MNIST.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\nildip.mukherjee\\appdata\\local\\continuum\\anaconda3\\envs\\python_36\\lib\\site-packages\\h5py\\__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n",
      "  from ._conv import register_converters as _register_converters\n",
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "from keras.datasets import mnist\n",
    "from keras.utils import np_utils, plot_model\n",
    "from keras.callbacks import EarlyStopping\n",
    "\n",
    "from numpy.random.mtrand import _rand as global_randstate\n",
    "global_randstate.seed(666)\n",
    "\n",
    "import os\n",
    "os.environ['PATH']+=os.pathsep+'C:\\\\Program Files (x86)\\\\Graphviz2.38\\\\bin'\n",
    "\n",
    "from models import * \n",
    "from utils import * "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\nildip.mukherjee\\appdata\\local\\continuum\\anaconda3\\envs\\python_36\\lib\\site-packages\\ipykernel_launcher.py:9: DeprecationWarning: Both axis > a.ndim and axis < -a.ndim - 1 are deprecated and will raise an AxisError in the future.\n",
      "  if __name__ == '__main__':\n",
      "c:\\users\\nildip.mukherjee\\appdata\\local\\continuum\\anaconda3\\envs\\python_36\\lib\\site-packages\\ipykernel_launcher.py:10: DeprecationWarning: Both axis > a.ndim and axis < -a.ndim - 1 are deprecated and will raise an AxisError in the future.\n",
      "  # Remove the CWD from sys.path while we load stuff.\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": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# loading the data\n",
    "(X_train, y_train), (X_test, y_test) = mnist.load_data()\n",
    "\n",
    "# normalizing the values\n",
    "X_train = X_train / 255\n",
    "X_test = X_test / 255\n",
    "\n",
    "# reshaping data for 2D CNN\n",
    "X_train_2D = np.expand_dims(X_train, axis=4)\n",
    "X_test_2D = np.expand_dims(X_test, axis=4)\n",
    "\n",
    "# reshaping data for 1D CNN\n",
    "X_train_1D = np.expand_dims(X_train.reshape(X_train.shape[0], X_train.shape[1]*X_train.shape[2]).astype('float32'), axis = 2)\n",
    "X_test_1D = np.expand_dims(X_test.reshape(X_test.shape[0], X_train.shape[1]*X_train.shape[2]).astype('float32'), axis = 2)\n",
    "\n",
    "# reshaping data for basic NN\n",
    "X_train_nn = X_train.reshape(X_train.shape[0], X_train.shape[1]*X_train.shape[2]).astype('float32')\n",
    "X_test_nn = X_test.reshape(X_test.shape[0], X_train.shape[1]*X_train.shape[2]).astype('float32')\n",
    "\n",
    "# one hot encode output\n",
    "Y_train = np_utils.to_categorical(y_train)\n",
    "Y_test = np_utils.to_categorical(y_test)\n",
    "\n",
    "# plotting the average image for each number\n",
    "for i in range(0,10):\n",
    "    plt.figure()\n",
    "    plt.title('Mean plot of all {0}s'.format(i))\n",
    "    plt.tight_layout()\n",
    "    plt.imshow(np.mean(X_train[np.where(y_train == i)], axis=0), cmap = plt.get_cmap('gray_r'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         (None, 784)               0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 500)               392500    \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 500)               0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 100)               50100     \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 100)               0         \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 25)                2525      \n",
      "_________________________________________________________________\n",
      "dropout_3 (Dropout)          (None, 25)                0         \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              (None, 10)                260       \n",
      "=================================================================\n",
      "Total params: 445,385\n",
      "Trainable params: 445,385\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/200\n",
      "60000/60000 [==============================] - 6s 96us/step - loss: 0.7483 - acc: 0.7679 - val_loss: 0.2015 - val_acc: 0.9419\n",
      "Epoch 2/200\n",
      "60000/60000 [==============================] - 5s 91us/step - loss: 0.3151 - acc: 0.9160 - val_loss: 0.1460 - val_acc: 0.9605\n",
      "Epoch 3/200\n",
      "60000/60000 [==============================] - 5s 91us/step - loss: 0.2431 - acc: 0.9370 - val_loss: 0.1257 - val_acc: 0.9642\n",
      "Epoch 4/200\n",
      "60000/60000 [==============================] - 6s 92us/step - loss: 0.2038 - acc: 0.9475 - val_loss: 0.1049 - val_acc: 0.9706\n",
      "Epoch 5/200\n",
      "60000/60000 [==============================] - 6s 92us/step - loss: 0.1774 - acc: 0.9544 - val_loss: 0.0993 - val_acc: 0.9735\n",
      "Epoch 6/200\n",
      "60000/60000 [==============================] - 6s 92us/step - loss: 0.1571 - acc: 0.9595 - val_loss: 0.0948 - val_acc: 0.9759\n",
      "Epoch 7/200\n",
      "60000/60000 [==============================] - 5s 92us/step - loss: 0.1447 - acc: 0.9625 - val_loss: 0.0953 - val_acc: 0.9765\n",
      "Epoch 8/200\n",
      "60000/60000 [==============================] - 5s 91us/step - loss: 0.1357 - acc: 0.9644 - val_loss: 0.0880 - val_acc: 0.9774\n",
      "Epoch 9/200\n",
      "60000/60000 [==============================] - 6s 92us/step - loss: 0.1269 - acc: 0.9668 - val_loss: 0.0900 - val_acc: 0.9765\n",
      "Epoch 10/200\n",
      "60000/60000 [==============================] - 6s 93us/step - loss: 0.1197 - acc: 0.9689 - val_loss: 0.0814 - val_acc: 0.9792\n",
      "Epoch 11/200\n",
      "60000/60000 [==============================] - 6s 92us/step - loss: 0.1129 - acc: 0.9702 - val_loss: 0.0804 - val_acc: 0.9806\n",
      "Epoch 12/200\n",
      "60000/60000 [==============================] - 6s 93us/step - loss: 0.1068 - acc: 0.9721 - val_loss: 0.0810 - val_acc: 0.9799\n",
      "Epoch 13/200\n",
      "60000/60000 [==============================] - 6s 92us/step - loss: 0.1028 - acc: 0.9732 - val_loss: 0.0755 - val_acc: 0.9802\n",
      "Epoch 14/200\n",
      "60000/60000 [==============================] - 6s 92us/step - loss: 0.0981 - acc: 0.9739 - val_loss: 0.0783 - val_acc: 0.9809\n",
      "Epoch 15/200\n",
      "60000/60000 [==============================] - 6s 93us/step - loss: 0.0919 - acc: 0.9757 - val_loss: 0.0793 - val_acc: 0.9822\n",
      "Model Error: 1.78%\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model = simple_nn(10,784)\n",
    "model.summary()\n",
    "# print model structure to png file\n",
    "plot_model(model, to_file='model_simple_nn.png', show_shapes = True)\n",
    "# complile the model\n",
    "model.compile(loss = 'categorical_crossentropy', optimizer = 'adam', metrics = ['accuracy'])\n",
    "# Fit the model\n",
    "early_stopping = EarlyStopping(monitor = 'val_loss', patience=2)\n",
