ionosphere_DL_classifier.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "<b>Ionosphere Data Set</b> \n",
    "(reference to UCI Machine Learning Repository <a href=\"https://archive.ics.uci.edu/ml/datasets/ionosphere\">description</a>)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# coding: utf-8\n",
    "import multiprocessing \n",
    "multiprocessing.set_start_method('spawn')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tb 2018-09-12 \n",
      "\n",
      "CPython 3.6.4\n",
      "IPython 6.2.1\n",
      "\n",
      "sys 3.6.4 |Anaconda custom (64-bit)| (default, Jan 16 2018, 12:04:33) \n",
      "[GCC 4.2.1 Compatible Clang 4.0.1 (tags/RELEASE_401/final)]\n",
      "numpy 1.14.2\n",
      "pandas 0.22.0\n",
      "sklearn 0.19.2\n",
      "torch 0.4.0a0+9692519\n",
      "IPython 6.2.1\n",
      "\n",
      "compiler   : GCC 4.2.1 Compatible Clang 4.0.1 (tags/RELEASE_401/final)\n",
      "system     : Darwin\n",
      "release    : 17.5.0\n",
      "machine    : x86_64\n",
      "processor  : i386\n",
      "CPU cores  : 24\n",
      "interpreter: 64bit\n",
      "\n",
      "GPU Name: TITAN Xp\n",
      "GPU Memory: 12.0GB\n",
      "CUDA Version: (9, 1, 0)\n",
      "GPU Free/Total Memory: 87%\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib.pylab import rcParams\n",
    "rcParams['figure.figsize'] = 14, 8\n",
    "\n",
    "from IPython.display import display\n",
    "import pickle\n",
    "\n",
    "import numpy as np\n",
    "np.set_printoptions(precision=4, threshold=10000, linewidth=100, edgeitems=999, suppress=True)\n",
    "\n",
    "import pandas as pd\n",
    "pd.set_option('display.max_columns', 30)\n",
    "pd.set_option('display.width', 200)\n",
    "pd.set_option('display.max_rows', 500)\n",
    "pd.set_option('display.max_columns', 500)\n",
    "pd.set_option('display.width', 1000)\n",
    "pd.set_option('precision', 6)\n",
    "\n",
    "from collections import Counter\n",
    "\n",
    "import seaborn as sns\n",
    "sns.set()\n",
    "\n",
    "from imblearn.over_sampling import RandomOverSampler, SMOTE\n",
    "from sklearn import metrics\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.model_selection import ShuffleSplit, KFold\n",
    "from sklearn.metrics import roc_curve, auc\n",
    "from sklearn.preprocessing import MinMaxScaler, StandardScaler, RobustScaler, Normalizer\n",
    "from sklearn.model_selection import learning_curve\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.model_selection import cross_validate\n",
    "from sklearn.model_selection import ShuffleSplit, KFold\n",
    "from sklearn.model_selection import GridSearchCV, cross_val_score, KFold, RandomizedSearchCV\n",
    "from sklearn.model_selection import train_test_split\n",
    "from imblearn.pipeline import Pipeline\n",
    "from sklearn.model_selection import validation_curve\n",
    "\n",
    "import torch\n",
    "from torch import nn\n",
    "\n",
    "from lib.utils import *\n",
    "\n",
    "from lib.classifier import Classifier, torch_weight_init, Stopping, OutlierSampler, exp_decay, to_var, PytorchLogger\n",
    "\n",
    "from lib.model import MLP\n",
    "\n",
    "%load_ext watermark\n",
    "%watermark -a \"tb\" -d -v -m -p sys,numpy,pandas,sklearn,torch,IPython\n",
    "gpu_stat()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Hyperparameters\n",
    "HP = HYPERPARAMETERS({\n",
    "    \"GPU\"                 : True ,\n",
    "    \"SEED\"                : 0\n",
    "})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# if not HP.SEED:\n",
    "#     torch.cuda.is_available = lambda : False\n",
    "\n",
    "# Seed random generators for repro \n",
    "np.random.seed(HP.SEED)  \n",
    "# torch.manual_seed(HP.SEED);\n",
    "# if torch.cuda.is_available():\n",
    "#     torch.cuda.manual_seed(HP.SEED)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "([0,\n",
       "  1,\n",
       "  2,\n",
       "  3,\n",
       "  4,\n",
       "  5,\n",
       "  6,\n",
       "  7,\n",
       "  8,\n",
       "  9,\n",
       "  10,\n",
       "  11,\n",
       "  12,\n",
       "  13,\n",
       "  14,\n",
       "  15,\n",
       "  16,\n",
       "  17,\n",
       "  18,\n",
       "  19,\n",
       "  20,\n",
       "  21,\n",
       "  22,\n",
       "  23,\n",
       "  24,\n",
       "  25,\n",
       "  26,\n",
       "  27,\n",
       "  28,\n",
       "  29,\n",
       "  30,\n",
       "  31,\n",
       "  32,\n",
       "  33],\n",
       " ['bad', 'good'])"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# load dataset https://archive.ics.uci.edu/ml/datasets/ionosphere\n",
    "data = pd.read_csv(\"data/ionosphere.csv\", header=None)\n",
    "col_names = list(data.columns)\n",
    "feature_names = list(col_names[:-1])\n",
    "target_names = ['bad', 'good']\n",
    "\n",
    "cols = list(data)\n",
    "cols.insert(0, cols.pop(cols.index(34)))\n",
    "data = data.loc[:, cols]\n",
    "data.columns = col_names\n",
    "\n",
    "feature_names, target_names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Dataset statistics:\n",
      "===================\n",
      "number of features:       34\n",
      "number of classes:        2\n",
      "data type:                float32\n",
      "number of train samples:  280 (pos=184, neg=96, size=0MB)\n",
      "number of test samples:   71 (pos=41, neg=30, size=0MB)\n",
      "labels                    ['b', 'g']\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/anaconda3/lib/python3.6/site-packages/sklearn/preprocessing/label.py:151: DeprecationWarning: The truth value of an empty array is ambiguous. Returning False, but in future this will result in an error. Use `array.size > 0` to check that an array is not empty.\n",
      "  if diff:\n"
     ]
    }
   ],
   "source": [
    "X = np.float32(data.drop(0, axis=1).values)\n",
    "y = data[0].values   # conversion to int64 for pytorch\n",
    "\n",
    "# Encode target class\n",
    "encoder = LabelEncoder()\n",
    "y = np.int64(encoder.fit_transform(y))\n",
    "target_names = list(set(encoder.inverse_transform(y)))\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, shuffle=True, \n",
    "                                                    random_state=HP.SEED)\n",
    "\n",
    "dataset_statistics(X_train, y_train, X_test, y_test, target_names)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix\n",
      " [[29  1]\n",
      " [ 1 40]]\n",
      "Accuracy  0.971830985915493\n",
      "Recall    0.975609756097561\n",
      "Precision 0.975609756097561\n",
      "F1        0.975609756097561\n",
      "ROC AUC   0.9711382113821138\n"
     ]
    }
   ],
   "source": [
    "params = {\n",
    "    'module'                       : MLP,\n",
