkwZgLLjOeeDfS8LfPDp16KJT3RcmEMAT/cK5OC3KOG5AP+IKZAhyqCN4vsuds2w6fdHPJR8OLcNUkgro1pEOnPeyMas+y5O/VPVmzkLUJe9txjjlFeosOFSu9ev7lKiCmX054L47lnHz1P7kmpcidKwcgWlS2TJiNkHOfUbDdaJpCnJD4Ucsf+CWx5+XDsPvOWhRpj2jZLPb6+elBL+m2ahTj7qxngtU9VaqGZqZwIJ+dzvakFwmUUy3iaiORdKKHs+9DtUBU2klMsGOGeDfQcue8bK+GF/RPa5eUcM06ce7BMmpsr7kau2jYYCw/ACZKt85kr98mIsEoRPHGYLvzd62+k7DhM8txbyALt8twny7XAiGUgmKY5NoTRhU5blfUg/UM4uL3/1x/bkG7bQD0qn8ega9Fz5P6WL/3ayyGTBIuAqmS5OHHuwYs3TVhi1I2cs8gbLCXmKZFDIeXNp6rJ5J6PGa25ZvEMPPNyMOa/0YRUfj3x61tJAa2Pc+e2TC00E5Vjc7jllM0ILozvUPXrDNXiEXLdmcO8utP+WqoJ1eOP2aDb0HMdqjrEfsSsfgCOnVFRKHf54k0zQDVKueeIhNzg4ghH2ZYJKnf+QPjz166J3GdBwR580Rfffjqe/8wblG1sNKlCYrRMC0mgbcq9ElbuQLr86PVG1DT65qMiY93OWVW0jBw1Y9Dd6Dly1wFPBpXPUchPL+ZzoRs2jXKPuvAty08prIrIAdwRqrxD1X2g8G2G+qJfslT+sx/jHCaJeqIwlB2HEcfaDm+21dEyHOWqHfLcnbjxoA8tIk20DFflLelQVRB0XYOYhI5Z8btBd0PrDiGiS4loCxFtJaJrFes/T0SPuf+eIaLwZKRdhIqroHMUJvc+wXPnr/6MAacum61Vd1QkiqPcfVtGnNNSOYjJ/c6Vezy5q5RltCeca7CjTt5HFLIdLeN/7pMGpgU7VJNJWS8rZHhZw/ncxeRfac+FwnQ3tky2kOi5E1EOwA0ALgEwDOAhIrqDMebF5zHGrhHKfxDAuha0tWng4Y85heIVZ9zJ5whVm6UKM4zKuGhZFJiyTfb6ZXi2jPtAGCpGnyqVLSP7pIG21HF36ijUtOsbQbs8dwBq5R6T1jjqTUkFHc+9Xvjb+xXVk1tGFgqG3LMBHeV+JoCtjLFtjLEygFsAXB5T/koANzejca0Cn2hDRRA5y7+IeVSLk7dFr+4o5Z63gp672Nml3N69C31bJjpnjajcr3nNGtx29bmxMwTVFQqpWJbsubdQubc6FFKoXn5LEN+2dFqh8zPEPTzrDoVU1VWXLROswwxiygZ0yH0pgJ3C92F3WQhEtBLAUQB+FrH+KiLaQEQbRkZG0ra1aeBWjOrVnkAhWyZNTvcoQZmzKDDaUVWOMf9GshlDtWbj2u8+CUDfcz9idh9OXzk3MFGDqi1A47ZMonJvsP44tFq5i8h5Fpr/1+9c1LFlktuqrqcxdveFRvoKVAPr4o7ZoPugQ+7qvj81rgBwK2NMGb/HGLuRMbaeMbZ+4cKFum1sGVSdco6fGlTVLI0tE0E6FpGnwp30sSptFPTc9437yc7mDRUj9ymmPMi5n+OUu6/iGqP3yHzuEhm0Au303HPCdQA4ZytuVqJUHaqeLRNe5y1r4qHWMxOTPBrXRMtkAzrkPgxgufB9GYBdEWWvQActmS0vxY/klKFW7uGLN51yV1/5OdGWyVuRr7hetAxjOCQkDIvLWSMqd+7QxBFrXa/VdURztLJDtfW5ZYQHrnRBiFkWdX7KuEmjvLlrlQ8JTqr1QfTc/Rm+9LZVTRtoEodlCzp3yEMAVhPRUURUhEPgd8iFiOg4AHMB/Ka5TdTHn9y0IVX5KPUndxyl6lCNUe48w2DeIuWNwiCqamDUJffXrl0c6zHLucYBBKJxZHgv663OCtlCFmi1egxGMIX37XeoxtgpXl112jIN2iDBaJn0W8uf/AnCDb1nAYnkzhirAvgAgHsAbAbwHcbYRiK6joguE4peCeAW1ui7fgM4PFVNVV6p3BVRAtUaSzHqVX3h53Pkx9dbwUROylBIxjA64ZD7By9aHTvvqeg/50O2TDRppBn5qNp7O1L+RtbdYoIRa0/bodpsW6beQ23kJwrYMtIbhOlQzQa00g8wxu4GcLe07JPS9081r1n1QY5HTkJUp5ysmH64cbd2nZEdqkSBjtyo13r+1WbAnsNTAID5M4qxHrMYLcM/xr1C12OXKN8AEn7uLJOA6oEb7FDlb0jJddWfXKwxpeyTcfp+FnGP3u4le8agu9FTI1Rn9KdLlaNW7uGOozjVLCPqRhR9W4fco+0bwLkJd+ybQDFv4YhZ/bHRIaJjwztUdUZQtsqWiYsBzw6C50tekyb9QPwIVbfOOOWeuAc1Gvn9g3OoBpcZWyYb6Clyn6qkSAIDtXIXE3lxgkw7WbVyOUnk7u3PhzgTk82A4QMTWDZ3AJZFgYfDm045MnKfsnJXoVmee9ItnmUKUCl38bvq/Km2VX0PIHYQU2Ps3kiHdpwt06bMDwYNoqdO00S5msqaiVbuDvhFPF52vPyFM/sS69RV7pYX1SIFRQqhkKMTFcwbdEIgxQfRF65Yh63Xvz60jVN30HOPnYWq1Z57hhWe2HL5OMVw2Xo7S0P7a6FyJ6EO/ZS/gnIPWYjZPa/TCT1F7uOlWmw8uIyoUEjZc+cTegwWo0eJcsSNUBX3G6WqRG90dKKCOYOFUFstaWYgVUy2790rbucm2TJRnNWg4OwKBB+YwXViKKRONsdGBzE1Y+astA/agHJ3/4rJ6Ay6Hz1D7jWbYbJSw8wUvnuS587XcuU+UEgm96iIRfEGzYsdqoFSLKC4D05WMHugGNlWr27J8nGOI1q5e5NAxBxHaBsFOXQyFLLVEJvujejlg5jIL6EOhQxCb7KO6GX1K/fwltrpB0j9GTCee1bQE+T++M5RHPNxJ5hn9kBBezv1CNXw0PLxkkPucSkAOKKzQgqfrfA8rd7+3b81xnBgouwrd7f86kUzwvsM1M3VnvNdSe5p39GhJpikezzLJKDynMXvcdPsyYddiBmjEJvy17ND6vsdVQORtEeoBmyZ4IPMKPdsoCdmYvrZ03u8z7916hI8tP2A1nY5SU1XbaYMAZtwp7pTTdogI26EqvdZ7JCj4M3Lt98/XsZEuYalcwac5RbhpveciROOnBWqW7wRPXL3OoNbN+wgkbxbQALf/+D5WuehUQR+UwW5DbgWnU6nZZxyj80K2eBwf4r4rLWt0pbh6wy7ZwE9Qe7zZ/g+uzwlWhxykg9etVngLuDrx8tVDBRyWjdyXOIw73OOhJtZVoXO3+f3jgMAVs4f9Na9crU6Hw8p2hxnT/n5axrNLRO/vhUUcNJSvbz6jUL1m3rrQBh0UzAr7RTpyOU5c5P2x+G/WNb3SzZiywSn2eMPt/A6g+5FT5A7Hxz07feeheH9k9rbiTdtIWehVLXdUMig+p0o1TBQzGl5yNrKXVFMDIU84CYN0+kgFvfJO25v/IP1uP2xXYGHA0c9ce5qTzjJc+8NEgiHQgJDrnLXSV5iXnsAABinSURBVBwWq9xj9tuwcle8hmpHyyiUu0kcli30hOfOPfHTV871hvjrQPRCxaHevmXifBorVbU6U/n2KsidngFbRlHOt4J0InTC2y+