    "model.fit(X_train_nn, Y_train, validation_data = (X_test_nn, Y_test), epochs=200, batch_size=200, callbacks=[early_stopping])\n",
    "# Final evaluation of the model\n",
    "scores = model.evaluate(X_test_nn, Y_test, verbose=0)\n",
    "print(\"Model Error: %.2f%%\" % (100-scores[1]*100))\n",
    "# plotting the confusion matrix\n",
    "plot_confusion_matrix(y_test, model.predict(X_test_nn).argmax(axis= 1), classes=[0,1,2,3,4,5,6,7,8,9])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Nildip.mukherjee\\AnacondaProjects\\Others\\MNIST\\models.py:29: UserWarning: Update your `Conv1D` call to the Keras 2 API: `Conv1D(kernel_size=15, strides=1, activation=\"relu\", filters=5)`\n",
      "  cnn_layer = Conv1D(nb_filter = 5, kernel_size = 15, strides = 1, activation = 'relu')(input_shape)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_2 (InputLayer)         (None, 784, 1)            0         \n",
      "_________________________________________________________________\n",
      "conv1d_1 (Conv1D)            (None, 770, 5)            80        \n",
      "_________________________________________________________________\n",
      "dropout_4 (Dropout)          (None, 770, 5)            0         \n",
      "_________________________________________________________________\n",
      "max_pooling1d_1 (MaxPooling1 (None, 385, 5)            0         \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 1925)              0         \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 500)               963000    \n",
      "_________________________________________________________________\n",
      "dropout_5 (Dropout)          (None, 500)               0         \n",
      "_________________________________________________________________\n",
      "dense_6 (Dense)              (None, 100)               50100     \n",
      "_________________________________________________________________\n",
      "dropout_6 (Dropout)          (None, 100)               0         \n",
      "_________________________________________________________________\n",
      "dense_7 (Dense)              (None, 10)                1010      \n",
      "=================================================================\n",
      "Total params: 1,014,190\n",
      "Trainable params: 1,014,190\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/200\n",
      "60000/60000 [==============================] - 28s 474us/step - loss: 0.4646 - acc: 0.8556 - val_loss: 0.1493 - val_acc: 0.9581\n",
      "Epoch 2/200\n",
      "60000/60000 [==============================] - 28s 474us/step - loss: 0.2067 - acc: 0.9382 - val_loss: 0.1124 - val_acc: 0.9693\n",
      "Epoch 3/200\n",
      "60000/60000 [==============================] - 28s 474us/step - loss: 0.1601 - acc: 0.9524 - val_loss: 0.0928 - val_acc: 0.9729\n",
      "Epoch 4/200\n",
      "60000/60000 [==============================] - 28s 473us/step - loss: 0.1395 - acc: 0.9581 - val_loss: 0.0780 - val_acc: 0.9771\n",
      "Epoch 5/200\n",
      "60000/60000 [==============================] - 28s 474us/step - loss: 0.1248 - acc: 0.9633 - val_loss: 0.0766 - val_acc: 0.9780\n",
      "Epoch 6/200\n",
      "60000/60000 [==============================] - 29s 476us/step - loss: 0.1119 - acc: 0.9666 - val_loss: 0.0668 - val_acc: 0.9792\n",
      "Epoch 7/200\n",
      "60000/60000 [==============================] - 29s 478us/step - loss: 0.1020 - acc: 0.9695 - val_loss: 0.0648 - val_acc: 0.9787\n",
      "Epoch 8/200\n",
      "60000/60000 [==============================] - 29s 477us/step - loss: 0.0951 - acc: 0.9714 - val_loss: 0.0588 - val_acc: 0.9823\n",
      "Epoch 9/200\n",
      "60000/60000 [==============================] - 29s 477us/step - loss: 0.0852 - acc: 0.9739 - val_loss: 0.0588 - val_acc: 0.9820\n",
      "Epoch 10/200\n",
      "60000/60000 [==============================] - 29s 477us/step - loss: 0.0837 - acc: 0.9739 - val_loss: 0.0543 - val_acc: 0.9836\n",
      "Epoch 11/200\n",
      "60000/60000 [==============================] - 29s 484us/step - loss: 0.0800 - acc: 0.9749 - val_loss: 0.0540 - val_acc: 0.9831\n",
      "Epoch 12/200\n",
      "60000/60000 [==============================] - 29s 478us/step - loss: 0.0762 - acc: 0.9762 - val_loss: 0.0552 - val_acc: 0.9832\n",
      "Epoch 13/200\n",
      "60000/60000 [==============================] - 29s 481us/step - loss: 0.0747 - acc: 0.9774 - val_loss: 0.0498 - val_acc: 0.9831\n",
      "Epoch 14/200\n",
      "60000/60000 [==============================] - 29s 479us/step - loss: 0.0666 - acc: 0.9788 - val_loss: 0.0518 - val_acc: 0.9850\n",
      "Epoch 15/200\n",
      "60000/60000 [==============================] - 29s 479us/step - loss: 0.0681 - acc: 0.9789 - val_loss: 0.0514 - val_acc: 0.9831\n",
      "Model Error: 1.69%\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model = simple_1Dcnn(10,784)\n",
    "model.summary()\n",
    "# print model structure to png file\n",
    "plot_model(model, to_file='model_simple_1Dcnn.png', show_shapes = True)\n",
    "# complile the model\n",
    "model.compile(loss = 'categorical_crossentropy', optimizer = 'adam', metrics = ['accuracy'])\n",
    "# Fit the model\n",
    "early_stopping = EarlyStopping(monitor = 'val_loss', patience=2)\n",
    "model.fit(X_train_1D, Y_train, validation_data = (X_test_1D, Y_test), epochs=200, batch_size=200, callbacks=[early_stopping])\n",
    "# Final evaluation of the model\n",
    "scores = model.evaluate(X_test_1D, Y_test, verbose=0)\n",
    "print(\"Model Error: %.2f%%\" % (100-scores[1]*100))\n",
    "# plotting the confusion matrix\n",
    "plot_confusion_matrix(y_test, model.predict(X_test_1D).argmax(axis= 1), classes=[0,1,2,3,4,5,6,7,8,9])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Nildip.mukherjee\\AnacondaProjects\\Others\\MNIST\\models.py:52: UserWarning: Update your `Conv2D` call to the Keras 2 API: `Conv2D(kernel_size=15, strides=1, activation=\"relu\", filters=5)`\n",
      "  cnn_layer = Conv2D(nb_filter = 5, kernel_size = 15, strides = 1, activation = 'relu')(input_shape)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_3 (InputLayer)         (None, 28, 28, 1)         0         \n",
      "_________________________________________________________________\n",
      "conv2d_1 (Conv2D)            (None, 14, 14, 5)         1130      \n",
      "_________________________________________________________________\n",
      "dropout_7 (Dropout)          (None, 14, 14, 5)         0         \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 7, 7, 5)           0         \n",
      "_________________________________________________________________\n",
      "flatten_2 (Flatten)          (None, 245)               0         \n",
      "_________________________________________________________________\n",