    "    'module_H'                     : 20,\n",
    "    'module_D_in'                  : X_train.shape[1] ,\n",
    "    'module_D_out'                 : 2 ,\n",
    "\n",
    "    'module_initialize'            : torch_weight_init ,\n",
    "\n",
    "    'max_epochs'                   : 1000,\n",
    "    'batch_size'                   : 100 ,\n",
    "    'batch_shuffle'                : True,\n",
    "\n",
    "    'optimizer'                    : torch.optim.SGD ,\n",
    "    'optimizer_lr'                 : 0.5 ,\n",
    "    'optimizer_momentum'           : 0.9 ,\n",
    "    'optimizer_weight_decay'       : 5e-4 ,\n",
    "    'optimizer_nesterov'           : True ,\n",
    "\n",
    "    'criterion'                    : nn.CrossEntropyLoss ,\n",
    "    'criterion_size_average'       : True ,\n",
    "\n",
    "    'scheduler'                    : torch.optim.lr_scheduler.LambdaLR ,\n",
    "    'scheduler_lr_lambda'          : exp_decay ,\n",
    "\n",
    "    'stopping'                     : Stopping ,\n",
    "    'stopping_patience'            : 140 ,\n",
    "\n",
    "    'scoring'                      : metrics.f1_score ,\n",
    "\n",
    "    'random_state'                 : 10\n",
    "}\n",
    "\n",
    "pipeline = Pipeline(steps=[\n",
    "                ('scaler',     MinMaxScaler(feature_range=(0,1), copy=True)),\n",
    "                ('outlier',    OutlierSampler(3.5)),\n",
    "                ('sampler',    SMOTE(random_state=HP.SEED)),\n",
    "                ('classifier', Classifier(**params))\n",
    "                ])\n",
    "\n",
    "y_pred = pipeline.fit(X_train, y_train).predict(X_test)\n",
    "y_prob = pipeline.predict_proba(X_test)\n",
    "\n",
    "print('Confusion Matrix\\n', metrics.confusion_matrix(y_test, y_pred))  # Confusion Matrix\n",
    "print('Accuracy ', metrics.accuracy_score(y_test, y_pred))            # Accuracy (TP + TN) / float(TP + TN + FP + FN)\n",
    "print('Recall   ', metrics.recall_score(y_test, y_pred))                # Recall (Sensitivity)   TP / float(FN + TP)\n",
    "print('Precision', metrics.precision_score(y_test, y_pred))     \n",
    "print('F1       ', metrics.f1_score(y_test, y_pred))     \n",
    "print('ROC AUC  ', metrics.roc_auc_score(y_test, y_pred))     "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test accuracy score: 0.9142857142857143 0.09476070829586856\n"
     ]
    }
   ],
   "source": [
    "#  Cross-validate score on test set\n",
    "cv = KFold(n_splits=10, shuffle=True, random_state=HP.SEED).split(X_test)\n",
    "scores = cross_val_score(pipeline, X_test, y_test, cv=cv, scoring='accuracy', \n",
    "                         verbose=0, n_jobs=-1)\n",
    "\n",
    "print(\"Test accuracy score:\", np.mean(scores), np.std(scores))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x107c7b0b8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "clf = pipeline.named_steps['classifier']\n",
    "plot_loss_curve(clf.loss_train_, clf.score_train_,  clf.loss_valid_, clf.score_valid_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "try:\n",
    "    cv = KFold(n_splits=10, shuffle=True, random_state=HP.SEED).split(X_train)\n",
    "    plot_learning_curve(pipeline, 'Learning Curves', X_train, y_train, cv=cv, shuffle=True, \n",
    "                        n_jobs=-1, random_state=random_state)\n",
    "except:\n",
    "    pass"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1aa81db128>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# param_name ='classifier__optimizer_lr'\n",
    "# param_range = [  0.01, 0.1, 0.2, 0.3, 0.5, 0.8, 1.0, 1.1, 2.0]\n",
    "\n",
    "# param_name ='classifier__module_H'\n",
    "# param_range = [  2, 3, 5, 7]\n",
    "\n",
    "param_name ='classifier__stopping_patience'\n",
    "param_range = [ 120, 150, 240]\n",
    "\n",
    "# param_name ='classifier__batch_size'\n",
    "# param_range = [ 70, 80, 90]\n",
    "\n",
    "cv = KFold(n_splits=3, shuffle=True, random_state=HP.SEED).split(X)\n",
    "train_scores, test_scores = validation_curve(\n",
    "                 estimator=pipeline,\n",
    "                 X=X,\n",
    "                 y=y,\n",
    "                 param_name=param_name,\n",
    "                 param_range=param_range,\n",
    "                 scoring = 'f1',\n",
    "                 cv=cv, n_jobs=-1)\n",
    "\n",
    "train_mean = np.mean(train_scores, axis=1)\n",
    "train_std = np.std(train_scores, axis=1)\n",
    "test_mean = np.mean(test_scores, axis=1)\n",
    "test_std = np.std(test_scores, axis=1)\n",
    "plt.plot(param_range, train_mean,\n",
    "          color='blue', marker='o',\n",
    "          markersize=5,\n",
    "          label='training accuracy')\n",
    "plt.fill_between(param_range, train_mean + train_std,\n",
    "                  train_mean - train_std, alpha=0.15,\n",
    "                  color='blue')\n",
    "plt.plot(param_range, test_mean,\n",
    "          color='green', linestyle='--',\n",
    "          marker='s', markersize=5,\n",
    "          label='validation accuracy')\n",
    "plt.fill_between(param_range,\n",
    "                  test_mean + test_std,\n",
    "                  test_mean - test_std,\n",
    "                  alpha=0.15, color='green')\n",
    "plt.grid()\n",
    "# plt.xscale('log')\n",
    "plt.legend(loc='lower right')\n",
    "plt.xlabel('Parameter')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.ylim([0.8, 1.0])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix\n",
      " [[29  1]\n",
      " [ 1 40]]\n",
      "Accuracy  0.971830985915493\n",
      "Recall    0.975609756097561\n",
      "Precision 0.975609756097561\n",
      "F1        0.975609756097561\n",
      "ROC AUC   0.9711382113821138\n"
     ]
    }
   ],
   "source": [
    "# http://www.ritchieng.com/machine-learning-evaluate-classification-model/\n",
    "# https://github.com/justmarkham/DAT8/blob/master/other/model_evaluation_comparison.md\\\n",
    "# Accuracy of a model that simply counts, i.e. looks at the prior distribution\n",
    "\n",
    "y_pred = pipeline.predict(X_test)\n",
    "try:\n",
    "    y_prob = pipeline.predict_proba(X_test)\n",
    "except:\n",
    "    pass\n",
    "\n",
    "print('Confusion Matrix\\n', metrics.confusion_matrix(y_test, y_pred))  # Confusion Matrix\n",
    "print('Accuracy ', metrics.accuracy_score(y_test, y_pred))            # Accuracy (TP + TN) / float(TP + TN + FP + FN)\n",
    "print('Recall   ', metrics.recall_score(y_test, y_pred))                # Recall (Sensitivity)   TP / float(FN + TP)\n",
    "print('Precision', metrics.precision_score(y_test, y_pred))     \n",
    "print('F1       ', metrics.f1_score(y_test, y_pred))     \n",