ZM4CrX3VMLMF2LE9EBhBU7uF1gyneEOMGoImZJmU0nH6A/6X02l9lT4ZDIg26GT1B7odLVRRzFvryOVRSkHtA8SpCC/nNVaraKXzeZOUeHS3jf+fT6+mQu2gDqOZTjWpho3OoJqGDaYYahqofRFzHlbtKTITOaZ1x7v410gzp7qCuaBkpMsjYMtlAT5D7eKnqpR5IQ+5y7DmAwIAP8SIejJnmTkTUdS+OUI2LWw4r93R+bZoBM6mod5rdz4G3KUX6Aa7ceabPyI0baYPkdafePlAX/6Sbz1303IP19Yrd1uvoCXIfm6p609BdfMJi7e1y4iQXwo2kuqkGCjkt1fPGk49ULpeVO69L3IeY8neSK/e8hi2j8NzjcOyimQCA9114TGJZg3BuIYvIs+nSiIm0UImMVNsLyj+t+o+zZYznng30BrmXat4E0scsnKEd665W7uHcMgDQX8xpRZesXjwD2z/7xlA8uhyZ4yWMQrBzle9zMoXnLj8gkjB7oIDtn30jXnfiERql06MXlF3gbUierMMCijFpLtIQaWyHaoO/o6od+raMoNy9jlRjy2QJPULulUDon+61J06Awb1qsfNJvKf785bWjRF1Y8u5ZVTK3StLvperkytHvNnShIKmQbtv5/OPXdDmPQYRmImJwuuKMX0babjPv6bCGzUzn/spy5wQ0gvW6E1vqVTuUrsMuhs9EQo5XqoFYt11EeW5Q6FQinkL9lQyu8flc1d9Dqhub3g7AYyhmLe08oqIRXSSm7ULjXSnfv3dZzjjDjqEQLRMqEPVj11vVp9xHIE3I5/7SUtnY9N1r9PuO7IU1yiRtMCgq9Ejyr0aUKy6l57SqyZSKpRiTlO5J8S5EzkqXrR45E34fvUzXDobtEq1A+nu52bc+vmcpWVJtQqxHapEPrkrHmFpjj/ukrJj3u50II9n0CV2eZ9mguxsoieUe4jcNe8GUbl7KVyF9SL5F3TJPWI5J3e+T8+WUWzhLGNaCaecdrrbtfCma5aPnBmIyl323ElMJqbYtA5fRrVFXAy8DprV98EPPy4PjkH3oTeU+1S1LtUqeu6islYNsy7kCUfO6U+sM3KEalRnlKJ4VAx80j7NLdc8JHvuXLk3aX+qmPSYdVp1NtSe8Oe4+V4Nug+ZV+481e9QHbZMZLQMf/0UHn2FnIXrf/tk3P7YrvhKI3bOX+09xacoI0+5lqSQvv/B82Ezhn1uqoJuQS/c+3FjB6yA5946Wyau010HjcTJB485bEsZdD8yr9x5rvW6omUscVRqWLmL6q2Ys7TeDqITh/H10o0i7Ucsk3QcJy2djVOWzfGtnsTW1Y+67ucM+zPi4YYShwmeu3LbJnEff3A02qFa17Yq5e5+N7ZMNpB55T42pcq1nt5zzwvRMqqgAJ3Mfs426n3Ls9jEDc0nxac48Lz03TLavxfu/fj0A35eItVPruqf+NffX6c1H4AIfyxEfWjkPFDMVdgL53c6IPPkzpOGBTtU9baVMzXK28sdqjqI2rVH7goVyCHn9tY9Dh6jn+VcLt2GYMd6cJ0zzZ4kZxPwplOWKJf7Hesx6+q2ZaLr1t3W+Ry8voznng1kntwPTVUAwMstA+hfzCK5i7G8qpCvQl43ciVJuYdtGbnBaW/KbrVlGs0X30kELeewco/l9jp+K9Xve9SCIZy5ah7+8vXHpa8Q6SKc4toj12JsmWwg8+Q+OuGQ+9xBfxCTtuJVdagKQ1QDg5i0bRn1cnlatjiRnTb7XjtsmUaIIouIi5ax/EtE3aGa4qeKewAW8xa+875z9CtroB2hbQPHHxQPpkM1G8h8h+rBSYfc56SYO5VDVO6BEXn8bz22TBS5W2rCJgJOXOJMnfeOs1e6ZeLrksFtmVZOqTftEFCuIdcZR8x2wmL/v1cd29Bu3vvKowEAK+cNNVRPHOohYxMKmX30jHLnE0kD+iozH2XLSJ2fQBrPXb1v2ccXFduimf3Y/tk3CmX9zl0ddJst0wv3fjAUMrjOIme0p3jOAtum2M/lr1iKy1+xNH0DNdCIwo57azS2TDagxVhEdCkRbSGirUR0bUSZtxHRJiLaSETfbm4zo7FvvIScRZjZL5C7tuL1D98fCBQmfKesXqVR+85JD4y4jrS0ExF7byBGuDcNwQ7V6E5w5bZdQn6NtELV5eCHQjZQsUHbkKjciSgH4AYAlwAYBvAQEd3BGNsklFkN4GMAzmOMHSCiRa1qsIyNuw5h2dyBYOeo5rbiRBjKafaEmuTkUVFI6lCVb3w1EaS7e+LynHQSWXaJAudF6uDOCrk1orBV0TLKlQZdCx3lfiaArYyxbYyxMoBbAFwulXkvgBsYYwcAgDG2p7nNVOPgZAX3bRnBghn1ZUIcKITz0RCCFg2Hpdk7EXXZex2qbj1xxCeXTQJ/sHUwiWIA3aJcG4FKuesO4umWo29VnHtWHm7THTr0sRTATuH7sLtMxBoAa4joV0R0PxFdqqqIiK4iog1EtGFkZKS+FgvYO1YCAFx0fPBFQZdcBop+1sGAcldEy+iqoKhicgRM3AAVlUUUh3bEuach7F6Lt0/LZd3ybGvIllFMuo4GR8watBc65K46k/LdmwewGsCrAFwJ4KtENCe0EWM3MsbWM8bWL1yoN2lAHCZKzmxFaxbPrGv7wQC5+4RK8ns49MktaYQqB4sx3VVvDnHwQiH1ihtoIO6tLSvk1phyD9djPPdsQSdaZhjAcuH7MgBy9qxhAPczxioAnieiLXDI/qGmtDICPK/MUDGY91v3oh4oKMhdUO5iNbqeexTkYBtxmj0ZabM8+sq9zsZpIKot337vWZgzEJwopTdsGVJ+BpKvr+4ZE9CI5y70N8WMqjboXugo94cArCaio4ioCOAKAHdIZb4H4NUAQEQL4Ng025rZUBl3PbEbn7l7MwB/JnoObXIXHgpJ2zSqVuQbgj8s4ia0ricvfbtx7jELsNaN05eRZXdGFefNkei5dwn3NUu5F6RBcll5c5nuSFTujLEqEX0AwD0AcgC+xhjbSETXAdjAGLvDXfdaItoEoAbgLxhj+1rZ8Pd/+xHvc0i5ayqWgHIXRqiqSLVRtSKr8TedeiSeGB7Fn71mTbisxfepV3det+e1Tei1Wz8c4dShhqREI80MzFImhQFn5PCnPbQGMTHG7gZwt7Tsk8JnBuDD7r+2o17lLg5M4kqaMabsEGxUHMvb9+Vz+PTlJynL+tOa6aEdyj0rhNYsqAYxZS0UspF87qoxHt5MTN2lJQwikPkRqgCwoI7JsYHgTZp0A+hOeRe9r2CUTHxZ3ia9fVoW4fSVc/FH5x1VZ+uSkebor7lkDbaOjOHMo+e1rD2tRiO5VbrlQdisZhQkNje2TDaQSXIXlfXK+YPoy8u2jB6icnarbt5GL2g51a9Ou9Ls8barz03ZotbhpKWz8fO/eHWnm9EQFGOY9LftEuOiIc89EC0U7LA3HarZQCZfsCo1n9zLVTu0XlvxBl694yNOoqqcP1TEecfODy3/o/ODKlp+tY9D2lDIdkD8TdevnNvBlrQHwVDALjoRKdDIdaQSMyZxWLaQSXKfqta8z7UGhmWq0rpGDeGPUu4PfPxifOuPzw4tv/LMFYHEUmmUf1Tu927Av//hetzaRW8JrQIF3uTSbtvkxtQJShAssdvGrOvG69IgjGySe8Und1WaW91rTxyFl8Zzj827EYE0aqebb50shzemgXgO0o476Jbz16xomWbWa9A+ZJLcSxXfilEpd90BR6obOMmWefJTr8VTn3qdsJ3WrlK92kclGeskeEumCbc35Ll3CxpK+au4sM1kHdlCJsldVO5Kctdk3EDumISc6LzszP5CYKLjpAt9tjuJSJpoG9UIWYP2ItDZnvIu6Rby461odnOM554NZDJaZkpQ7irLXd8qEayWFGXT4CcfvhAjh0t48sVR7W0aiU9uFfzcUdNFu/tIS9bdctqaff0wkzgsU8imchc6VBfNDKf71Zw0SRqoErZlRBrTrVPGwpl9WLtkViqC8AfLdNNN1E1taS+yeuStCsk0g5iygUwq97EpJ2HYO85egasVc1hqe+5CsZxny6iVabPSD2iVTZl+wKC1SDMADeie8+bbe81pkIlzzxYy+QwenSwDAN5z/tFYOmcgtF47Pa9iSj1Ruas6XOtFGp+SdxiPHC41tM9WYPqZMvWEQnYH+bWqFd31RmkQhUwqd29S7IGCcr1+h6r4OX6bRjuR0nSoPrtnDACw++BUYzttIqbz/awbCnnnB87HPRtfan2DdNHgOfvX31/nvSU3sVqDNiFz5M4Ywz/86BkAwKwIcq8nPNGfuDrdICZddIuaaxTTsD9Vm8xOXjYbJy+b3dK2pIGXgK7OS+9NpywJfPcSh/XItdzryJwt84tn92Ks5KiJKDWse/EFlLuUP6PeOnX2lUX4zZ9+7M4fzH/9prVYPKtPaQV2I3g2x2qtuecs69fydEHmlHtJiHGPgv58p2IoZIIt0+BjMOtqJ+PNbwj82F99/CI88PHXdLYxKdDvJtQr18L5l+qB6VDNFjKn3HVQT3peP7dM1Pr2Rcu8+RVLkgsZtA1ZfTD3F5zbW5Vcrx6YOVSzhZ4k9zTpdTlyCltGJPp22jIfvuS4hvbVSkwnz52f8qySWb8701ipmvy2mwb13F8G7UdPknuujmuPIj5zND4TU/r0A92EbslR3k4U3ZFr3Xg+dNDnKvdSk5Q7R0Z/jmmHniT3RlR22pS/uuCevY7q6WZlNI2Euzd9YVY9Zq7cm/W2xSPJsvp7TDf0JrnXQ45C+gHVvdCsUEiderqR21WDvHod3qTpHW5HveiXZihrFrrx+jQIoyfJXTf9gAi+BVMsAxqPlsl55K5ftpvQhU1qOXg/TFY7VAv1+JMx8DtUs/l7TDf0JLnXQ8RJ12uzomV06jGvvd0B/pDN6ulo+nXkTbOX0R9kmiFzce464BffX7zuOMwfKuK4I2Zi71gZ7/3GhuSNWzRClSv2rNoyHFF9Er0IK+PKHQA+/3un4sQlzRk160/W0ZTqDFqMniR3/jq9ZE4/fnvdMq1tkjL/NUq4nueu8VbRjWQyHaNlvEnNM3zoutd/GmT595hO6Ely//gbToDNgEtPPFJ7G89zj4pzb5DdUyn3Lpbu06lD1bNlpuGDLQ7dKD4MwuhJcl88qx//cuW6urZlYBFx7o1d0LkUYXVdye0JI3h7EZ4t05M9U+lhZmLKFsxl6yIp1K9ZhKsTwNCNN0/3taj18B7I0/Loo9GV4sMgBC1yJ6JLiWgLEW0lomsV699FRCNE9Jj774+b39TWghI893y98+y5sFNEGnQjuU9HpAlfnQ7wO1TND5IFJDIWEeUA3ADg9QDWAriSiNYqiv4XY+wV7r+vNrmdHUdeusPPPGpe4PsJR86K3d5O8UrbzTbAdJog20sD3eF2dAuYJ1A62w4DPeh47mcC2MoY2wYARHQLgMsBbGplwzqFqBGqMrnf8t6zA9/v+uD5sfXa3tDt5DZ0o3KfjmqNK3d7Gj3QdDAdr4UsQkcjLgWwU/g+7C6T8TtE9AQR3UpEy5vSOgVadZt5nruwB/ESltMIWxYFolrk7zI4P+ikI+5Gcp+O4OezZhtyF2GuzmxAh9xV51K+2u8EsIoxdgqAnwD4T2VFRFcR0QYi2jAyMpKupXzHLbrPkjrNGlUrnCCyOoipC5vUcvBuFru5SRUzCy58jPbIBnTIfRiAqMSXAdglFmCM7WOMldyvXwFwuqoixtiNjLH1jLH1CxcurKe9rfd8I+LcG4XnuWswdze+9k7HxGHclqlNp4PWgIkeygZ0yP0hAKuJ6CgiKgK4AsAdYgEiEkcLXQZgc/OaGESr3pB9W6Y1KtWboqwFdRu0Bvwha2wZB+YZly0kdqgyxqpE9AEA9wDIAfgaY2wjEV0HYANj7A4AHyKiywBUAewH8K5WNbhVKsofodqa+rnXPqM/2+PGplNumVkDBQDGhuDw51DtbDsM9KDFNIyxuwHcLS37pPD5YwA+1tymRbalJfW2+oI9/9gFuOY1a/DOc1a2dkctwnS8n//xbafilgdfwLrlczrdFAOD1MicjGxVWBr3EVv16mlZhD99zerWVG7QEiyY0YcPXGTOmUE20cXDZdRoVeQCV+7GXlXDG8Frfh8Dg0wgc+TejsgFw19hqLJmGkwvsBQD8Qw6j8yReyOe+8lLZ+Nt6+PzW0cNYjIwMHDQjaG6BmFk0HOvf9s73RQB39kwHFqnsh2MSA3D/CbTF+bcZwuZU+6t6lA9bYUTEfHK1QuMYlfgohMWAQBOXBKfIM2gd/E7pzlvvbPdEFGD7kb2lHuLejzXrZiLTde9DoPFPO58fFfyBtMMbzplCS46fhEGi5m7ZAyahA9fsgbvf/WxGCjmOt0UAw1kULm3rm5DXPEwv8/0hmWRIfYMIXN3a7vTrz74VxebCBEDA4PMIXPk3u48H4tm9rd1fwYGBgbNQObIfbqo6P+66mwsmmUeLAYGBvUhc+TeDluGz7rU1+C8qY3grKPnd2zfBgYG2UfmyP2oBUMAgP/8ozPrruPr7z4DY6UqAODb7z0LLx2cCqy/ZO1iXP2qY/AnFxxdf0MNDAwMOgjq1ITH69evZxs2bOjIvg0MDAyyCiJ6mDG2Pqlc5kIhDQwMDAySYcjdwMDAoAdhyN3AwMCgB2HI3cDAwKAHYcjdwMDAoAdhyN3AwMCgB2HI3cDAwKAHYcjdwMDAoAfRsUFMRDQCYEedmy8AsLeJzckCzDFPD5hjnh5o5JhXMsYWJhXqGLk3AiLaoDNCq5dgjnl6wBzz9EA7jtnYMgYGBgY9CEPuBgYGBj2IrJL7jZ1uQAdgjnl6wBzz9EDLjzmTnruBgYGBQTyyqtwNDAwMDGKQOXInokuJaAsRbSWiazvdnmaBiJYT0b1EtJmINhLRn7rL5xHRj4noWffvXHc5EdEX3N/hCSI6rbNHUB+IKEdEjxLR993vRxHRA+7x/hcRFd3lfe73re76VZ1sd70gojlEdCsRPe2e63OmwTm+xr2mnyKim4movxfPMxF9jYj2ENFTwrLU55aI/tAt/ywR/WG97ckUuRNRDsANAF4PYC2AK4lobWdb1TRUAXyEMXYCgLMBvN89tmsB/JQxthrAT93vgPMbrHb/XQXgS+1vclPwpwA2C9//L4DPu8d7AMB73OXvAXCAMXYsgM+75bKIfwbwQ8bY8QBOhXPsPXuOiWgpgA8BWM8YOwlADsAV6M3z/HUAl0rLUp1bIpoH4G8AnAXgTAB/wx8IqcEYy8w/AOcAuEf4/jEAH+t0u1p0rLcDuATAFgBHusuOBLDF/fxlAFcK