      "dense_8 (Dense)              (None, 500)               123000    \n",
      "_________________________________________________________________\n",
      "dropout_8 (Dropout)          (None, 500)               0         \n",
      "_________________________________________________________________\n",
      "dense_9 (Dense)              (None, 100)               50100     \n",
      "_________________________________________________________________\n",
      "dropout_9 (Dropout)          (None, 100)               0         \n",
      "_________________________________________________________________\n",
      "dense_10 (Dense)             (None, 10)                1010      \n",
      "=================================================================\n",
      "Total params: 175,240\n",
      "Trainable params: 175,240\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/200\n",
      "60000/60000 [==============================] - 33s 549us/step - loss: 0.4969 - acc: 0.8425 - val_loss: 0.1678 - val_acc: 0.9621\n",
      "Epoch 2/200\n",
      "60000/60000 [==============================] - 33s 544us/step - loss: 0.1912 - acc: 0.9441 - val_loss: 0.1172 - val_acc: 0.9743\n",
      "Epoch 3/200\n",
      "60000/60000 [==============================] - 33s 546us/step - loss: 0.1475 - acc: 0.9567 - val_loss: 0.1002 - val_acc: 0.9776\n",
      "Epoch 4/200\n",
      "60000/60000 [==============================] - 33s 546us/step - loss: 0.1265 - acc: 0.9624 - val_loss: 0.0855 - val_acc: 0.9812\n",
      "Epoch 5/200\n",
      "60000/60000 [==============================] - 33s 548us/step - loss: 0.1101 - acc: 0.9668 - val_loss: 0.0780 - val_acc: 0.9834\n",
      "Epoch 6/200\n",
      "60000/60000 [==============================] - 33s 546us/step - loss: 0.0991 - acc: 0.9710 - val_loss: 0.0748 - val_acc: 0.9837\n",
      "Epoch 7/200\n",
      "60000/60000 [==============================] - 33s 546us/step - loss: 0.0932 - acc: 0.9732 - val_loss: 0.0660 - val_acc: 0.9853\n",
      "Epoch 8/200\n",
      "60000/60000 [==============================] - 33s 546us/step - loss: 0.0868 - acc: 0.9740 - val_loss: 0.0599 - val_acc: 0.9856\n",
      "Epoch 9/200\n",
      "60000/60000 [==============================] - 33s 545us/step - loss: 0.0837 - acc: 0.9751 - val_loss: 0.0578 - val_acc: 0.9871\n",
      "Epoch 10/200\n",
      "60000/60000 [==============================] - 33s 546us/step - loss: 0.0758 - acc: 0.9777 - val_loss: 0.0586 - val_acc: 0.9866\n",
      "Epoch 11/200\n",
      "60000/60000 [==============================] - 33s 547us/step - loss: 0.0739 - acc: 0.9779 - val_loss: 0.0523 - val_acc: 0.9868\n",
      "Epoch 12/200\n",
      "60000/60000 [==============================] - 33s 551us/step - loss: 0.0732 - acc: 0.9782 - val_loss: 0.0561 - val_acc: 0.9852\n",
      "Epoch 13/200\n",
      "60000/60000 [==============================] - 33s 545us/step - loss: 0.0688 - acc: 0.9796 - val_loss: 0.0475 - val_acc: 0.9874\n",
      "Epoch 14/200\n",
      "60000/60000 [==============================] - 33s 550us/step - loss: 0.0653 - acc: 0.9805 - val_loss: 0.0463 - val_acc: 0.9877\n",
      "Epoch 15/200\n",
      "60000/60000 [==============================] - 33s 551us/step - loss: 0.0643 - acc: 0.9806 - val_loss: 0.0489 - val_acc: 0.9868\n",
      "Epoch 16/200\n",
      "60000/60000 [==============================] - 33s 549us/step - loss: 0.0612 - acc: 0.9819 - val_loss: 0.0482 - val_acc: 0.9867\n",
      "Model Error: 1.33%\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model = simple_2Dcnn(10,28,28)\n",
    "model.summary()\n",
    "# print model structure to png file\n",
    "plot_model(model, to_file='model_simple_2Dcnn.png', show_shapes = True)\n",
    "# complile the model\n",
    "model.compile(loss = 'categorical_crossentropy', optimizer = 'adam', metrics = ['accuracy'])\n",
    "# Fit the model\n",
    "early_stopping = EarlyStopping(monitor = 'val_loss', patience=2)\n",
    "model.fit(X_train_2D, Y_train, validation_data = (X_test_2D, Y_test), epochs=200, batch_size=200, callbacks=[early_stopping])\n",
    "# Final evaluation of the model\n",
    "scores = model.evaluate(X_test_2D, Y_test, verbose=0)\n",
    "print(\"Model Error: %.2f%%\" % (100-scores[1]*100))\n",
    "# plotting the confusion matrix\n",
    "plot_confusion_matrix(y_test, model.predict(X_test_2D).argmax(axis= 1), classes=[0,1,2,3,4,5,6,7,8,9])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Nildip.mukherjee\\AnacondaProjects\\Others\\MNIST\\models.py:75: UserWarning: Update your `Conv1D` call to the Keras 2 API: `Conv1D(kernel_size=15, strides=1, activation=\"relu\", filters=5)`\n",
      "  cnn_layer = Conv1D(nb_filter = 5, kernel_size = 15, strides = 1, activation = 'relu')(input_shape)\n",
      "C:\\Users\\Nildip.mukherjee\\AnacondaProjects\\Others\\MNIST\\models.py:78: UserWarning: Update your `Conv1D` call to the Keras 2 API: `Conv1D(kernel_size=10, strides=1, activation=\"relu\", filters=4)`\n",
      "  cnn_layer = Conv1D(nb_filter = 4, kernel_size = 10, strides = 1, activation = 'relu')(input_shape)\n",
      "C:\\Users\\Nildip.mukherjee\\AnacondaProjects\\Others\\MNIST\\models.py:81: UserWarning: Update your `Conv1D` call to the Keras 2 API: `Conv1D(kernel_size=5, strides=1, activation=\"relu\", filters=4)`\n",
      "  cnn_layer = Conv1D(nb_filter = 4, kernel_size = 5, strides = 1, activation = 'relu')(input_shape)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_4 (InputLayer)         (None, 784, 1)            0         \n",
      "_________________________________________________________________\n",
      "conv1d_4 (Conv1D)            (None, 780, 4)            24        \n",
      "_________________________________________________________________\n",
      "dropout_12 (Dropout)         (None, 780, 4)            0         \n",
      "_________________________________________________________________\n",
      "max_pooling1d_4 (MaxPooling1 (None, 390, 4)            0         \n",
      "_________________________________________________________________\n",
      "flatten_3 (Flatten)          (None, 1560)              0         \n",
      "_________________________________________________________________\n",
      "dense_11 (Dense)             (None, 500)               780500    \n",
      "_________________________________________________________________\n",
      "dropout_13 (Dropout)         (None, 500)               0         \n",
      "_________________________________________________________________\n",
      "dense_12 (Dense)             (None, 100)               50100     \n",
      "_________________________________________________________________\n",
      "dropout_14 (Dropout)         (None, 100)               0         \n",