    "print('ROC AUC  ', metrics.roc_auc_score(y_test, y_pred))     "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test accuracy score: 0.9126667665189905 0.056394631946857274\n"
     ]
    }
   ],
   "source": [
    "#  Cross-validate score \n",
    "cv = KFold(n_splits=10, shuffle=True, random_state=HP.SEED).split(X_train)\n",
    "scores = cross_val_score(pipeline, X_train, y_train, scoring=\"f1\", cv=cv,  verbose=0, n_jobs=-1)\n",
    "\n",
    "print(\"Test accuracy score:\", np.mean(scores), np.std(scores))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1acf14a0b8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_roc_curve(y_test, y_pred, y_prob)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1acf1c2828>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_classifier_summary(y_test, y_pred, target_names)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1acf181390>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_scatter_plots(X_test, y_pred, y_prob, y_test, target_names, figsize=(18,6))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Note: weighted average f1-score \n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "          b       0.95      0.97      0.96       126\n",
      "          g       0.98      0.97      0.98       225\n",
      "\n",
      "avg / total       0.97      0.97      0.97       351\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'Data points=351'"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "'Features=34'"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "'Class dist.=0.641026'"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "'F1 valid=0.977679'"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "'ACC=0.971510'"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "'ROC_AUC=0.970794'"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "'LOG_LOSS=0.984020'"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "'Misclassified=10'"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "'Data points=[34, 63, 133, 137, 143, 144, 234, 236, 271, 285]'"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "classifier_summary_report(X, y, pipeline.predict(X), target_names)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Summary for model: MLP\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Shape                     Param #   \n",
      "=================================================================\n",
      "fc1.0 (Linear)               ((20, 34),)               680       \n",
      "_________________________________________________________________\n",
      "fc1.1 (BatchNorm1d)          ((20,), (20,))            40        \n",
      "_________________________________________________________________\n",
      "fc1.2 (ReLU)                 ()                        0         \n",
      "_________________________________________________________________\n",
      "fc2.0 (Linear)               ((20, 34),)               680       \n",
      "_________________________________________________________________\n",
      "fc2.1 (BatchNorm1d)          ((20,), (20,))            40        \n",
      "_________________________________________________________________\n",
      "fc2.2 (ReLU)                 ()                        0         \n",
      "_________________________________________________________________\n",
      "classifier.0 (Dropout)       ()                        0         \n",
      "_________________________________________________________________\n",
      "classifier.1 (Linear)        ((2, 20), (2,))           42        \n",
      "=================================================================\n",
      "Total params:         1,482\n",
      "Trainable params:     1,482\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "print_model_summary(pipeline.named_steps['classifier'].module_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1acf469e80>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1acf53e518>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1acf3baef0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1acf469cf8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ad556d5f8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ad561d860>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ad54fd898>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ad5739f60>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ad5725048>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlYAAAERCAYAAAC5EOlWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAIABJREFUeJzt3XmcHFW5//FPyETWAYOOCm5cLvAIF2RVghBBfyiCRBRBJbIIKKCiKCgCAm4ogoC4IbJENhGEECGybyoBEVkUEP1iUFREvQM3hEBkScjvj3OaVJpeamYqM90z3/frlVe6q6qrntPLmadOnTpn3KJFizAzMzOzoVtmpAMwMzMzGy2cWJmZmZlVxImVmZmZWUWcWJmZmZlVxImVmZmZWUWcWJmZmZlVpGekA7D2ImIN4C/ALyVtXbfuLGAvoE/SIwPY58+AiyWd1WKbbYDvSlp/4FE/v4/PA/sD10nap8k2fcAPgLVI38nLgc9Jeq7J9q8GbgU2HEiZBxDzl4HZks5psc2HgF0k7dhg3TuBzSUdXXVsZqOJ67YXbN+2bouIRTR4TyLiXcC2kj45yCJZRdxi1T2eAiIiXltbEBErAluOXEil7AtMbVbxZN8E7pP0emATYHPgQ402jIg9gV8Cq1cc5/MkHd0qqSrhDcCqVcVjNsq5bmPodZuky5xUdQa3WHWPhcCFwAeBr+VlOwOXAofUNoqI/YBP5u3/DRwo6f6IWB04m/Sj/SvwssJrljgDqj0vHjwiXgQcB2wNjAfuAj4p6fGIOADYTNKH615zIfAq4MyIOBr4BXAq8DrgOeBUSd8GZgA3A0h6KiLuBV5LnVyGdwPbAWr2RkXEycA8SUdFxGrAw8BbJd0YEbsDUyS9PyL2BT5GOsF4NL9Xf8xnyvdKOiEidsjlXgj8FtgW2CofarWIuBx4DbAAmAqsDBwAjI+IuZI+3yxOMwNct5Wu2wq+GhFvINVdR0r6WbEVPSImAccDywKrAddK2jcieoDvkJLWZ4E/A3tLeqLEMa0kt1h1l3OAPQrP9wLOqj2JiLcChwJvkbQhcD7w04gYB3wPuFXS/5Aqp9cN8NiHkZKHTfO+Hwa+DiDp1PqKJy9/f97ug5IuBE4B7pf0OmALYL+IWEvSdEn/ymXYmJSgzGiwv4cl7Szp/jaxXgJsnx+/A/gX8Lb8/F3A9IjYmvT+TZa0MakSWuKYEfES4Fxgd0kbATcCryxssiZwkKQNSGean5H0a1IFe6GTKrPSXLeVq9tq/ixpE2B34Ox8ybHoIOBoSZsD6wHviohNc2zbkC41bkpKrF5f8phWkhOrLiLpDmBhRGyar8X3Srq3sMk7SH/Q+/P2Z5ESgTVILS1n5eWzgRsGePgdgZ2AuyLit6Szq/UGuI9tgdNyDHMlrZ9jASAitgOuAT4h6bcD3HfRLOBVEfFy0ntyDPC2fGa6NXAF8E5Sv4dbcnmOByZGRPES3ptJzfi/yzGfDTxeWH9bIf7fUjhTNrPyXLcN2Kn5WPcC95ESpqK9gBdHxBGkpG95YCXgHlKL368j4ivAdEm3VBCPFfhSYPc5l3SW0p8fF40HnqlbNg6YACzKj2sWNNiu1izeyHhS68yVebuVgOUGGPuCHAd5H2sCj+Qm94NJZ467SbpuIDvNnTa/nJ8+LGmH3IF1B1Kfhj2Aw4FdgVskPRER44FzJX0u72MZ0qWEOXXxFt8zSM38Nc8WHte/v2Y2MK7b6jSq2/LjhYXNlmHJughSC/rdwFXAT0j14DhJj0XEhqRLgW8FLoyIb0g6ZSBxWWtuseo+55EShPeTmsOLrgI+UGsWjoi9SX2HZud1++XlrwHeUnhdP7BZfjy1yXGvBg6MiBflJOR04NgBxn4dsHeOYRXgemDtiPg48HFg0kArHni+0+ZG+V+t4rmEdOngHknPkM5ijwWmF8qzW+6DBalf1PV1u74ZWCciXp9jfi/wYgoVaBMLSBW+mZXnuq1Ok7oNcgf4iNiE1PL+69qKiHgx6Qaaz0m6hNQXbC1Sv88dc2y3SPoi6RLsGwYal7XmxKrLSPoH8AfgT5L+r27dtaS7UG6IiN+TmoN3zLf2fhxYLyL+AJxJunRV80ngexFxJ7Au8M8Gh/4K8CCpY+d9pLPAQwAi4oCIOKNE+AcC60bE3aSk5VhS0/TXSWeIl0TEb/O/z+d9fzkPfzBQ15FaoK7Nz68GXg7MBJB0DanD6rU5nqnAzpKeT5ry+7sbcE5+b7YjJU3z2xz7BmC7iPjOIOI2G5Nctw3ImhFxF3AG8IHi+yXpsXz8O3Nn+cNyTGsBVwK/B+6NiNuBNwFfGmQM1sS4RYvanXybjU0RsTJwJPBFSfPz2eHlwOrFBMzMzKzGiZVZCxFxDPAeUh+GZ4GDJd00slGZmVmncmJlZmZmVhH3sTIzMzOriIdbGIKI+DlpXKT/lvTnwvJtSINJflbSCcMQRw9pvKV1IuLNwEck7dFgu8NJnT57SHfgfKlRX6F899sRpFF7/wrsKenRPETBUaRBNlckjQd1sKRFeRTgk/Py8cBxks6r2++7SUMc9DYpx+uAE4FX50VzgM9LmhURhwEfyMvXIt3tMzc/fy+p0+prC8smkPpDfUXSvCbH2w14vaTDI2Ia8CNJ19dtsxbwfdJozS8CzpR0YqP95e1fBNwEXFT77CNiCmlk6L8VNp3cLK4W+275+UXE24Hj82CmZg11W73V4HVl67GG27X6TUfE+4AvkG5SeQj4mKS/5nV3kMaDqg378CNJ32hw3HeS+maukI/9e1I9+VBEXEyqvwA2BO4lDZ0wR9JbIo0MX1s2LsdxZqvhECLiB8Alkq7OHfg3kfSfum22ItXPPaQphA6UdHuDfb2GNO7VK/O2n8n7Lda/5PeuV9LKzeJqEuvapLr6pcATpL8tf8zrDgH2yWXuB/aX9MBA9t8p3GI1dH8jjb1StCdpyoXhshlwR378FuDn9RtEmprlfcCmwPp5u10bbLcZ8F3gvUoTlN4PfDWvPog0au+WpNF6twDen0c/ng58If9R3x44Kf+IavtdGziB1mM9TQfOkPR6pbm1jgIuj4hVJX29dtsxcDup8q/dhlz78X22sM1GQC8vvG27aOvCe7Ul0GigvLNIAxNulMu7f6RRoJs5mTQie9GbgBMK8W40iKSq6ecXEcvnvmAX4pMlK6cr6q16A6jHWm13Fg1+07mO+gHpLrsNSXcNX5z3tyLw36QRy2u/4UZJVW16nQ9I2lBpNPjfkcaSQtIuhToK0kjyG0kqDhFRW7Yh6U7kPXPS0cyWwM2R5lp8pD6pys4DDs3HPS7H2MhM4HKl2Sj2BC6IiGXr6t9tgCdJw2IM1I9I0/2sR0pgL46IcRGxLWnuxS1yuS8BfjiI/XcEV8JDdx6pgvoyQESsQJpL7vkxSyLilaRk5TWklpQLJH0trzuCNOrv8qTWns9ImhERXySNKrwaqSXmH6SpVf5Z2O/qpFaj1YH5kUYNDuBPEXGv0vQqNe8Bzpf0ZH7tD3PcP6krz+6kM6QH8/MvAi/Jj/fM8f0n7+O9pLO3ZUlng9cB5DOzftL4KX/K78l5wMG0TnRWy+8BeT+/zGeQC5u/pDFJz+aB+f4VEa+rnRXluA8gjVu1HjApIr5NGorhp6SKrOhMUsKCpLkRMZsGc33l/e4BrEJqKSt6E/BsRLyfNHL75yX9Mr9mCuns9kWkYRw+I+lXDXbf6vPbjvS+7cXiudbMWumWeqte2Xqs1XbNftOrAr+TdE9e98uIWCMi1gD+i9TCclVEvCy/T0c0SGJeSvotr1RYdjIpuRowSY/kemx6RJxU10J9BukzW50020QfsCgivqoXTqc1HpiYH/eSWq2WEBEbAatK+n4+9l25peu5uk1PAK7U4gFVVwG+BWxA+p5cTzrJXWKg1vx9eh1wQd7/lRHxfWBj0rRjH5VUm9niduBzZd6jTuQWq6G7C3gqIjbPz3cGLmPJ0X/PBaYpzc30RmDbiHhfPsPYFtgmt9B8nsWj7AJMBnZVmn/qSVIy8Dyl+aVqc9jtlh//k3RWVV85vRr4e+H5Q6TEp946QE9EXBoRvyPNwzWvsG69iLg+j9fyUeD/JD0l6czaDiJNltoL3JoX/SD/u7vB8Yo+DnwnIh6OiJ9ExIHAbyTNbfO6hnKldz/pB19cfirpLO+B/J59FjhbUn1ShaQfSpqfy/UOUpJ0Vf12EbEBqUVvvwahPEqagmIj0gjwMyLiVfkM+WvADvkMcT/SeDcrNthH089P0k8lfZolp9sxa6Vb6q16ZeuxVr+XZr/pu4D1c4JRO+l5CSlJ7M3x7koaUPM1NBhEVNLdpAFG74qI+yLidGAKDeqMAfgd8ApS0lY81oeBTwM/zu/hhaQpcxrNUboPaTy+h0iX+g5ssM06wIMRcVJE/DoibgZWk/T8qO4RsR5pyp+jC6/7JnBH/p5snOM8uMH+X00aPb6YqD0EvErSvZJ+kY+xLGn8r4savx2dz4lVNWpTMcALJw9dkXTJ6Sv5zOxW0o9yo3ztfk/ggxHxdVIFVDzT+Xkhg7+LdEbVyAbA3bnC+5saj7G0DEuOGD6Oxi1BE0gVwf4sPpM4vbBuEmmqmC1JZ0ufKL44X4v/EjBF0n8i4mPAAknTmsT+PEk/JlViewJ/JFUG9+UzxsFaROMBPTdg8VnkxqT3t6mI2JN0lr9L8ew7r1uFNILxnrUz5CKlyVUvlrRI0izSJce35X+rAdfn78aPSGeHa9Xvg/Kfn1lZ3VBv1Sv7O2i7Xf1vOncp2Ac4NdLgm5uS6ohnlEZA30PS/0l6inRC9J5GAUo6hPS7Pgr4D/AN4BeR+qgORq0cjS7xbcjiuqthPRZpztTTga0lvYr0mV/c4ARuAqle/4XS5M2fJk15s3phm08B36072d2RdDn1t6RLu2+k7mQ2q/9MoO5ziTSy/jWk1sEjGuyjK/hSYDV+BNwREScBK0u6NyJq68aTvjxvKpwlvZR0trgJcCkp478G+AWpU2VN8Yf0grno8hf+alLT+82ky1Ar5i/4/nVnf38jNRnXrE46W6j3MHC3Fs/I/kMWT2r6MOns6Gng6Yi4iDRR8cn5LOMs0uW1LQqXEj8ErJBjehGwfH68g6SHC2V5HfAhSYeRmtmvA46OiOuAXUjNzwOSL2+sS+o8Wlx+AOny2zI5lv8C/hkRu9W3WkXqP3ZCjmFbNZ5AdTtSM/v5+XN/DWnS55WBk4CPAccW/nCMI42JtRJwvaTn+ypEmoD24YioXSqBdHZY9vMzK6vj660h/A6abtfsN53rsNmSJhWefwr4S269mlu7hM/i3/ASIs3t9xJJPyT1GZ2eL5s+REp8XtBhvIQ3AH+R9ETdsc7IZfjfiPgo6XLqTyPi8rpWq8nAX5U7q0v6aUScTKobi/E8DDwm6dK83W0R8WdS8vZwTgzfS0o4i8aTWij/kON6MemS5BLzHJJa5FeLiHGFurD4ubye1Go6g3RpuWtPHN1iVYGcINwNTKNu8tB85nYruWk0f+luJvVPeDNwu6STSJXTu0lf0oEc97PAebkp+FLgAKVOhvVN6peSzjBXzBXGh0h9iupdDOwYEbV+VTsDvyms2z0ilomICaQzldq684CVSRXxg4UY36g00/tGpJau/+T4nk+qsn8D+0XELrUFEbEq6e6UO0u+Jc+LiOVJfRuuLMaTYzqV1CdhF1Jl9yiwXqNLgcDxpM9psyZJFZJ+ImkNLe7ceRnwTUlHky6jfpz0PhIRG5PO6K4i9UV4e04qax1u7waWl7SDFneSvYzyn59ZKd1Qbw3hd9Bqu2a/6WVJncBrdyV/CpilNF3Mq4ATIt0oMj6/Lxc2OO484Nh8yaxmTdIl1gHf4ZaT0ONocGKZLwU+CvwPqcXxmvw+1V8KvJt0iXOdvM/NSXcs3l+33S2kxHnHvN3rSB32a104NiDdvfhg3euuBj4dqRP6sqT670DVzXMo6SHS3I7vz/vfjtRCf09EvIp0Av9lSZ/u5qQKnFhV6RzStfpGnbOnkjpJ30OaLPPHkn4E/Bh4aaRbZO8jNX+uGhENhyNoYjKpcoN0ae6XjTaSNJN0p8VtpNt578gxLzEfVt7uZFLT9X25TLV+Q0eSEqB7Sa1Afya1Vm1BSlLWIlVMtTmxGiUqDUmaQ5ptfd+IeDDSfGDXAV+TdEPrVz/vG/m4d5Le5ydIFU4jm5KSwvWB39dd9wcg/9gPJvUZuLZQrtpkq2fk1q9W5VpI+mP0mUjzdv0QeL+kRyTdR3pvL4jUn+0rwLvqz0zzfpp+fmZD0NH1Vr0B1mMv2K7Vbzonkx8Brsxl25w82TGpj+gvSCd5f8xlfsE8f5JuJPVfOjsi/pT3czKphX5OyffmxhzTHaQk5Rw1GG4hUh/NP+c+UMX3sz6m+0mXa6dH6hv7LdK8qI9HxOr5WKvnKxHbsbiuuhjYR2kOR4C1SXMq1vsk6QaGe0hJ2D2k5LWR3YAD8v6/Smrpeo502XRF4JOFz6Rdf7uO5ZHXzczMzCriFiszMzOzijixMjMzM6uIEyszMzOzijixMjMzM6tIx4xj1d8/b8i96CdOXIE5cxqNBdndRmu5YPSWbbSWC6orW19fb6t5I7tKFfVXFUbz964Rl3d06/TyNqvDRlWLVU/PYAe27WyjtVwwess2WssFo7ts3W6sfTYu7+jWreUdVYmVmZmZ2UhyYmVmZmZWESdWZmZmZhVxYmVmZmZWESdWZmZmZhVxYmVmZmZWkbbjWEXEBGAasAawLHCMpMsK66cARwMLgGmSTo+I5YHzgJcB84C9JPVXH76Z2cBFxHjgdCCAhcDekh4orH9BvTYigZpZ1ynTYrU78KikycD2wHdrK3LS9U3g7cDWwH4R8Qrgo8A9+TXnAEdWHbiZ2RBMAZC0JSmBOqm2okW9ZmbWVpnE6iLgqMLzBYXH6wKzJc2R9AwwC5gMbAVclbe5Eti2gljNzCoh6afAfvnpa4F/F1Y3q9fMzNpqeylQ0hMAEdELXMySrU8rA3MLz+cBq9Qtry1raeLEFSoZZbWvr3fI++hE3VCuKYdcusTzmSfuVOp13Vg2aF++bijXYI2GsklaEBFnA+8BdimsalavNVVV/VWF0fDZDITLO7IGUzcORKeVt4xScwVGxKuBGcApks4vrHocKJa6F3isbnltWUsVzT1Gf/+8Ie+n03RrucrE3K1lg9bl6+ZytVNV2TqhwpS0V0R8Dvh1RKwn6Uma12tNdcp8ZqP5e9eIy9uZqoqx08vbrA4r03n95cA1wIGSrq9b/Qdg7YhYFXgCeDNwAqlpfQfgNlK/rJsGHbmZWcUiYg/gVZKOBeYDz5E6sUPzes3MrK0yLVZHABOBoyKi1tfqdGBFSadFxMHA1aT+WtMk/SMivg+cHRGzgGeAqUshdjOzwboE+GFE/BKYAHwK2DkiVmpWr41grGbWRcr0sToIOKjF+pnAzLpl84FdhxydmdlSkC/5va/F+hfUa2ZmZXiAUDMzM7OKOLEyMzMzq4gTKzMzM7OKOLEyMzMzq4gTKzMzM7OKOLEyMzMzq4gTKzMzM7OKOLEyMzMzq4gTKzMzM7OKOLEyMzMzq4gTKzMzM7OKOLEyMzMzq4gTKzMzM7OKOLEyMzMzq4gTKzMzM7OKOLEyMzMzq4gTKzMzM7OKOLEyMzMzq0hPmY0iYnPgOEnbFJa9ArigsNlGwGHAD4CHgD/l5b+SdHgl0ZqZmZl1sLaJVUQcCuwBPFlcLulfwDZ5my2ArwKnA/8N3ClpStXBmpmZmXWyMpcCHwB2brYyIsYB3wE+KmkhsCnwyoi4MSKuiIioJlQzMzOzzta2xUrS9IhYo8UmU4DfS1J+/k/gWEkXRcRWwHnAG9odZ+LEFejpGV8i5Nb6+nqHvI9O1I3lKhtzN5YN2sfdreUqYzSXzcxsKEr1sWpjd+Bbhee3AwsAJM2KiFdGxDhJi1rtZM6c+UMOpK+vl/7+eUPeT6fp1nKViblbywaty9fN5WqnqrI5OTOz0aiKuwI3BW4pPP8C8CmAiNgQ+Fu7pMrMzMxsNBhwi1VETAVWknRaRPQB8+oSp68D50XEO0ktVx+qJFIzMzOzDlcqsZL0IDApPz6/sLyfNMxCcds5wDurC9HMzMysO1TRx8rMrGtExARgGrAGsCxwjKTLCusPBvYF+vOi/Qs355iZteTEyszGmt2BRyXtEREvAe4CLius3wTYU9IdIxKdmXU1J1ZmNtZcBFxceL6gbv2mwOF5donLJR07bJGZWddzYmVmY4qkJwAiopeUYB1Zt8kFwPeAx4EZEbGjpJ+12mdV4/BVYawNY+Hydp4qY+yG8tZzYmVmY05EvBqYAZxSvCEnzyRxsqS5+fnlwMZAy8SqinH4qjCax09rxOXtTFXF2OnlbZb0ObEyszElIl4OXAMcKOn6utUrA/dGxLqk+VHfSurobmZWihMrMxtrjgAmAkdFxFF52enAinl8viOAG4GngeslXTFCcZpZF3JiZWZjiqSDgINarD8XOHf4IjKz0aSKKW3MzMzMDCdWZmZmZpVxYmVmZmZWESdWZmZmZhVxYmVmZmZWESdWZmZmZhVxYmVmZmZWESdWZmZmZhVxYmVmZmZWkVIjr0fE5sBxkrapW34wsC/QnxftD/wNOA94GTAP2EtSP2ZmZmajXNsWq4g4FDgDWK7B6k2APSVtk/8J+Chwj6TJwDnAkVUGbGZmZtapylwKfADYucm6TYHDI2JWRByel20FXJUfXwlsO7QQzczMzLpD20uBkqZHxBpNVl8AfA94HJgRETsCKwNz8/p5wCplApk4cQV6esaX2bSlvr7eIe+jE3VjucrG3I1lg/Zxd2u5yhjNZTMzG4pSfawaiYhxwMmS5ubnlwMbk5KsWq3bCzxWZn9z5swfbCjP6+vrpb9/3pD302m6tVxlYu7WskHr8nVzudqpqmxOzsxsNBp0YkVqmbo3ItYFngTeCkwD5gM7ALcB2wM3DTVIMzMzs24w4MQqIqYCK0k6LSKOAG4Engaul3RFRPwcODsiZgHPAFOrDNjMzMysU5VKrCQ9CEzKj88vLD8XOLdu2/nArtWFaGZmZtYdPEComZmZWUWcWJmZmZlVxImVmZmZWUWcWJmZmZlVxImVmZmZWUWcWJmZmZlVxImVmZmZWUWcWJmZmZlVxImVmZmZWUWcWJmZmZlVZCiTMJuZdZ2ImECaMH4NYFngGEmXFdZPAY4GFgDTJJ0+EnGaWXdyi5WZjTW7A49KmgxsD3y3tiInXd8E3g5sDewXEa8YkSjNrCs5sTKzseYi4KjC8wWFx+sCsyXNkfQMMAuYPJzBmVl386VAMxtTJD0BEBG9wMXAkYXVKwNzC8/nAau02+fEiSvQ0zO+yjAHra+vd9iONeWQS1+wbOaJOw3b8aF7y1u/r0b76YT3dzCq/EyG8/OtihMrMxtzIuLVwAzgFEnnF1Y9DhRr8l7gsXb7mzNnfrUBDlJfXy/9/fNGNIbhPP5oKm/Z/Yx0ecuoKsZO+HxbaZb0ObEyszElIl4OXAMcKOn6utV/ANaOiFWBJ4A3AycMc4hm1sWcWJnZWHMEMBE4KiJqfa1OB1aUdFpEHAxcTeqDOk3SP0YoTjPrQk6szGxMkXQQcFCL9TOBmcMXkZmNJqUSq4jYHDhO0jZ1y3cDPgUsBO4GPibpuYi4i8UdQP8iae/qQjYzMzPrTG0Tq4g4FNgDeLJu+fLAMcAGkuZHxI+BHSPiGoD6JMzMzMxstCszjtUDwM4Nlj8NvElS7XaYHuApYENghYi4JiJuiIhJ1YRqZmZm1tnatlhJmh4RazRY/hzwb4CI+ASwEnAtsD7pLpozgLWBKyMiJC2o30dRVePAdOOYF2V0Y7nKxtyNZYP2cXdrucoYzWUzMxuKIXVej4hlgOOBdYD3SloUEfeTRi5eBNwfEY8CqwF/b7WvKsaB6fQxLwarW8tVJuZuLRu0Ll83l6udqsrm5MzMRqOhTmnzA2A54N2FS4L7ACcCRMTqpJGM/znE45iZmZl1vAG3WEXEVNJlv9uBfYGbgBsiAuBbwJnAWRExC1gE7NPuMqCZmZnZaFAqsZL0IDApPy5O/9CsxWvq0MIyMzMz6z5DvRRoZmZmZpkTKzMzM7OKOLEyMzMzq4gTKzMzM7OKOLEyMzMzq4gTKzMzM7OKOLEyMzMzq4gTKzMzM7OKOLEyMzMzq4gTKzMzM7OKOLEyMzMzq4gTKzMzM7OKOLEyMzMzq4gTKzMzM7OKOLEyMzMzq4gTKzMzM7OKOLEyMzMzq4gTKzMzM7OK9JTZKCI2B46TtE3d8inA0cACYJqk0yNieeA84GXAPGAvSf2VRm1mNkQt6rWDgX2BWr21vyQNc3hm1qXaJlYRcSiwB/Bk3fIJwDeBN+R1N0fETGAqcI+kL0bEB4AjgYOqDtzMbLCa1WvZJsCeku4Y3qjMbDQocynwAWDnBsvXBWZLmiPpGWAWMBnYCrgqb3MlsG0VgZqZVahZvQawKXB4RMyKiMOHMSYzGwXatlhJmh4RazRYtTIwt/B8HrBK3fLasrYmTlyBnp7xZTZlyiGXvmDZzBN3AqCvr7ftNgNV5b4Gq1aumvqYGsWzNONutO969TEPdbt2x6/y8y2j3WfSKKbBxl02xsHsazDfnb6+3o74XQxWi3oN4ALge8DjwIyI2FHSz1rtbyD119I2mN9TNx9/tJR3adaXw63KGLuhvPVK9bFq4nGgWOJe4LG65bVlbc2ZM38IoUB//zz6+nrp75/XcpuqVLmvdtqVC8rHM5xxlzlWmbJVebwqlTleVduUNZh9DfS70+ozG8jxO7HCjIhxwMmS5ubnlwMbAy0Tq6HWX1Wp8vc0WJ1WNy5tw11/jXR5y6gqxk74fFtpVocNJbH6A7B2RKwKPAG8GTgBeC2wA3AbsD1w0xCOYWY2nFYG7o2IdUn9r94KTBvZkMysmww4sYqIqcBKkk7Ld89cTeqrNU3SPyLi+8DZETELeIbUmd3MrGPV1WtHADcCTwPXS7piZKMzs25SKrGS9CAwKT8+v7B8JjCzbtv5wK7VhWhmVr0W9dq5wLkjFJaZdTkPEGpmZmZWESdWZmZmZhWawNQVAAAOpElEQVRxYmVmZmZWESdWZmZmZhVxYmVmZmZWESdWZmZmZhVxYmVmZmZWESdWZmZmZhVxYmVmZmZWESdWZmZmZhVxYmVmZmZWESdWZmZmZhVxYmVmZmZWESdWZmZmZhVxYmVmZmZWESdWZmZmZhVxYmVmZmZWESdWZmZmZhXpabdBRCwDnAJsCDwNfFjS7LxuI+DkwuaTgHcDtwH3A/fm5TMkfavCuM3MzMw6TtvEipQoLSdpi4iYBJwI7AQg6bfANgARsSvwsKSrImJb4MeSPrF0wjYzMzPrPGUSq62AqwAk3RoRm9VvEBErAl8C3pwXbQpsEhG/AP4X+KSkf1YTspmZmVlnKpNYrQzMLTxfGBE9khYUlu0LXCTpkfz8j8Adkq6LiA8C3wF2aXWQiRNXoKdn/ABCX1JfX+8S/7fapgpV7quK45WNZzjjHu6YOu0zqXKbsgazr8F8Ts1eM9yfgZlZpymTWD0OFGvLZeqSKoAPsmTidAMwPz+eAXy53UHmzJnfbpOW+vvn0dfXS3//vJbbVKXKfbXTrlxQPp7hjLvMscqUrcrjVanM8arapqzB7Gug351Wn9lAju8kzMxGozJ3Bd4M7ACQ+1jdU1wZEasAy0r6e2HxGcB78+P/B9wx9FDNzMzMOluZFqsZwNsi4hZgHLB3RBwMzJZ0GbAO8GDdaw4DpkXEx4AngQ9XF7KZmZlZZ2qbWEl6DjigbvEfC+t/Q7pzsPiavwBvqSJAMzMzs27hAULNzMzMKuLEyszGpIjYPCJ+3mD5lIj4TUT8KiI+MgKhmVkXc2JlZmNORBxKuslmubrlE4BvAm8Htgb2i4hXDH+EZtatnFiZ2Vj0ALBzg+Xrkm7MmSPpGWAWMHlYIzOzrlbmrkAzs1FF0vSIWKPBqvoBkecBq7Tb30AHOJ5yyKVLPJ954k6lX9vOSI8PVtXx698jaPw+DeZ4ZffdaLt2xy/zmjL7Gep2RWVjqup7ONKDHpe1tH6HTqzMzBarHxC5F3is3YuqGOC4ClUOuDtYS/P49fvuhAGGh/v4w/n+jvR+hvv7PNBjNUv6nFiZmS32B2DtiFgVeII0/+kJIxuSmXUTJ1ZmNuZFxFRgJUmn5QGQryb1QZ0m6R8jG52ZdRMnVmY2Jkl6EJiUH59fWD4TmDlCYZlZl/NdgWZmZmYVcWJlZmZmVhEnVmZmZmYVcWJlZmZmVhEnVmZmZmYVcWJlZmZmVhEnVmZmZmYVcWJlZmZmVhEnVmZmZmYVaTvyekQsA5wCbAg8DXxY0uzC+m8DW5JmgQfYCZgAnA8sDzwM7C1paLOUmpmZmXW4Mi1W7waWk7QFcBhwYt36TYDtJG2T/80FjgbOlzQZuAvYv8qgzczMzDpRmcRqK+AqAEm3ApvVVuTWrLWB0yLi5ojYp/41wJXAtpVFbGZmZtahykzCvDIwt/B8YUT0SFoArAh8BzgJGA/cGBG3171mHrBKu4NMnLgCPT3jBxL7Evr6epf4v9U2VahyX1Ucr2w8wxn3cMfUaZ9JlduUNZh9DeZzavaa4f4MzMw6TZnE6nGgWFsuk5MqgPnAt2r9pyLiBlJfrNpr/pP/f6zdQebMGVoXrP7+efT19dLfP6/lNlWpcl/ttCsXlI9nOOMuc6wyZavyeFUqc7yqtilrMPsa6Hen1Wc2kOM7CTOz0ajMpcCbgR0AImIScE9h3TrArIgYHxETSJcA7yy+BtgeuKmyiM3MzMw6VJkWqxnA2yLiFmAcsHdEHAzMlnRZRPwIuBV4FjhH0u8j4hjg7Ij4CPAIMHUpxW9mZmbWMdomVpKeAw6oW/zHwvrjgePrXvNv4B1VBGhmZmbWLTxAqJmZmVlFnFiZmZmZVcSJlZmZmVlFnFiZmZmZVcSJlZmZmVlFnFiZmZmZVcSJlZmZmVlFnFiZmZmZVcSJlZmZmVlFykxpY2Y2qkTEMsAppEnjnwY+LGl2Yf23gS2B2qzSO0maO+yBmlnXcWJlZmPRu4HlJG2RJ5c/EdipsH4TYDtJj4xIdGbWtXwp0MzGoq2AqwAk3QpsVluRW7PWBk6LiJsjYp+RCdHMupFbrMxsLFoZKF7aWxgRPZIWACsC3wFOAsYDN0bE7ZLubraziRNXoKdn/KCD6evrHfRrl+a+Ou34jfZd1fEGu5/hPv5wv78juZ+q9zVcx3JiZWZj0eNAsRZdJidVAPOBb0maDxARN5D6YjVNrObMmT+kYPr757XfqIS+vt7K9jVYS/P49fuusryD3c9wH38439+R3s9wf58HeqxmiZgvBZrZWHQzsANA7mN1T2HdOsCsiBgfERNIlw3vHP4QzawbucXKzMaiGcDbIuIWYBywd0QcDMyWdFlE/Ai4FXgWOEfS70cwVjPrIk6szGzMkfQccEDd4j8W1h8PHD+sQZnZqOBLgWZmZmYVadtiVWIgvU8DH8hPr5D0pYgYBzwE/Ckv/5WkwyuN3MzMzKzDlLkU2HQgvYhYE/ggsDmwCLgpImaQ7qq5U9KUpRO2mZmZWecpcymw6UB6wN+Bd0hamPssTACeAjYFXhkRN0bEFRERFcdtZmZm1nHKtFg1HUhP0rPAI/nS3zeAuyTdHxGvAI6VdFFEbAWcB7yh1UGqGmCv1QBf3TpoWZnjdcLgcoM9VicOSlfV8arapqzB7Gswn1Oz14z04JRmZiOtTGLVaiA9ImI5YBppstKP5cW3AwsAJM2KiFdGxDhJi5odpIoB9toNJlblQGPDOWhZmUHSOmFwucEcqxMG+Fuax6tqm7IGs6+BfndafWYDOb6TMDMbjcpcCmw6kF5uqboU+J2k/SUtzKu+AHwqb7Mh8LdWSZWZmZnZaFCmxarpQHqkebS2BpaNiO3z9ocDXwfOi4h3klquPlR14GZmZmadpm1i1W4gPWC5Ji9952CDMjMzM+tGHiDUzMzMrCJOrMzMzMwq4sTKzMzMrCJOrMzMzMwq4sTKzMzMrCJOrMzMzMwq4sTKzMzMrCJOrMzMzMwq4sTKzMzMrCJOrMzMzMwq4sTKzMzMrCJOrMzMzMwq4sTKzMzMrCJOrMzMzMwq4sTKzMzMrCJOrMzMzMwq4sTKzMzMrCJOrMzMzMwq0tNug4hYBjgF2BB4GviwpNmF9R8B9gcWAMdI+llEvBQ4H1geeBjYW9L8pRC/mdmADaZeG5FAzazrlGmxejewnKQtgMOAE2srIuIVwCeBLYHtgGMjYlngaOB8SZOBu0gVlJlZpxhMvWZm1laZxGor4CoASbcCmxXWvRG4WdLTkuYCs4HXF18DXAlsW1nEZmZDN5h6zcysrXGLFi1quUFEnAFMl3Rlfv43YE1JCyJid2ADSZ/L684BzgFOzcv/ExFrAudI2mppFsTMrKzB1GuSrhu5iM2sW5RpsXoc6C2+RtKCJut6gcfqlteWmZl1isHUa2ZmbZVJrG4GdgCIiEnAPYV1twGTI2K5iFgFWBe4t/gaYHvgpsoiNjMbusHUa2ZmbZW5FFi7e+b1wDhgb1KFNFvSZfnumf1ISdrXJE2PiJcDZ5PO9B4Bpkp6cukVw8ysvMHUayMWrJl1lbaJlZmZmZmV4wFCzczMzCrixMrMzMysIm1HXu80JUZM/jTwgfz0CklfGv4oB6dd2QrbXA5cKunU4Y9y4Ep8ZtsDX8hP7wQ+LqkrrlGXKNtngN2A50h9dWaMSKCDFBGbA8dJ2qZu+RTSQMALgGmSTh+B8Ma8iFgeOA94GTAP2EtSf9023yCN29UDnNbNn1WZ8ubt1gJ+Kmn9YQ6xEmNpZoCSf/f6gFtIw6A8NfxRDkw3tli1GjF5TeCDwJuALYC3R0Q3DezXtGwFxwCrDmtUQ9fqM+sFvgHsKGkS8CDw0pEIcpBale3FpBG8twDeDpw8IhEOUkQcCpwBLFe3fALwTVKZtgb2y6OV2/D7KHBPnuXiHODI4sqIeAuwVv5+bgV8LiImDn+YlWlZXoCI2AO4gO6qR+qNpZkBWv7di4jtgGuAl49AbIPSjYlVqxGT/w68Q9JCSc8BE4COz24LWpWNiNiF1PJx5fCHNiStyvUm0q3uJ0bETcC/G52BdrBWZXsS+CuwYv733LBHNzQPADs3WL4u6e65OZKeAWYBk4c1MqtpN8vFr4B98uNFwHjg2eEJbakoM6vHHFLC383G0swALf/ukerNbYH/G+a4Bq3rLgUCKwNzC88XRkSPpAWSngUeiYhxpFaQuyTdPyJRDk7TskXE+sBUYBfSJZhu0rRcpLPKtwAbAU8AN0XEr7roc2tVNkjJ/n2kP2jHDndwQ5GHTlmjwar6Ms8DVhmWoMawiNgX+HTd4n+z+LN4weeQL5s8lVsZzyZdCnxiacdahcGUF6B2WSwilmp8S1mremW0/f5a1qGSroXu+jy7MbFqNWIyEbEcMI30ZfvYMMc2VK3KtifwSuAGYA3gmYh4UNJVdL5W5XoU+I2kfwFExC9JSVa3JFatyrY9sBrwX/n51RFxs6TbhjPApcAjk48ASWcCZxaXRcQltJnlIl/6uxj4uaSuSe4HW95RYizNDNDyb3o36sZLgU1HTM4tVZcCv5O0v6SFIxPioDUtm6RDJW2eOxGfBZzUJUkVtB7l+g5g/Yh4aUT0AJNILTzdolXZ5gD/AZ7OLQePAS8e9gir9wdg7YhYNSJeBLyZdMnJhl/LWS5yZ+/rSTcYfGWYY1saxsqsHmNpZoBWZe1K3dhiNQN4W0TcQh4xOSIOJl1nHk+6tr5svtMM4HBJ3VLpNy2bpMtGNrQhaVmuiDgcuDpv+xNJ3VRJtCvbtsCtEfEcqS/StSMY65BExFRgJUmn5TJeTTo5mybpHyMb3Zj1feDsiJgFPEPqLkBEHE9qpdoSWBP4SL6TDGBvSX8ZiWAr0LK8o6A1uKZdvfJtUlK5DPD5brhTroVR93fPI6+bmZmZVaQbLwWamZmZdSQnVmZmZmYVcWJlZmZmVhEnVmZmZmYVcWJlZmZmVhEnVmZmZmYVcWJlZmZmVhEnVmZmZmYV+f8WB9ML23EI0gAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1acf243780>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ad584ec50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ad57ebf60>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ad5813860>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1acf14ec88>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1aa8210cf8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_layer_stats(pipeline.named_steps['classifier'].module_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/svg+xml": [