5b1yWfkHYJl7wV8E4PsACM7Ajrx8vgHcA+Ac93PeLUedPoaUxzsLwPNyu3v8HC8FsBPAPPe8fR/A63r1PANYBeCpes8tgCsBfFlYHiiX5l+mlDv8C4Vj2F3WU3BfRdcBeADAYsbYbgBw/y5yi/XCb/FPAD4KwHa/zwcwyhirut/FY/KO111/0C2fJRwNYATAf7hW1FeJaAg9fI4ZYy8C+HsALwDYDee8PYzePs8i0p7bpp3zrJE7KZb1VLgPEc0AcBuAP2OMHYorqliWmd+CiN4EYA9j7GFxsaIo01iXFeQBnAbgS4yxdQDG4b+mq5D5Y3YthcsBHAVgCYAhOJaEjF46zzqIOs6mHX/WyH0YwHLh+zIAuzrUlqaDiApwiP1bjLHvuotfJqIj3fVHAtjjLs/6b3EegMuIaDuAW+BYM/8EYA4R5d0y4jF5x+uunw1gfzsb3AQMAxhmjD3gfr8VDtn36jkGgNcAeJ4xNsIYqwD4LoBz0dvnWUTac9u0c541cn8IwGq3p70Ip2Pmjg63qSkgIgLw7wA2M8b+UVh1BwDeY/6HcLx4vvydbq/72QAO8te/LIAx9jHG2DLG2Co45/FnjLG3A7gXwFvdYvLx8t/hrW75TCk6xthLAHYS0XHuoosBbEKPnmMXLwA4m4gG3WucH3PPnmcJac/tPQBeS0Rz3bee17rL0qPTHRB1dFi8AcAzAJ4D8Fedbk8Tj+t8OK9fTwB4zP33Bjh+408BPOv+neeWJziRQ88BeBJONELHj6POY38VgO+7n48G8CCArQD+G0Cfu7zf/b7VXX90p9td57G+AsAG9zx/D8DcXj/HAD4N4GkATwG4CUBfL55nADfD6VeowFHg76nn3AL4I/f4twJ4d73tMSNUDQwMDHoQWbNlDAwMDAw0YMjdwMDAoAdhyN3AwMCgB2HI3cDAwKAHYcjdwMDAoAdhyN3AwMCgB2HI3cDAwKAHYcjdwMDAoAfx/wB1s3kTVn2UmAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "save_path = saver.save(sess,model_path)\n",
    "\n",
    "print('Model Saved in file : ',save_path)\n",
    "plt.plot(mse_history)\n",
    "plt.show()\n",
    "plt.plot(accuracy_history)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test Accuracy=  0.8333333\n",
      "MSE=  16.56451111439289\n"
     ]
    }
   ],
   "source": [
    "correct_prediction = tf.equal(tf.argmax(y,1),tf.argmax(y_,1))\n",
    "accuracy = tf.reduce_mean(tf.cast(correct_prediction,tf.float32))\n",
    "print('Test Accuracy= ',(sess.run(accuracy, feed_dict={x: test_x, y_:test_y})))\n",
    "\n",
    "pred_y = sess.run(y, feed_dict={x:test_x})\n",
    "mse = tf.reduce_mean(tf.square(pred_y - test_y))\n",
    "print('MSE= ',sess.run(mse))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}