      "_________________________________________________________________\n",
      "dense_13 (Dense)             (None, 10)                1010      \n",
      "=================================================================\n",
      "Total params: 831,634\n",
      "Trainable params: 831,634\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/200\n",
      "60000/60000 [==============================] - 21s 356us/step - loss: 0.4616 - acc: 0.8569 - val_loss: 0.1568 - val_acc: 0.9506\n",
      "Epoch 2/200\n",
      "60000/60000 [==============================] - 21s 349us/step - loss: 0.2105 - acc: 0.9374 - val_loss: 0.1153 - val_acc: 0.9672\n",
      "Epoch 3/200\n",
      "60000/60000 [==============================] - 21s 350us/step - loss: 0.1630 - acc: 0.9516 - val_loss: 0.0912 - val_acc: 0.9724\n",
      "Epoch 4/200\n",
      "60000/60000 [==============================] - 21s 350us/step - loss: 0.1415 - acc: 0.9578 - val_loss: 0.0764 - val_acc: 0.9764\n",
      "Epoch 5/200\n",
      "60000/60000 [==============================] - 21s 351us/step - loss: 0.1217 - acc: 0.9629 - val_loss: 0.0779 - val_acc: 0.9766\n",
      "Epoch 6/200\n",
      "60000/60000 [==============================] - 21s 351us/step - loss: 0.1088 - acc: 0.9673 - val_loss: 0.0668 - val_acc: 0.9807\n",
      "Epoch 7/200\n",
      "60000/60000 [==============================] - 21s 350us/step - loss: 0.1003 - acc: 0.9702 - val_loss: 0.0614 - val_acc: 0.9816\n",
      "Epoch 8/200\n",
      "60000/60000 [==============================] - 21s 351us/step - loss: 0.0900 - acc: 0.9726 - val_loss: 0.0616 - val_acc: 0.9800\n",
      "Epoch 9/200\n",
      "60000/60000 [==============================] - 21s 352us/step - loss: 0.0852 - acc: 0.9739 - val_loss: 0.0567 - val_acc: 0.9819\n",
      "Epoch 10/200\n",
      "60000/60000 [==============================] - 21s 352us/step - loss: 0.0800 - acc: 0.9754 - val_loss: 0.0567 - val_acc: 0.9826\n",
      "Epoch 11/200\n",
      "60000/60000 [==============================] - 21s 353us/step - loss: 0.0765 - acc: 0.9762 - val_loss: 0.0513 - val_acc: 0.9843\n",
      "Epoch 12/200\n",
      "60000/60000 [==============================] - 21s 352us/step - loss: 0.0700 - acc: 0.9785 - val_loss: 0.0539 - val_acc: 0.9829\n",
      "Epoch 13/200\n",
      "60000/60000 [==============================] - 21s 353us/step - loss: 0.0681 - acc: 0.9790 - val_loss: 0.0502 - val_acc: 0.9843\n",
      "Epoch 14/200\n",
      "60000/60000 [==============================] - 21s 352us/step - loss: 0.0660 - acc: 0.9792 - val_loss: 0.0511 - val_acc: 0.9837\n",
      "Epoch 15/200\n",
      "60000/60000 [==============================] - 21s 352us/step - loss: 0.0624 - acc: 0.9799 - val_loss: 0.0474 - val_acc: 0.9858\n",
      "Epoch 16/200\n",
      "60000/60000 [==============================] - 21s 354us/step - loss: 0.0580 - acc: 0.9817 - val_loss: 0.0484 - val_acc: 0.9850\n",
      "Epoch 17/200\n",
      "60000/60000 [==============================] - 21s 352us/step - loss: 0.0592 - acc: 0.9814 - val_loss: 0.0491 - val_acc: 0.9845\n",
      "Model Error: 1.55%\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model = stacked_1Dcnn(10,784)\n",
    "model.summary()\n",
    "# print model structure to png file\n",
    "plot_model(model, to_file='model_stacked_1Dcnn.png', show_shapes = True)\n",
    "# complile the model\n",
    "model.compile(loss = 'categorical_crossentropy', optimizer = 'adam', metrics = ['accuracy'])\n",
    "# Fit the model\n",
    "early_stopping = EarlyStopping(monitor = 'val_loss', patience=2)\n",
    "model.fit(X_train_1D, Y_train, validation_data = (X_test_1D, Y_test), epochs=200, batch_size=200, callbacks=[early_stopping])\n",
    "# Final evaluation of the model\n",
    "scores = model.evaluate(X_test_1D, Y_test, verbose=0)\n",
    "print(\"Model Error: %.2f%%\" % (100-scores[1]*100))\n",
    "# plotting the confusion matrix\n",
    "plot_confusion_matrix(y_test, model.predict(X_test_1D).argmax(axis= 1), classes=[0,1,2,3,4,5,6,7,8,9])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Nildip.mukherjee\\AnacondaProjects\\Others\\MNIST\\models.py:104: UserWarning: Update your `Conv2D` call to the Keras 2 API: `Conv2D(kernel_size=15, strides=1, activation=\"relu\", filters=5)`\n",
      "  cnn_layer = Conv2D(nb_filter = 5, kernel_size = 15, strides = 1, activation = 'relu')(input_shape)\n",
      "C:\\Users\\Nildip.mukherjee\\AnacondaProjects\\Others\\MNIST\\models.py:107: UserWarning: Update your `Conv2D` call to the Keras 2 API: `Conv2D(kernel_size=10, strides=1, activation=\"relu\", filters=4)`\n",
      "  cnn_layer = Conv2D(nb_filter = 4, kernel_size = 10, strides = 1, activation = 'relu')(input_shape)\n",
      "C:\\Users\\Nildip.mukherjee\\AnacondaProjects\\Others\\MNIST\\models.py:110: UserWarning: Update your `Conv2D` call to the Keras 2 API: `Conv2D(kernel_size=5, strides=1, activation=\"relu\", filters=4)`\n",
      "  cnn_layer = Conv2D(nb_filter = 4, kernel_size = 5, strides = 1, activation = 'relu')(input_shape)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_5 (InputLayer)         (None, 28, 28, 1)         0         \n",
      "_________________________________________________________________\n",
      "conv2d_4 (Conv2D)            (None, 24, 24, 4)         104       \n",
      "_________________________________________________________________\n",
      "dropout_17 (Dropout)         (None, 24, 24, 4)         0         \n",
      "_________________________________________________________________\n",
      "max_pooling2d_4 (MaxPooling2 (None, 12, 12, 4)         0         \n",
      "_________________________________________________________________\n",
      "flatten_4 (Flatten)          (None, 576)               0         \n",
      "_________________________________________________________________\n",
      "dense_14 (Dense)             (None, 500)               288500    \n",
      "_________________________________________________________________\n",
      "dropout_18 (Dropout)         (None, 500)               0         \n",
      "_________________________________________________________________\n",
      "dense_15 (Dense)             (None, 100)               50100     \n",
      "_________________________________________________________________\n",
      "dropout_19 (Dropout)         (None, 100)               0         \n",
      "_________________________________________________________________\n",
      "dense_16 (Dense)             (None, 10)                1010      \n",
      "=================================================================\n",
      "Total params: 339,714\n",
      "Trainable params: 339,714\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/200\n",