       "<?xml version=\"1.0\" encoding=\"UTF-8\" standalone=\"no\"?>\n",
       "<!DOCTYPE svg PUBLIC \"-//W3C//DTD SVG 1.1//EN\"\n",
       " \"http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd\">\n",
       "<!-- Generated by graphviz version 2.40.1 (20161225.0304)\n",
       " -->\n",
       "<!-- Title: %3 Pages: 1 -->\n",
       "<svg width=\"482pt\" height=\"456pt\"\n",
       " viewBox=\"0.00 0.00 482.32 456.00\" xmlns=\"http://www.w3.org/2000/svg\" xmlns:xlink=\"http://www.w3.org/1999/xlink\">\n",
       "<g id=\"graph0\" class=\"graph\" transform=\"scale(1 1) rotate(0) translate(4 452)\">\n",
       "<title>%3</title>\n",
       "<polygon fill=\"#ffffff\" stroke=\"transparent\" points=\"-4,4 -4,-452 478.3242,-452 478.3242,4 -4,4\"/>\n",
       "<!-- 115249301488 -->\n",
       "<g id=\"node1\" class=\"node\">\n",
       "<title>115249301488</title>\n",
       "<polygon fill=\"#d3d3d3\" stroke=\"#000000\" points=\"307.3027,-20 203.3457,-20 203.3457,0 307.3027,0 307.3027,-20\"/>\n",
       "<text text-anchor=\"middle\" x=\"255.3242\" y=\"-6.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\">AddmmBackward</text>\n",
       "</g>\n",
       "<!-- 115144293008 -->\n",
       "<g id=\"node2\" class=\"node\">\n",
       "<title>115144293008</title>\n",
       "<polygon fill=\"#d3d3d3\" stroke=\"#000000\" points=\"185.1254,-76 83.523,-76 83.523,-56 185.1254,-56 185.1254,-76\"/>\n",
       "<text text-anchor=\"middle\" x=\"134.3242\" y=\"-62.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\">ExpandBackward</text>\n",
       "</g>\n",
       "<!-- 115144293008&#45;&gt;115249301488 -->\n",
       "<g id=\"edge1\" class=\"edge\">\n",
       "<title>115144293008&#45;&gt;115249301488</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M155.9566,-55.9883C174.9379,-47.2036 202.771,-34.3222 224.1549,-24.4255\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"225.866,-27.4903 233.4712,-20.1138 222.9259,-21.1376 225.866,-27.4903\"/>\n",
       "</g>\n",
       "<!-- 115146658424 -->\n",
       "<g id=\"node3\" class=\"node\">\n",
       "<title>115146658424</title>\n",
       "<polygon fill=\"#add8e6\" stroke=\"#000000\" points=\"161.3242,-144 107.3242,-144 107.3242,-112 161.3242,-112 161.3242,-144\"/>\n",
       "<text text-anchor=\"middle\" x=\"134.3242\" y=\"-130.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\">None</text>\n",
       "<text text-anchor=\"middle\" x=\"134.3242\" y=\"-118.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\"> (2)</text>\n",
       "</g>\n",
       "<!-- 115146658424&#45;&gt;115144293008 -->\n",
       "<g id=\"edge2\" class=\"edge\">\n",
       "<title>115146658424&#45;&gt;115144293008</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M134.3242,-111.7102C134.3242,-104.0144 134.3242,-94.744 134.3242,-86.5691\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"137.8243,-86.3512 134.3242,-76.3512 130.8243,-86.3513 137.8243,-86.3512\"/>\n",
       "</g>\n",
       "<!-- 114488233568 -->\n",
       "<g id=\"node4\" class=\"node\">\n",
       "<title>114488233568</title>\n",
       "<polygon fill=\"#d3d3d3\" stroke=\"#000000\" points=\"307.4655,-76 203.183,-76 203.183,-56 307.4655,-56 307.4655,-76\"/>\n",
       "<text text-anchor=\"middle\" x=\"255.3242\" y=\"-62.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\">DropoutBackward</text>\n",
       "</g>\n",
       "<!-- 114488233568&#45;&gt;115249301488 -->\n",
       "<g id=\"edge3\" class=\"edge\">\n",
       "<title>114488233568&#45;&gt;115249301488</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M255.3242,-55.9883C255.3242,-48.9098 255.3242,-39.1714 255.3242,-30.4779\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"258.8243,-30.3038 255.3242,-20.3039 251.8243,-30.3039 258.8243,-30.3038\"/>\n",
       "</g>\n",
       "<!-- 115248866080 -->\n",
       "<g id=\"node5\" class=\"node\">\n",
       "<title>115248866080</title>\n",
       "<polygon fill=\"#d3d3d3\" stroke=\"#000000\" points=\"301.1353,-138 209.5131,-138 209.5131,-118 301.1353,-118 301.1353,-138\"/>\n",
       "<text text-anchor=\"middle\" x=\"255.3242\" y=\"-124.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\">AddBackward1</text>\n",
       "</g>\n",
       "<!-- 115248866080&#45;&gt;114488233568 -->\n",
       "<g id=\"edge4\" class=\"edge\">\n",
       "<title>115248866080&#45;&gt;114488233568</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M255.3242,-117.762C255.3242,-109.185 255.3242,-96.6836 255.3242,-86.1154\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"258.8243,-86.0475 255.3242,-76.0475 251.8243,-86.0476 258.8243,-86.0475\"/>\n",
       "</g>\n",
       "<!-- 115248868376 -->\n",
       "<g id=\"node6\" class=\"node\">\n",
       "<title>115248868376</title>\n",
       "<polygon fill=\"#d3d3d3\" stroke=\"#000000\" points=\"242.1254,-200 122.5231,-200 122.5231,-180 242.1254,-180 242.1254,-200\"/>\n",
       "<text text-anchor=\"middle\" x=\"182.3242\" y=\"-186.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\">ThresholdBackward1</text>\n",
       "</g>\n",
       "<!-- 115248868376&#45;&gt;115248866080 -->\n",
       "<g id=\"edge5\" class=\"edge\">\n",
       "<title>115248868376&#45;&gt;115248866080</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M194.3787,-179.762C205.6467,-170.1918 222.6674,-155.7359 235.8601,-144.5312\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"238.1378,-147.1887 243.4941,-138.0475 233.6063,-141.8533 238.1378,-147.1887\"/>\n",
       "</g>\n",
       "<!-- 115248945696 -->\n",
       "<g id=\"node7\" class=\"node\">\n",
       "<title>115248945696</title>\n",
       "<polygon fill=\"#d3d3d3\" stroke=\"#000000\" points=\"241.4594,-256 89.189,-256 89.189,-236 241.4594,-236 241.4594,-256\"/>\n",
       "<text text-anchor=\"middle\" x=\"165.3242\" y=\"-242.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\">CudnnBatchNormBackward</text>\n",
       "</g>\n",
       "<!-- 115248945696&#45;&gt;115248868376 -->\n",
       "<g id=\"edge6\" class=\"edge\">\n",
       "<title>115248945696&#45;&gt;115248868376</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M168.3635,-235.9883C170.5601,-228.7525 173.6004,-218.7373 176.2832,-209.9\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"179.6405,-210.8894 179.1963,-200.3039 172.9423,-208.856 179.6405,-210.8894\"/>\n",
       "</g>\n",
       "<!-- 115249025936 -->\n",
       "<g id=\"node8\" class=\"node\">\n",
       "<title>115249025936</title>\n",
       "<polygon fill=\"#d3d3d3\" stroke=\"#000000\" points=\"84.4734,-318 .1751,-318 .1751,-298 84.4734,-298 84.4734,-318\"/>\n",
       "<text text-anchor=\"middle\" x=\"42.3242\" y=\"-304.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\">MmBackward</text>\n",
       "</g>\n",
       "<!-- 115249025936&#45;&gt;115248945696 -->\n",
       "<g id=\"edge7\" class=\"edge\">\n",
       "<title>115249025936&#45;&gt;115248945696</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M62.36,-297.9006C82.3736,-287.8125 113.3529,-272.1969 136.1055,-260.7281\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"137.9427,-263.7216 145.297,-256.095 134.7918,-257.4708 137.9427,-263.7216\"/>\n",
       "</g>\n",
       "<!-- 115249026104 -->\n",
       "<g id=\"node9\" class=\"node\">\n",
       "<title>115249026104</title>\n",
       "<polygon fill=\"#d3d3d3\" stroke=\"#000000\" points=\"78.2988,-380 6.3496,-380 6.3496,-360 78.2988,-360 78.2988,-380\"/>\n",
       "<text text-anchor=\"middle\" x=\"42.3242\" y=\"-366.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\">TBackward</text>\n",
       "</g>\n",
       "<!-- 115249026104&#45;&gt;115249025936 -->\n",
       "<g id=\"edge8\" class=\"edge\">\n",
       "<title>115249026104&#45;&gt;115249025936</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M42.3242,-359.762C42.3242,-351.185 42.3242,-338.6836 42.3242,-328.1154\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"45.8243,-328.0475 42.3242,-318.0475 38.8243,-328.0476 45.8243,-328.0475\"/>\n",
       "</g>\n",
       "<!-- 115248989016 -->\n",
       "<g id=\"node10\" class=\"node\">\n",
       "<title>115248989016</title>\n",
       "<polygon fill=\"#add8e6\" stroke=\"#000000\" points=\"70.8164,-448 13.832,-448 13.832,-416 70.8164,-416 70.8164,-448\"/>\n",
       "<text text-anchor=\"middle\" x=\"42.3242\" y=\"-434.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\">None</text>\n",
       "<text text-anchor=\"middle\" x=\"42.3242\" y=\"-422.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\"> (20, 34)</text>\n",
       "</g>\n",
       "<!-- 115248989016&#45;&gt;115249026104 -->\n",
       "<g id=\"edge9\" class=\"edge\">\n",
       "<title>115248989016&#45;&gt;115249026104</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M42.3242,-415.7102C42.3242,-408.0144 42.3242,-398.744 42.3242,-390.5691\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"45.8243,-390.3512 42.3242,-380.3512 38.8243,-390.3513 45.8243,-390.3512\"/>\n",
       "</g>\n",
       "<!-- 115249026384 -->\n",
       "<g id=\"node11\" class=\"node\">\n",
       "<title>115249026384</title>\n",
       "<polygon fill=\"#add8e6\" stroke=\"#000000\" points=\"156.3242,-324 102.3242,-324 102.3242,-292 156.3242,-292 156.3242,-324\"/>\n",
       "<text text-anchor=\"middle\" x=\"129.3242\" y=\"-310.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\">None</text>\n",
       "<text text-anchor=\"middle\" x=\"129.3242\" y=\"-298.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\"> (20)</text>\n",
       "</g>\n",
       "<!-- 115249026384&#45;&gt;115248945696 -->\n",
       "<g id=\"edge10\" class=\"edge\">\n",
       "<title>115249026384&#45;&gt;115248945696</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M138.7828,-291.7102C143.5052,-283.5771 149.2488,-273.6855 154.1835,-265.1867\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"157.3192,-266.7566 159.3138,-256.3512 151.2657,-263.2416 157.3192,-266.7566\"/>\n",
       "</g>\n",
       "<!-- 115249027616 -->\n",
       "<g id=\"node12\" class=\"node\">\n",
       "<title>115249027616</title>\n",
       "<polygon fill=\"#add8e6\" stroke=\"#000000\" points=\"228.3242,-324 174.3242,-324 174.3242,-292 228.3242,-292 228.3242,-324\"/>\n",
       "<text text-anchor=\"middle\" x=\"201.3242\" y=\"-310.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\">None</text>\n",
       "<text text-anchor=\"middle\" x=\"201.3242\" y=\"-298.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\"> (20)</text>\n",
       "</g>\n",
       "<!-- 115249027616&#45;&gt;115248945696 -->\n",
       "<g id=\"edge11\" class=\"edge\">\n",
       "<title>115249027616&#45;&gt;115248945696</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M191.8656,-291.7102C187.1432,-283.5771 181.3996,-273.6855 176.4649,-265.1867\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"179.3828,-263.2416 171.3346,-256.3512 173.3292,-266.7566 179.3828,-263.2416\"/>\n",
       "</g>\n",
       "<!-- 115248868488 -->\n",
       "<g id=\"node13\" class=\"node\">\n",
       "<title>115248868488</title>\n",
       "<polygon fill=\"#d3d3d3\" stroke=\"#000000\" points=\"387.1254,-200 267.5231,-200 267.5231,-180 387.1254,-180 387.1254,-200\"/>\n",
       "<text text-anchor=\"middle\" x=\"327.3242\" y=\"-186.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\">ThresholdBackward1</text>\n",
       "</g>\n",
       "<!-- 115248868488&#45;&gt;115248866080 -->\n",
       "<g id=\"edge12\" class=\"edge\">\n",
       "<title>115248868488&#45;&gt;115248866080</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M315.4349,-179.762C304.426,-170.2821 287.8499,-156.0083 274.891,-144.8492\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"276.8539,-141.9206 266.9923,-138.0475 272.2862,-147.225 276.8539,-141.9206\"/>\n",
       "</g>\n",
       "<!-- 115248946424 -->\n",
       "<g id=\"node14\" class=\"node\">\n",
       "<title>115248946424</title>\n",
       "<polygon fill=\"#d3d3d3\" stroke=\"#000000\" points=\"431.4594,-256 279.189,-256 279.189,-236 431.4594,-236 431.4594,-256\"/>\n",
       "<text text-anchor=\"middle\" x=\"355.3242\" y=\"-242.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\">CudnnBatchNormBackward</text>\n",
       "</g>\n",
       "<!-- 115248946424&#45;&gt;115248868488 -->\n",
       "<g id=\"edge13\" class=\"edge\">\n",
       "<title>115248946424&#45;&gt;115248868488</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M350.3184,-235.9883C346.6218,-228.5952 341.4745,-218.3005 336.9869,-209.3253\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"340.0788,-207.6829 332.4762,-200.3039 333.8178,-210.8134 340.0788,-207.6829\"/>\n",
       "</g>\n",
       "<!-- 115248988624 -->\n",
       "<g id=\"node15\" class=\"node\">\n",
       "<title>115248988624</title>\n",