      "60000/60000 [==============================] - 22s 370us/step - loss: 0.4090 - acc: 0.8720 - val_loss: 0.1187 - val_acc: 0.9648\n",
      "Epoch 2/200\n",
      "60000/60000 [==============================] - 22s 366us/step - loss: 0.1466 - acc: 0.9564 - val_loss: 0.0752 - val_acc: 0.9793\n",
      "Epoch 3/200\n",
      "60000/60000 [==============================] - 22s 366us/step - loss: 0.1076 - acc: 0.9677 - val_loss: 0.0618 - val_acc: 0.9814\n",
      "Epoch 4/200\n",
      "60000/60000 [==============================] - 22s 371us/step - loss: 0.0907 - acc: 0.9730 - val_loss: 0.0539 - val_acc: 0.9842\n",
      "Epoch 5/200\n",
      "60000/60000 [==============================] - 22s 367us/step - loss: 0.0794 - acc: 0.9767 - val_loss: 0.0492 - val_acc: 0.9860\n",
      "Epoch 6/200\n",
      "60000/60000 [==============================] - 22s 367us/step - loss: 0.0685 - acc: 0.9795 - val_loss: 0.0452 - val_acc: 0.9869\n",
      "Epoch 7/200\n",
      "60000/60000 [==============================] - 22s 368us/step - loss: 0.0617 - acc: 0.9815 - val_loss: 0.0427 - val_acc: 0.9872\n",
      "Epoch 8/200\n",
      "60000/60000 [==============================] - 22s 371us/step - loss: 0.0585 - acc: 0.9823 - val_loss: 0.0403 - val_acc: 0.9875\n",
      "Epoch 9/200\n",
      "60000/60000 [==============================] - 22s 367us/step - loss: 0.0512 - acc: 0.9844 - val_loss: 0.0419 - val_acc: 0.9873\n",
      "Epoch 10/200\n",
      "60000/60000 [==============================] - 22s 368us/step - loss: 0.0475 - acc: 0.9858 - val_loss: 0.0369 - val_acc: 0.9891\n",
      "Epoch 11/200\n",
      "60000/60000 [==============================] - 22s 367us/step - loss: 0.0446 - acc: 0.9866 - val_loss: 0.0372 - val_acc: 0.9881\n",
      "Epoch 12/200\n",
      "60000/60000 [==============================] - 22s 366us/step - loss: 0.0441 - acc: 0.9868 - val_loss: 0.0375 - val_acc: 0.9887\n",
      "Model Error: 1.13%\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model = stacked_2Dcnn(10,28,28)\n",
    "model.summary()\n",
    "# print model structure to png file\n",
    "plot_model(model, to_file='model_stacked_2Dcnn.png', show_shapes = True)\n",
    "# complile the model\n",
    "model.compile(loss = 'categorical_crossentropy', optimizer = 'adam', metrics = ['accuracy'])\n",
    "# Fit the model\n",
    "early_stopping = EarlyStopping(monitor = 'val_loss', patience=2)\n",
    "model.fit(X_train_2D, Y_train, validation_data = (X_test_2D, Y_test), epochs=200, batch_size=200, callbacks=[early_stopping])\n",
    "# Final evaluation of the model\n",
    "scores = model.evaluate(X_test_2D, Y_test, verbose=0)\n",
    "print(\"Model Error: %.2f%%\" % (100-scores[1]*100))\n",
    "# plotting the confusion matrix\n",
    "plot_confusion_matrix(y_test, model.predict(X_test_2D).argmax(axis= 1), classes=[0,1,2,3,4,5,6,7,8,9])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Nildip.mukherjee\\AnacondaProjects\\Others\\MNIST\\models.py:133: UserWarning: Update your `Conv2D` call to the Keras 2 API: `Conv2D(kernel_size=15, strides=1, activation=\"relu\", filters=4)`\n",
      "  tower1 = Conv2D(nb_filter = 4, kernel_size = 15, strides = 1, activation = 'relu')(input_shape)\n",
      "C:\\Users\\Nildip.mukherjee\\AnacondaProjects\\Others\\MNIST\\models.py:137: UserWarning: Update your `Conv2D` call to the Keras 2 API: `Conv2D(kernel_size=15, strides=1, activation=\"relu\", filters=4)`\n",
      "  tower2 = Conv2D(nb_filter = 4, kernel_size = 15, strides = 1, activation = 'relu')(input_shape)\n",
      "C:\\Users\\Nildip.mukherjee\\AnacondaProjects\\Others\\MNIST\\models.py:141: UserWarning: Update your `Conv2D` call to the Keras 2 API: `Conv2D(kernel_size=15, strides=1, activation=\"relu\", filters=4)`\n",
      "  tower3 = Conv2D(nb_filter = 4, kernel_size = 15, strides = 1, activation = 'relu')(input_shape)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_6 (InputLayer)            (None, 28, 28, 1)    0                                            \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_5 (Conv2D)               (None, 14, 14, 4)    904         input_6[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_6 (Conv2D)               (None, 14, 14, 4)    904         input_6[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_7 (Conv2D)               (None, 14, 14, 4)    904         input_6[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "dropout_20 (Dropout)            (None, 14, 14, 4)    0           conv2d_5[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "dropout_21 (Dropout)            (None, 14, 14, 4)    0           conv2d_6[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "dropout_22 (Dropout)            (None, 14, 14, 4)    0           conv2d_7[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "flatten_5 (Flatten)             (None, 784)          0           dropout_20[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "flatten_6 (Flatten)             (None, 784)          0           dropout_21[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "flatten_7 (Flatten)             (None, 784)          0           dropout_22[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_1 (Concatenate)     (None, 2352)         0           flatten_5[0][0]                  \n",
      "                                                                 flatten_6[0][0]                  \n",
      "                                                                 flatten_7[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "dense_17 (Dense)                (None, 500)          1176500     concatenate_1[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "dropout_23 (Dropout)            (None, 500)          0           dense_17[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "dense_18 (Dense)                (None, 100)          50100       dropout_23[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "dropout_24 (Dropout)            (None, 100)          0           dense_18[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "dense_19 (Dense)                (None, 10)           1010        dropout_24[0][0]                 \n",
      "==================================================================================================\n",
      "Total params: 1,230,322\n",
      "Trainable params: 1,230,322\n",
      "Non-trainable params: 0\n",
      "__________________________________________________________________________________________________\n",
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/200\n",
      "60000/60000 [==============================] - 86s 1ms/step - loss: 0.4200 - acc: 0.8707 - val_loss: 0.1020 - val_acc: 0.9699\n",
      "Epoch 2/200\n",
      "60000/60000 [==============================] - 86s 1ms/step - loss: 0.1609 - acc: 0.9538 - val_loss: 0.0722 - val_acc: 0.9777\n",