       "<polygon fill=\"#d3d3d3\" stroke=\"#000000\" points=\"330.4734,-318 246.1751,-318 246.1751,-298 330.4734,-298 330.4734,-318\"/>\n",
       "<text text-anchor=\"middle\" x=\"288.3242\" y=\"-304.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\">MmBackward</text>\n",
       "</g>\n",
       "<!-- 115248988624&#45;&gt;115248946424 -->\n",
       "<g id=\"edge14\" class=\"edge\">\n",
       "<title>115248988624&#45;&gt;115248946424</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M299.3879,-297.762C309.6323,-288.2821 325.0572,-274.0083 337.1162,-262.8492\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"339.5039,-265.4083 344.4664,-256.0475 334.7496,-260.2706 339.5039,-265.4083\"/>\n",
       "</g>\n",
       "<!-- 115248500920 -->\n",
       "<g id=\"node16\" class=\"node\">\n",
       "<title>115248500920</title>\n",
       "<polygon fill=\"#d3d3d3\" stroke=\"#000000\" points=\"324.2988,-380 252.3496,-380 252.3496,-360 324.2988,-360 324.2988,-380\"/>\n",
       "<text text-anchor=\"middle\" x=\"288.3242\" y=\"-366.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\">TBackward</text>\n",
       "</g>\n",
       "<!-- 115248500920&#45;&gt;115248988624 -->\n",
       "<g id=\"edge15\" class=\"edge\">\n",
       "<title>115248500920&#45;&gt;115248988624</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M288.3242,-359.762C288.3242,-351.185 288.3242,-338.6836 288.3242,-328.1154\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"291.8243,-328.0475 288.3242,-318.0475 284.8243,-328.0476 291.8243,-328.0475\"/>\n",
       "</g>\n",
       "<!-- 115248501088 -->\n",
       "<g id=\"node17\" class=\"node\">\n",
       "<title>115248501088</title>\n",
       "<polygon fill=\"#add8e6\" stroke=\"#000000\" points=\"316.8164,-448 259.832,-448 259.832,-416 316.8164,-416 316.8164,-448\"/>\n",
       "<text text-anchor=\"middle\" x=\"288.3242\" y=\"-434.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\">None</text>\n",
       "<text text-anchor=\"middle\" x=\"288.3242\" y=\"-422.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\"> (20, 34)</text>\n",
       "</g>\n",
       "<!-- 115248501088&#45;&gt;115248500920 -->\n",
       "<g id=\"edge16\" class=\"edge\">\n",
       "<title>115248501088&#45;&gt;115248500920</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M288.3242,-415.7102C288.3242,-408.0144 288.3242,-398.744 288.3242,-390.5691\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"291.8243,-390.3512 288.3242,-380.3512 284.8243,-390.3513 291.8243,-390.3512\"/>\n",
       "</g>\n",
       "<!-- 115248991816 -->\n",
       "<g id=\"node18\" class=\"node\">\n",
       "<title>115248991816</title>\n",
       "<polygon fill=\"#add8e6\" stroke=\"#000000\" points=\"402.3242,-324 348.3242,-324 348.3242,-292 402.3242,-292 402.3242,-324\"/>\n",
       "<text text-anchor=\"middle\" x=\"375.3242\" y=\"-310.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\">None</text>\n",
       "<text text-anchor=\"middle\" x=\"375.3242\" y=\"-298.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\"> (20)</text>\n",
       "</g>\n",
       "<!-- 115248991816&#45;&gt;115248946424 -->\n",
       "<g id=\"edge17\" class=\"edge\">\n",
       "<title>115248991816&#45;&gt;115248946424</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M370.0695,-291.7102C367.5305,-283.8395 364.4603,-274.3218 361.7802,-266.0134\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"365.0644,-264.7938 358.6633,-256.3512 358.4024,-266.9428 365.0644,-264.7938\"/>\n",
       "</g>\n",
       "<!-- 115248990976 -->\n",
       "<g id=\"node19\" class=\"node\">\n",
       "<title>115248990976</title>\n",
       "<polygon fill=\"#add8e6\" stroke=\"#000000\" points=\"474.3242,-324 420.3242,-324 420.3242,-292 474.3242,-292 474.3242,-324\"/>\n",
       "<text text-anchor=\"middle\" x=\"447.3242\" y=\"-310.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\">None</text>\n",
       "<text text-anchor=\"middle\" x=\"447.3242\" y=\"-298.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\"> (20)</text>\n",
       "</g>\n",
       "<!-- 115248990976&#45;&gt;115248946424 -->\n",
       "<g id=\"edge18\" class=\"edge\">\n",
       "<title>115248990976&#45;&gt;115248946424</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M423.1523,-291.7102C409.4393,-282.4688 392.3571,-270.9569 378.7663,-261.7979\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"380.6586,-258.8526 370.4099,-256.1664 376.7466,-264.6575 380.6586,-258.8526\"/>\n",
       "</g>\n",
       "<!-- 115144328976 -->\n",
       "<g id=\"node20\" class=\"node\">\n",
       "<title>115144328976</title>\n",
       "<polygon fill=\"#d3d3d3\" stroke=\"#000000\" points=\"397.2988,-76 325.3496,-76 325.3496,-56 397.2988,-56 397.2988,-76\"/>\n",
       "<text text-anchor=\"middle\" x=\"361.3242\" y=\"-62.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\">TBackward</text>\n",
       "</g>\n",
       "<!-- 115144328976&#45;&gt;115249301488 -->\n",
       "<g id=\"edge19\" class=\"edge\">\n",
       "<title>115144328976&#45;&gt;115249301488</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M342.3735,-55.9883C325.97,-47.3223 302.0202,-34.6696 283.3921,-24.8283\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"284.9451,-21.6904 274.4682,-20.1138 281.6752,-27.8798 284.9451,-21.6904\"/>\n",
       "</g>\n",
       "<!-- 115248868544 -->\n",
       "<g id=\"node21\" class=\"node\">\n",
       "<title>115248868544</title>\n",
       "<polygon fill=\"#add8e6\" stroke=\"#000000\" points=\"388.3242,-144 334.3242,-144 334.3242,-112 388.3242,-112 388.3242,-144\"/>\n",
       "<text text-anchor=\"middle\" x=\"361.3242\" y=\"-130.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\">None</text>\n",
       "<text text-anchor=\"middle\" x=\"361.3242\" y=\"-118.4\" font-family=\"Times,serif\" font-size=\"12.00\" fill=\"#000000\"> (2, 20)</text>\n",
       "</g>\n",
       "<!-- 115248868544&#45;&gt;115144328976 -->\n",
       "<g id=\"edge20\" class=\"edge\">\n",
       "<title>115248868544&#45;&gt;115144328976</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M361.3242,-111.7102C361.3242,-104.0144 361.3242,-94.744 361.3242,-86.5691\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"364.8243,-86.3512 361.3242,-76.3512 357.8243,-86.3513 364.8243,-86.3512\"/>\n",
       "</g>\n",
       "</g>\n",
       "</svg>\n"
      ],
      "text/plain": [
       "<graphviz.dot.Digraph at 0x1ad564c198>"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "plot_model_graph(pipeline.named_steps['classifier'].module_(to_var(torch.from_numpy(X_test).float())), \n",
    "                 pipeline.named_steps['classifier'].module_.state_dict())"
   ]
  },
  {
   "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": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}