      "Epoch 3/200\n",
      "60000/60000 [==============================] - 90s 1ms/step - loss: 0.1233 - acc: 0.9639 - val_loss: 0.0649 - val_acc: 0.9799\n",
      "Epoch 4/200\n",
      "60000/60000 [==============================] - 89s 1ms/step - loss: 0.1032 - acc: 0.9708 - val_loss: 0.0576 - val_acc: 0.9810\n",
      "Epoch 5/200\n",
      "60000/60000 [==============================] - 86s 1ms/step - loss: 0.0872 - acc: 0.9742 - val_loss: 0.0539 - val_acc: 0.9824\n",
      "Epoch 6/200\n",
      "60000/60000 [==============================] - 86s 1ms/step - loss: 0.0777 - acc: 0.9771 - val_loss: 0.0461 - val_acc: 0.9862\n",
      "Epoch 7/200\n",
      "60000/60000 [==============================] - 86s 1ms/step - loss: 0.0700 - acc: 0.9790 - val_loss: 0.0479 - val_acc: 0.9846\n",
      "Epoch 8/200\n",
      "60000/60000 [==============================] - 86s 1ms/step - loss: 0.0677 - acc: 0.9804 - val_loss: 0.0454 - val_acc: 0.9866\n",
      "Epoch 9/200\n",
      "60000/60000 [==============================] - 86s 1ms/step - loss: 0.0602 - acc: 0.9820 - val_loss: 0.0419 - val_acc: 0.9871\n",
      "Epoch 10/200\n",
      "60000/60000 [==============================] - 86s 1ms/step - loss: 0.0580 - acc: 0.9826 - val_loss: 0.0420 - val_acc: 0.9881\n",
      "Epoch 11/200\n",
      "60000/60000 [==============================] - 87s 1ms/step - loss: 0.0534 - acc: 0.9841 - val_loss: 0.0434 - val_acc: 0.9875\n",
      "Model Error: 1.25%\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model = multitower_2Dcnn(10,28,28)\n",
    "model.summary()\n",
    "# print model structure to png file\n",
    "plot_model(model, to_file='model_multitower_2Dcnn.png', show_shapes = True)\n",
    "# complile the model\n",
    "model.compile(loss = 'categorical_crossentropy', optimizer = 'adam', metrics = ['accuracy'])\n",
    "# Fit the model\n",
    "early_stopping = EarlyStopping(monitor = 'val_loss', patience=2)\n",
    "model.fit(X_train_2D, Y_train, validation_data = (X_test_2D, Y_test), epochs=200, batch_size=200, callbacks=[early_stopping])\n",
    "# Final evaluation of the model\n",
    "scores = model.evaluate(X_test_2D, Y_test, verbose=0)\n",
    "print(\"Model Error: %.2f%%\" % (100-scores[1]*100))\n",
    "# plotting the confusion matrix\n",
    "plot_confusion_matrix(y_test, model.predict(X_test_2D).argmax(axis= 1), classes=[0,1,2,3,4,5,6,7,8,9])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA5oAAADvCAYAAACT8qtkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAHnpJREFUeJzt3X2UZHV95/H3VxodHgWcURRcRnQGI/ExfXzCTVhAHQgBd3U9o2JQTFjwiA/HPOD6RJRdlXiS3axGJUKGRYJGxJVjIIGDT+tGiC2CQkAYEWRkgEYiMIDIw3f/uLel6Kma7r71q+rf9Lxf59SZ6rr3d++3btWne7733roVmYkkSZIkSaU8ZrELkCRJkiQtLTaakiRJkqSibDQlSZIkSUXZaEqSJEmSirLRlCRJkiQVZaMpSZIkSSrKRnMRRcQBEXFdRGyKiFf1mb5fRHw/Iu6OiLdHxKcj4v3ttAMjYsP4q56f3lqlrY3ZlOpkNqU6LfFsXhARRy92HVulzNxqbsCbgB8C9wK3AJ8CdlvA+BuAQwrWM9TygIuBd2xh+mnAXw6YdiCwYVTPrV3m24Ap4H5g3WK//lt4T3x7RMveA/gycA9wI/D6xX6+td7M5qOmjTSbwOPa9d8I3A18Hzh0sd8DA94TZrOO18Fs5tj+bn4O2AjcBVwL/MFivwcGvCfMZh2vg9nM8WSzZ9mrgF8Cn1vs90Cf2k4aVV3ASuDr7fvtmlFt39m3reaIZkS8G/gY8MfA44EXA/sAF0XEYxeztiHsA1w1xPQiotHvvXAzcDJw+qhrqNQngV8BTwLeAHwqIvZf3JLqYzZHZ0A2J4CbgN+h2d7vB/4+IlaOup6KmM15MJujs4W/mx8BVmbmrsARwMkR8VujrqciZnMezObobCGbMz4JfHfUdVTobJod008A3gucExErRr7Wxe7e59mF7wpsAl476/GdgduAY9qf1wEn90w/kHYPCXAm8DBwX7usP6Hp7hM4lqap2gi8u2f8gpY3oPY/BNYDdwDnAU9pH//xrPGPmzXua8BDNHtdNgGre+uZTy00v7j+GfgFcAVwYM/yvwH8N+D/teOesYXtfzILPKLZr1bg3e3rtRF486x5Pw1cRHOE5pvAPj17YBKYmFX7HwC/0W6fh9rn/YuC77mdaP5Yru557Ezgo4udh5pumM1FzWbPmB8Ar57na7ZZrZjNJXfDbC56NoH92u3z2i3N12/bYTaX7A2zuWjZBNYCf88CjxzO5Ke9/ybg28DHgX8DfkLPWUXtvB8B/gW4E/gKsMfs59kz/w3AIcCaNj8PtM/7ioLvudU0Zyfu0vPY/wWOG/X7fWs5ovlSYBlwbu+DmbkJuAB4+VwLyMw3Aj8Ffi8zd87MU3om/weaQ+mvAE6MiEOGXB4AEXEQzZvttcCTaU4j+Xw7/umzxt8/a/kH0bwJ3tZOv3YhtUTEXsA/0DSJewB/BHxp1t6LN9L8QtqlrW2U9qTZa7cX8BbgkxGxe8/0NwAfBpYDlwNnzbXAzLwaOA74Tvu8d+s3X0T8dUT8YsDtBwMWvxp4aNZ2vwJwz+yjmc1FzmZEPInm/dp1T7HZXJrM5iJls31fz5yethE4f45NM4jZXJrM5iJkMyJ2BT5Es/NmWC8CfkSTvVOA0yIieqb/PnAM8BTgQeCv5lpgZv4j8N+BL7TP+7n95ouIr24hm18dsPj9gesz8+6ex8aSza2l0VwO3J6ZD/aZtrGdPow/y8x7MvOHwN8CrxtyeTPeAJyemZe1oXsP8JIxneJ2FHB+Zp6fmQ9n5kU0n7c8rGeedZl5VWY+mJkPjLieB4APZeYDmXk+zd6a/Xqm/0NmfqvdTu+l2U5PLbHizHxrZu424PacAcN2ptkT1etOml9geoTZXLhi2YyI7Wn+c3lGZl7TsR6zuTSZzYUrks3MfCvN+/Hf0zQT9/ebbx7M5tJkNheuRDY/DJyWmTcVqOfGzPybzHwIOIOm8X5Sz/QzM/PKzLyH5uMtr42I7Qqsl8w8fAvZPHzAsEXL5tbSaN4OLI+IiT7TntxOH0bvm+5Gmj0QJTyFnr0q7d6qn9PsnRy1fYD/3LunA3gZzfaaUSJs8/XzWb9U76V5429WS7ud7qDc69DFJprTW3rtSnOKkh5hNheuSDbbz6CcSXOqzduGqMdsLk1mc+GK/d3MzIcy89vA3sDxHesxm0uT2Vy4obIZEc+jOT31LwvVc8vMncy8t73bN5s022x7ht+BMIxFy+bW0mh+h2aP4H/qfTAidgIOpbnSFTRXOduxZ5Y9Zy0nByy/dw/gv6M5t32Y5c24mSYcvfU+AfjZHOO6mF3LTTR7VHr3dOyUmR/dwpjF9OvXICJ2pjk14maa1wAGvw5zPof2EtqbBtwGnW54LTAREat6HnsuY/gg+1bGbM6teDbbU3ROo9mD+uoRn5FgNrdOZnNu4/i7OQE8fagqBzObWyezObfS2TyQ5jOsP42IW2hOvX11RFxWsuges1+DB2h2IDzqNWiPcvae/jufbF6whWxeMGDYVcC+EdF7BHMs2dwqGs3MvBP4M+B/RcSaiNi+PVT/RZoPy5/Zzno5cFhE7BERewLvnLWoW4F9+6zi/RGxYzRXRnsz8IUhlzfj74A3R8TzIuJxNOdeX5qZN8z5pBdudi2fA34vIl4ZEdtFxLJovqdo7/kuMCImImIZsB0ws4yJnukZEQcWqv+wiHhZNFdb+zDNdropM6dpfokd1T6PY3j0H+1bgb1jC1dpy8zjsjnfvd+t7/np7ekO5wIfioidIuIA4Egeea8JszlPxbNJcxn836D5DMt9syeaTZnNeSmazYh4YkSsjYid2/GvpDlt8Ws985jNbZzZnJfSfzdPpcnA89rbp2k+8/lKgIhY2WZzZaH6j4qIZ0XEjjSfCz2nPc32WmBZRPxuNB99eR/NV5bNuBVYGVu4am5mHrqFbB46YMy1NK//B9tt9x+B5wBfKvN0B9sqGk2AbD6Y/F9prvJ0F3ApzR6Og/ORDx2fSfPh1huAC3kkXDM+ArwvmsPuf9Tz+DdprqJ1MfDxzLxwyOXN1HwxzbnZX6I57/7pNFe8GoVH1dKeg34kzTabptlWf8zCXvP30Vy960Sa8+Pvax+jDfcmmu+AKuHvgA/SnPrzWzSfBZjxhzS1/5zmg8v/3DPtazR7ZG6JiGFPN5ntrcAONFeBOxs4PjPdMzuL2ZxT0WxGxD7Af6H5Y3lLPLIn8w3tdLMpwGzOQ+m/m0lzmuwGmqtRfhx4Z2Z+BcymHmE251Q0m5l5b2beMnOjyeEv250y0ByBvJFyR2fPpLmq7i00F356e1vHnTQZ+Wy7rntofl/M+GL778+j/NHWtcAkze+mjwKv6Xn+IxOZNZ09OV7tnoufANtn/w9la4CIOArYPzPfU2BZ62gu9/y+oQvTkmA2uzObGiWz2Z3Z1CiZze4i4n3AdGZ+psCyvkHz1SmfHbqwJaDfB5GlOWXm5xa7BkmbM5tSncymVKfMPHmxa1iqtppTZyVJkiRJW4dt+tRZSZIkSVJ5HtGUJEmSJBU11s9oLl++PFeuXDnOVUrVuOGGG7j99ttjsevox2xqW2Y2pTqZTalO883mWBvNlStXMjU1Nc5VStWYnJxc7BIGMpvalplNqU5mU6rTfLPpqbOSJEmSpKJsNCVJkiRJRQ3VaEbEmoj4UUSsj4gTSxUlaThmU6qT2ZTqZDal8jo3mhGxHfBJ4FDgWcDrIuJZpQqT1I3ZlOpkNqU6mU1pNIY5ovlCYH1mXp+ZvwI+DxxZpixJQzCbUp3MplQnsymNwDCN5l7ATT0/b2gfe5SIODYipiJianp6eojVSZonsynVyWxKdTKb0ggM02j2++6U3OyBzFMzczIzJ1esWDHE6iTNk9mU6mQ2pTqZTWkEhmk0NwBP7fl5b+Dm4cqRVIDZlOpkNqU6mU1pBIZpNL8LrIqIp0XEY4G1wHllypI0BLMp1clsSnUym9IITHQdmJkPRsTbgH8CtgNOz8yrilUmqROzKdXJbEp1MpvSaHRuNAEy83zg/EK1SCrEbEp1MptSncymVN4wp85KkiRJkrQZG01JkiRJUlE2mpIkSZKkomw0JUmSJElF2WhKkiRJkoqy0ZQkSZIkFWWjKUmSJEkqykZTkiRJklSUjaYkSZIkqSgbTUmSJElSUTaakiRJkqSibDQlSZIkSUXZaEqSJEmSirLRlCRJkiQVZaMpSZIkSSpqYrELkKT5OOmkkzqNW79+fadxH/jABzqNW716dadxkiSpTp/4xCc6jbv++us7jTv++OMXPGbVqlWd1jVKHtGUJEmSJBVloylJkiRJKspGU5IkSZJUVOdGMyKeGhFfj4irI+KqiHhHycIkdWM2pTqZTalOZlMajWEuBvQg8O7MvCwidgG+FxEXZea/FqpNUjdmU6qT2ZTqZDalEeh8RDMzN2bmZe39u4Grgb1KFSapG7Mp1clsSnUym9JoFPmMZkSsBJ4PXNpn2rERMRURU9PT0yVWJ2mezKZUJ7Mp1clsSuUM3WhGxM7Al4B3ZuZds6dn5qmZOZmZkytWrBh2dZLmyWxKdTKbUp3MplTWUI1mRGxPE8izMvPcMiVJGpbZlOpkNqU6mU2pvGGuOhvAacDVmfkX5UqSNAyzKdXJbEp1MpvSaAxzRPMA4I3AQRFxeXs7rFBdkrozm1KdzKZUJ7MpjUDnrzfJzG8DUbAWSQWYTalOZlOqk9mURmOY79GUpE5OPvnkBY855ZRTOq3rvvvu6zTukEMO6TRu9erVncZJkqTR+sxnPtNp3AknnNBp3DOf+cxO44444ogFj1m1alWndY1Ska83kSRJkiRpho2mJEmSJKkoG01JkiRJUlE2mpIkSZKkomw0JUmSJElF2WhKkiRJkoqy0ZQkSZIkFWWjKUmSJEkqykZTkiRJklSUjaYkSZIkqSgbTUmSJElSUTaakiRJkqSibDQlSZIkSUVNLHYBkrZeP/vZzzqNu+aaaxY8Zvvtt++0ruc+97mdxq1du7bTOEmSVKeNGzd2GvfsZz+707iDDjqo07gDDjig07jaeERTkiRJklSUjaYkSZIkqSgbTUmSJElSUUM3mhGxXUR8PyK+WqIgSWWYTalOZlOqk9mUyipxRPMdwNUFliOpLLMp1clsSnUym1JBQzWaEbE38LvAZ8uUI6kEsynVyWxKdTKbUnnDHtH8H8CfAA8PmiEijo2IqYiYmp6eHnJ1kubJbEp1MptSncymVFjnRjMiDgduy8zvbWm+zDw1Myczc3LFihVdVydpnsymVCezKdXJbEqjMcwRzQOAIyLiBuDzwEER8bkiVUkahtmU6mQ2pTqZTWkEOjeamfmezNw7M1cCa4GvZeZRxSqT1InZlOpkNqU6mU1pNPweTUmSJElSURMlFpKZ3wC+UWJZksoxm1KdzKZUJ7MpleMRTUmSJElSUUWOaEraNm3atKnTuNWrVy94zAknnNBpXccff3ynccuWLes0TpIk1emkk07qNO64447rNG7PPffsNG6p8IimJEmSJKkoG01JkiRJUlE2mpIkSZKkomw0JUmSJElF2WhKkiRJkoqy0ZQkSZIkFWWjKUmSJEkqykZTkiRJklSUjaYkSZIkqSgbTUmSJElSUTaakiRJkqSibDQlSZIkSUXZaEqSJEmSippY7AIkbb1WrFjRadzrX//6BY95xjOe0WldkqT+brzxxgWPOffcczuta82aNQse88tf/rLTuqTS9txzz8UuYavkEU1JkiRJUlE2mpIkSZKkomw0JUmSJElFDdVoRsRuEXFORFwTEVdHxEtKFSapO7Mp1clsSnUym1J5w14M6H8C/5iZr4mIxwI7FqhJ0vDMplQnsynVyWxKhXVuNCNiV+C3gTcBZOavgF+VKUtSV2ZTqpPZlOpkNqXRGObU2X2BaeBvI+L7EfHZiNhp9kwRcWxETEXE1PT09BCrkzRPZlOqk9mU6mQ2pREYptGcAF4AfCoznw/cA5w4e6bMPDUzJzNzsut37klaELMp1clsSnUym9IIDNNobgA2ZOal7c/n0IRU0uIym1KdzKZUJ7MpjUDnRjMzbwFuioj92ocOBv61SFWSOjObUp3MplQnsymNxrBXnT0BOKu9Otf1wJuHL0lSAWZTqpPZlOpkNqXChmo0M/NyYLJQLZIKMZtSncymVCezKZU37BFNSUvAww8/zD333LPgcXvssUen9XUdJ0kqZ+XKlQses3z58k7rete73rXgMcuWLeu0Lkl1GOZiQJIkSZIkbcZGU5IkSZJUlI2mJEmSJKkoG01JkiRJUlE2mpIkSZKkomw0JUmSJElF2WhKkiRJkoqy0ZQkSZIkFWWjKUmSJEkqykZTkiRJklSUjaYkSZIkqSgbTUmSJElSUTaakiRJkqSiJha7AEmL7zGPeQw77bTTYpchDW1qamrBY+69994RVCKNz3333ddp3Ite9KIFj7nkkks6rUvStscjmpIkSZKkomw0JUmSJElF2WhKkiRJkooaqtGMiHdFxFURcWVEnB0Ry0oVJqk7synVyWxKdTKbUnmdG82I2At4OzCZmb8JbAesLVWYpG7MplQnsynVyWxKozHsqbMTwA4RMQHsCNw8fEmSCjCbUp3MplQnsykV1rnRzMyfAR8HfgpsBO7MzAtnzxcRx0bEVERMTU9Pd69U0ryYTalOZlOqk9mURmOYU2d3B44EngY8BdgpIo6aPV9mnpqZk5k5uWLFiu6VSpoXsynVyWxKdTKb0mgMc+rsIcBPMnM6Mx8AzgVeWqYsSUMwm1KdzKZUJ7MpjcAwjeZPgRdHxI4REcDBwNVlypI0BLMp1clsSnUym9IIDPMZzUuBc4DLgB+2yzq1UF2SOjKbUp3MplQnsymNxsQwgzPzg8AHC9UiqRCzKdXJbEp1MptSecN+vYkkSZIkSY8y1BFNSZJGYf369Z3GffOb31zwmLvvvrvTuqRa7LDDDp3GXXLJJYUrkaRHeERTkiRJklSUjaYkSZIkqSgbTUmSJElSUTaakiRJkqSibDQlSZIkSUXZaEqSJEmSirLRlCRJkiQVZaMpSZIkSSrKRlOSJEmSVJSNpiRJkiSpKBtNSZIkSVJRNpqSJEmSpKJsNCVJkiRJRU0sdgGSpKXt/vvvX/CYDRs2dFrX/vvvv+AxO+ywQ6d1SZKkwTyiKUmSJEkqykZTkiRJklSUjaYkSZIkqag5G82IOD0ibouIK3se2yMiLoqI69p/dx9tmZJmM5tSncymVCezKY3XfI5orgPWzHrsRODizFwFXNz+LGm81mE2pRqtw2xKNVqH2ZTGZs5GMzO/Bdwx6+EjgTPa+2cArypcl6Q5mE2pTmZTqpPZlMar62c0n5SZGwHaf584aMaIODYipiJianp6uuPqJM2T2ZTqZDalOplNaURGfjGgzDw1Myczc3LFihWjXp2keTKbUp3MplQnsyktTNdG89aIeDJA++9t5UqSNASzKdXJbEp1MpvSiHRtNM8Djm7vHw18pUw5koZkNqU6mU2pTmZTGpH5fL3J2cB3gP0iYkNEvAX4KPDyiLgOeHn7s6QxMptSncymVCezKY3XxFwzZObrBkw6uHAtkhbAbEp1MptSncymNF4jvxiQJEmSJGnbMucRTUmShvHQQw8teMyBBx5YvpABdt1117GtS5KkbYVHNCVJkiRJRdloSpIkSZKKstGUJEmSJBVloylJkiRJKspGU5IkSZJUlI2mJEmSJKkoG01JkiRJUlE2mpIkSZKkomw0JUmSJElF2WhKkiRJkoqy0ZQkSZIkFWWjKUmSJEkqamKxC5AkLW077rjjYpcgSZLGzCOakiRJkqSibDQlSZIkSUXZaEqSJEmSipqz0YyI0yPitoi4suexP4+IayLiBxHx5YjYbbRlSprNbEp1MptSncymNF7zOaK5Dlgz67GLgN/MzOcA1wLvKVyXpLmtw2xKNVqH2ZRqtA6zKY3NnI1mZn4LuGPWYxdm5oPtj5cAe4+gNklbYDalOplNqU5mUxqvEp/RPAa4YNDEiDg2IqYiYmp6errA6iTNk9mU6mQ2pTqZTamgoRrNiHgv8CBw1qB5MvPUzJzMzMkVK1YMszpJ82Q2pTqZTalOZlMqb6LrwIg4GjgcODgzs1xJkoZhNqU6mU2pTmZTGo1OjWZErAH+FPidzLy3bEmSujKbUp3MplQnsymNzny+3uRs4DvAfhGxISLeAnwC2AW4KCIuj4hPj7hOSbOYTalOZlOqk9mUxmvOI5qZ+bo+D582glokLYDZlOpkNqU6mU1pvEpcdVaSJEmSpF/rfDEgSZIkSdLie+ihhzqNO/zwwxc85rrrrpvXfB7RlCRJkiQVZaMpSZIkSSrKRlOSJEmSVJSNpiRJkiSpKBtNSZIkSVJRNpqSJEmSpKJsNCVJkiRJRdloSpIkSZKKstGUJEmSJBVloylJkiRJKspGU5IkSZJUlI2mJEmSJKkoG01JkiRJUlGRmeNbWcQ0cOOAycuB28dWzGC11AH11GIdm+tSyz6ZuWIUxQzLbC5ILXVAPbXUUgeYzcVgHZurpZatvQ6zORzr2FwttWztdcwrm2NtNLckIqYyc9I6HlFLLdaxuZpqGbVanqt1bK6WWmqpA+qqZdRqea7WsblaarGOxVHL87WOzdVSy7ZSh6fOSpIkSZKKstGUJEmSJBVVU6N56mIX0KqlDqinFuvYXE21jFotz9U6NldLLbXUAXXVMmq1PFfr2FwttVjH4qjl+VrH5mqpZZuoo5rPaEqSJEmSloaajmhKkiRJkpYAG01JkiRJUlFjbzQjYk1E/Cgi1kfEiX2mPy4ivtBOvzQiVo6ghqdGxNcj4uqIuCoi3tFnngMj4s6IuLy9faB0He16boiIH7brmOozPSLir9rt8YOIeMGI6tiv57leHhF3RcQ7Z80zkm0SEadHxG0RcWXPY3tExEURcV377+4Dxh7dznNdRBw9olr+PCKuabf/lyNitwFjt/ha1qyGXLbrMZubr2ebz+a2mkswmwNqMZtmc9GZzb61LHo2FzOX7bLNZq/MHNsN2A74MbAv8FjgCuBZs+Z5K/Dp9v5a4AsjqOPJwAva+7sA1/ap40Dgq2PYJjcAy7cw/TDgAiCAFwOXjul1uoXmy1hHvk2A3wZeAFzZ89gpwInt/ROBj/UZtwdwffvv7u393UdQyyuAifb+x/rVMp/XstZbLblsl202536ttrlsbou57Hm9zeYCX1OzaTbH9HqbzQW+puPO5rhz2S7bbPbcxn1E84XA+sy8PjN/BXweOHLWPEcCZ7T3zwEOjogoWURmbszMy9r7dwNXA3uVXEdBRwL/OxuXALtFxJNHvM6DgR9n5o0jXg8Amfkt4I5ZD/e+D84AXtVn6CuBizLzjsz8N+AiYE3pWjLzwsx8sP3xEmDvYdZRoSpyCWZzHrbJbG6juQSz2ZXZNJujZja7GXc2x5pLMJuzjbvR3Au4qefnDWwehl/P026MO4EnjKqg9lSG5wOX9pn8koi4IiIuiIj9R1RCAhdGxPci4tg+0+ezzUpbC5w9YNo4tgnAkzJzIzS/RIEn9plnMbbNMTR74/qZ67WsVXW5BLM5gNnsbynmEszmIGazP7M5Pmazv9qyWUMuYRvO5sQwgzvotydn9verzGeeIiJiZ+BLwDsz865Zky+jOdS+KSIOA/4PsGoEZRyQmTdHxBOBiyLimnYvxK/L7DNmZN9JExGPBY4A3tNn8ri2yXyNe9u8F3gQOGvALHO9lrWqKpdgNvsxmwNWtHRzCWZzELPZndksw2z2V002t7JcwhLN5riPaG4Antrz897AzYPmiYgJ4PFsfgh6aBGxPU0gz8rMc2dPz8y7MnNTe/98YPuIWF66jsy8uf33NuDLNKdj9JrPNivpUOCyzLx19oRxbZPWrTOnU7T/3tZnnrFtm/ZD2YcDb8jMvsGfx2tZq2py2S7fbPZnNmdZ4rkEs9mX2RzIbI6P2eyjsmzWkkvYhrM57kbzu8CqiHhau6dhLXDerHnOA2autPQa4GuDNkRX7TnypwFXZ+ZfDJhnz5lz6SPihTTb6ueF69gpInaZuU/zId0rZ812HvD70XgxcOfM4fcReR0DTjMYxzbp0fs+OBr4Sp95/gl4RUTsHs0VvF7RPlZURKwB/hQ4IjPvHTDPfF7LWlWRSzCbczCbPbaBXILZ7LcOszmY2Rwfs7n5OmrLZi25hG05mznCqz31u9Fccepamqt1vbd97EPtkwZYBnwRWA/8C7DvCGp4Gc3h6B8Al7e3w4DjgOPaed4GXEVzJbFLgJeOoI592+Vf0a5rZnv01hHAJ9vt9UNgcoSvzY40QXt8z2Mj3yY0vwg2Ag/Q7NF5C83nGC4Grmv/3aOddxL4bM/YY9r3ynrgzSOqZT3NefMz75WZq8g9BTh/S6/l1nKrIZftesxm/3q26Wxuq7lsn4PZfHQdZjPNZg03s7lZHdVkc7Fy2S7bbPbcol2oJEmSJElFjPvUWUmSJEnSEmejKUmSJEkqykZTkiRJklSUjaYkSZIkqSgbTUmSJElSUTaakiRJkqSibDQlSZIkSUX9f2AuWbwGgGVkAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA5oAAADvCAYAAACT8qtkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAHqZJREFUeJzt3X2QZXV95/H3V2aEDMiTMyowyKjLUsGYBTPlKpqEFVEgBNg1obAEUYisWq5okRhcUTfqajRWspusG5cVMgTwITwolAHDFEpYd5XYIipkEBB5Ck8NCCOgwMB3/zin9dJze/o+/M7t33S/X1W3uvue8zvne8+9n7nzPefccyMzkSRJkiSplGcsdAGSJEmSpMXFRlOSJEmSVJSNpiRJkiSpKBtNSZIkSVJRNpqSJEmSpKJsNCVJkiRJRdloLqCIeGVE3BgRD0fEUX2m7xMR342In0bEuyLiMxHxgXbagRFxx+SrHkxvrdLWxmxKdTKbUp0WeTYvjYjjF7qOrVJmbjU34M3AD4BHgbuBvwZ2HmL8LcBrCtYz1vKAy4GTtzD9DOAv5ph2IHBHV4+tXeY7gSngMWDdQj//W3hNfKOjZT886/Yk8FcL/ZhrvJnNp03rNJvAtu36bwV+CnwXOHShXwNzvCbMZh3Pg9nMib1vngPcBWwEbgD+YKFfA3O8JsxmHc+D2czJZLNn2XsDPwfOWejXQJ/a/ksXdQHP75PNBE7p+jFtNUc0I+IU4BPAHwE7AS8H9gLWR8QzF7K2MewFXDfG9CKi0e+1cCfwUeDMrmuoUWbuMHMDngv8DDhvgcuqjtnszhzZXAbcDvw2zfb+APB3EbGm63pqYTYHYza7s4X3zY8DazJzR+AI4KMR8Rtd11MLszkYs9mdLWRzxqeBb3ddR00y87ZZ2XwJ8BRwwSRWXv0N2JGm+z561v07APcCJ7R/rwM+2jP9QNo9JMDZ7Ub9Wbus9wJraDr6k2iaqrvo6e6HXd4ctb8VuAl4ALgY2L29/0ezxm87a9zXaPYE/ryd/q976xmkFpp/uP4f8CDwPeDAnuVfAfxX4P+24/7VFrb/RxnyiGa/WoFT2ufrLuAts+b9DLCe5gjNPwJ7tdNmnqNls2r/A+BX2+3zZPu4H+zwNXg8cDMQC52Hmm5mc2Gz2TPm+8DrB3zOzOYSuJnNhc8msE+7fY7e0nz9th1mc9HeMJsLlk3gGODvGPLI4Ux+2t/fDHwD+BTwE+DH9JxV1M77ceCfgIeAi4BdZz/OnvlvAV4DHAI8DjzRPu7vdfga/BDw9Um83reWI5oHANsBF/bemZkPA5cCB8+3gMw8DrgN+N1sOvpP9kz+dzSH0l8LnBoRrxlzeQBExKtpXmxHA7vRnOr2hXb8i2aNf2zW8l8N/B/gne30G4apJSL2AP6epkncFfhD4IKIWNUz9Diaf5Ce1dbWpefR7LXbAzgR+HRE7NIz/Y3AR4CVwDXAufMtMDM3AG8Dvtk+7p37zRcR/zMiHpzj9v0B6z8e+NtsE6pfMJsLnM2IeC7Nm/aoe4rN5uJkNhcom+3r+lHgepr/7F8yz6aZi9lcnMzmAmQzInYEPkyz82Zc/xb4IU32PgmcERHRM/1NwAnA7sAm4C/nW2BmfhX4GPDF9nH/m37zRcRXtpDNrwxY/5uAswacdyxbS6O5ErgvMzf1mXZXO30cf5KZj2TmD4C/Ad4w5vJmvBE4MzOvbkP3PuAVEzrF7Vjgksy8JDOfysz1NJ+3PKxnnnWZeV1mbsrMJzqu5wngw5n5RGZeQrO3Zp+e6X+fmVe22+n9NNtpzxIrzsx3ZObOc9x+fb7xEfF8mtMUJxLKrYzZHF6xbEbEcpr/XJ6VmdePWI/ZXJzM5vCKZDMz30Hzn93fpGkmHus33wDM5uJkNodXIpsfAc7IzNsL1HNrZv7vzHyS5jW+G82p4jPOzsxrM/MRmo+3HB0R2xRYL5l5+Bayefh84yPiN9tazy9Rz3y2lkbzPmBlRCzrM223dvo4el90t9LsgShhd3r2qrR7q+6n2TvZtb2A3+/d0wG8imZ7zSgRtkHdP+sf1UdpThPZrJZ2Oz1AuedhXG+iuXDCjxe6kAqZzeEVyWb7GZSzaU61eecY9ZjNxclsDq/Y+2ZmPpmZ3wBWA28fsR6zuTiZzeGNlc2I2I/m9NS/KFTP3TO/ZOaj7a99s0mzzZYz/g6EUo4HLmifv85tLY3mN2n2CP6H3jsjYnvgUJorXQE8AqzomeV5s5Yz1+kbvXsAn09zbvs4y5txJ004eut9NvAv84wbxexabqfZo9K7p2P7zPzTLYxZSL94DiJiB5pTI+6keQ5g7udh3sfQXkL74Tlug5xuOLFTDLZCZnN+xbPZnqJzBs1eydd3fEaC2dw6mc35TeJ9cxnworGqnJvZ3DqZzfmVzuaBNJ9hvS0i7qY59fb1EXF1yaJ7zH4OnqDZgfC056A9ytl7+u8g2bx0C9m8dJ6xvwL8PhPM5lbRaGbmQ8CfAH8VEYdExPL2UP15NB+WP7ud9RrgsIjYNSKeB7x71qLuAV7YZxUfiIgVEfFi4C3AF8dc3ozPAW+JiP0iYluac6+vysxb5n3Qw5tdyznA70bE6yJim4jYLprvKVo96AIjYllEbAdsA8wsY1nP9IyIAwvVf1hEvCqaq619hGY73Z6Z0zT/iB3bPo4TePqb9j3A6tjCVdoy823Zc7WtWbcXb6moiDiAZm+dV83rw2wOpHg2aS6D/6s0n2H52eyJZlNmcyBFsxkRz4mIYyJih3b862hOW/xazzxmc4kzmwMp/b55Ok0G9mtvn6H5zOfrACJiTZvNNYXqPzYi9o2IFTSfCz2/Pc32BmC7iPidaD76chrNV5bNuAdYE1u4am5mHrqFbB46T13/nuZiSl8f69ENYatoNAGy+WDyf6a5ytNG4CqaPRwH5S8/dHw2zZWobgEu45fhmvFx4LRoDrv/Yc/9/0hzFa3LgU9l5mVjLm+m5stpzs2+gOa8+xfRXPGqC0+rpT0H/UiabTZNs63+iOGe89Nort51Ks358T9r76MN98M03wFVwudoroL1APAbNJ8FmPFWmtrvB15Mc9WxGV+juQjK3REx7ukm/RwPXJiZP+1g2YuC2ZxX0WxGxF7Af6R5s7y7Z0/mG9vpZlOA2RxA6ffNpDlN9g6aq1F+Cnh3Zl4EZlO/ZDbnVTSbmfloZt49c6PJ4c/bnTLQHIG8lXJHZ8+muaru3TQXfnpXW8dDwDuAz7breoTm34sZMztn7o9ujrZO/AJdMcF1Vafdc/FjYHn2/1C25hARxwIvzsz3FVjWOprLPZ82dmFaFMzm6MymumQ2R2c21SWzObqIOA2Yzsz/VWBZV9B8dcpnxy5sEej3QWRpXpl5zkLXIGlzZlOqk9mU6pSZH13oGharrebUWUmSJEnS1mFJnzorSZIkSSrPI5qSJEmSpKIm+hnNlStX5po1aya5Sqkat9xyC/fdd18sdB39mE0tZWZTqpPZlOo0aDYn2miuWbOGqampSa5SqsbatWsXuoQ5mU0tZWZTqpPZlOo0aDY9dVaSJEmSVJSNpiRJkiSpqLEazYg4JCJ+GBE3RcSppYqSNB6zKdXJbEp1MptSeSM3mhGxDfBp4FBgX+ANEbFvqcIkjcZsSnUym1KdzKbUjXGOaL4MuCkzb87Mx4EvAEeWKUvSGMymVCezKdXJbEodGKfR3AO4vefvO9r7niYiToqIqYiYmp6eHmN1kgZkNqU6mU2pTmZT6sA4jWa/707Jze7IPD0z12bm2lWrVo2xOkkDMptSncymVCezKXVgnEbzDmDPnr9XA3eOV46kAsymVCezKdXJbEodGKfR/Dawd0S8ICKeCRwDXFymLEljMJtSncymVCezKXVg2agDM3NTRLwT+AdgG+DMzLyuWGWSRmI2pTqZTalOZlPqxsiNJkBmXgJcUqgWSYWYTalOZlOqk9mUyhvn1FlJkiRJkjZjoylJkiRJKspGU5IkSZJUlI2mJEmSJKkoG01JkiRJUlE2mpIkSZKkomw0JUmSJElF2WhKkiRJkoqy0ZQkSZIkFWWjKUmSJEkqykZTkiRJklSUjaYkSZIkqSgbTUmSJElSUTaakiRJkqSibDQlSZIkSUXZaEqSJEmSirLRlCRJkiQVZaMpSZIkSSrKRlOSJEmSVNTIjWZE7BkRX4+IDRFxXUScXLIwSaMxm1KdzKZUJ7MpdWPZGGM3Aadk5tUR8SzgOxGxPjP/uVBtkkZjNqU6mU2pTmZT6sDIRzQz867MvLr9/afABmCPUoVJGo3ZlOpkNqU6mU2pG0U+oxkRa4D9gav6TDspIqYiYmp6errE6iQNyGxKdTKbUp3MplTO2I1mROwAXAC8OzM3zp6emadn5trMXLtq1apxVydpQGZTqpPZlOpkNqWyxmo0I2I5TSDPzcwLy5QkaVxmU6qT2ZTqZDal8sa56mwAZwAbMvPPy5UkaRxmU6qT2ZTqZDalboxzRPOVwHHAqyPimvZ2WKG6JI3ObEp1MptSncym1IGRv94kM78BRMFaJBVgNqU6mU2pTmZT6sY436MpaZH4yU9+wnnnnTf0uL333nuk9e23334jjZMkSYvL1NTU0GOuuOKKkda1yy67jDTuxBNPHGncUlfk600kSZIkSZphoylJkiRJKspGU5IkSZJUlI2mJEmSJKkoG01JkiRJUlE2mpIkSZKkomw0JUmSJElF2WhKkiRJkoqy0ZQkSZIkFWWjKUmSJEkqykZTkiRJklSUjaYkSZIkqSgbTUmSJElSUcsWugBJC+/mm2/m6KOPHnrce9/73pHWt99++400TpIkDeaee+4ZadxXv/rVkcZdeOGFI427/PLLhx7zyCOPjLSu4447bqRxJ5544kjjljqPaEqSJEmSirLRlCRJkiQVZaMpSZIkSSpq7EYzIraJiO9GxFdKFCSpDLMp1clsSnUym1JZJY5ongxsKLAcSWWZTalOZlOqk9mUChqr0YyI1cDvAJ8tU46kEsymVCezKdXJbErljXtE878B7wWemmuGiDgpIqYiYmp6enrM1Uka0FDZnFxZ0pLn+6ZUJ7MpFTZyoxkRhwP3ZuZ3tjRfZp6emWszc+2qVatGXZ2kAY2SzQmVJi1pvm9KdTKbUjfGOaL5SuCIiLgF+ALw6og4p0hVksZhNqU6mU2pTmZT6sDIjWZmvi8zV2fmGuAY4GuZeWyxyiSNxGxKdTKbUp3MptQNv0dTkiRJklTUshILycwrgCtKLEtSOWZTqpPZlOpkNqVyPKIpSZIkSSqqyBHNWl100UVDj/nc5z430rr233//kcadeuqpI42TStp+++15yUteMvS4U045pYNqJEmqW2byxBNPDD1u06ZNI63v/vvvH3rMddddN9K6br755pHGPfbYYyON23333Yces+eee460rg9+8IMjjdNoPKIpSZIkSSrKRlOSJEmSVJSNpiRJkiSpKBtNSZIkSVJRNpqSJEmSpKJsNCVJkiRJRdloSpIkSZKKstGUJEmSJBVloylJkiRJKspGU5IkSZJUlI2mJEmSJKkoG01JkiRJUlE2mpIkSZKkopYtdAGDOOecc0Ya97GPfWzoMRs2bBhpXdtvv/1I47R1euyxx4Yek5kdVFLGPvvsw5VXXjn0uOXLl3dQjaQZDz74IF/+8peHHnfUUUd1UI2kGREx0nvgqO+bq1evnsgYgH333XekcYceeuhI45566qmhxxxwwAEjrUuT5RFNSZIkSVJRNpqSJEmSpKJsNCVJkiRJRY3VaEbEzhFxfkRcHxEbIuIVpQqTNDqzKdXJbEp1MptSeeNeDOi/A1/NzN+LiGcCKwrUJGl8ZlOqk9mU6mQ2pcJGbjQjYkfgt4A3A2Tm48DjZcqSNCqzKdXJbEp1MptSN8Y5dfaFwDTwNxHx3Yj4bERs9h0fEXFSRExFxNT09PQYq5M0oKGzed99902+SmnpGTqbGzdunHyV0tLj/2mlDozTaC4DXgr8dWbuDzwCnDp7psw8PTPXZubaVatWjbE6SQMaOpsrV66cdI3SUjR0NnfcccdJ1ygtRf6fVurAOI3mHcAdmXlV+/f5NCGVtLDMplQnsynVyWxKHRi50czMu4HbI2Kf9q6DgH8uUpWkkZlNqU5mU6qT2ZS6Me5VZ/8TcG57da6bgbeMX5KkAsymVCezKdXJbEqFjdVoZuY1wNpCtUgqxGxKdTKbUp3MplTeuEc0J2LnnXceadzBBx889Ji3vvWtI63rPe95z0jjtHXadttthx4TER1UUkZEsHz58oUuQ9Ist99+OyeffPLQ44444oiR1veMZ4xz6QZJi8Gee+450XFavHxHkSRJkiQVZaMpSZIkSSrKRlOSJEmSVJSNpiRJkiSpKBtNSZIkSVJRNpqSJEmSpKJsNCVJkiRJRdloSpIkSZKKstGUJEmSJBVloylJkiRJKspGU5IkSZJUlI2mJEmSJKkoG01JkiRJUlHLFrqAQRx++OETHSdJUg0ef/xxbrvttqHHPeMZ7keWJC0s34kkSZIkSUXZaEqSJEmSirLRlCRJkiQVNVajGRHviYjrIuLaiPh8RGxXqjBJozObUp3MplQnsymVN3KjGRF7AO8C1mbmrwHbAMeUKkzSaMymVCezKdXJbErdGPfU2WXAr0TEMmAFcOf4JUkqwGxKdTKbUp3MplTYyI1mZv4L8CngNuAu4KHMvGz2fBFxUkRMRcTU9PT06JVKGojZlOo0SjYnXaO0FPm+KXVjnFNndwGOBF4A7A5sHxHHzp4vM0/PzLWZuXbVqlWjVyppIGZTqtMo2Zx0jdJS5Pum1I1xTp19DfDjzJzOzCeAC4EDypQlaQxmU6qT2ZTqZDalDozTaN4GvDwiVkREAAcBG8qUJWkMZlOqk9mU6mQ2pQ6M8xnNq4DzgauBH7TLOr1QXZJGZDalOplNqU5mU+rGsnEGZ+aHgA8VqkVSIWZTqpPZlOpkNqXyxv16E0mSJEmSnmasI5qSpKXjoYceGmncTjvtVLiSpWP33Xfn7W9/+0KXIUnS0DyiKUmSJEkqykZTkiRJklSUjaYkSZIkqSgbTUmSJElSUTaakiRJkqSibDQlSZIkSUXZaEqSJEmSirLRlCRJkiQVZaMpSZIkSSrKRlOSJEmSVJSNpiRJkiSpKBtNSZIkSVJRNpqSJEmSpKKWLXQBkqStw0477bTQJSw5u+22G6eddtpClyFJ0tA8oilJkiRJKspGU5IkSZJUlI2mJEmSJKmoeRvNiDgzIu6NiGt77ts1ItZHxI3tz126LVPSbGZTqpPZlOpkNqXJGuSI5jrgkFn3nQpcnpl7A5e3f0uarHWYTalG6zCbUo3WYTaliZm30czMK4EHZt19JHBW+/tZwFGF65I0D7Mp1clsSnUym9JkjfoZzedm5l0A7c/nzDVjRJwUEVMRMTU9PT3i6iQNyGxKdTKbUp3MptSRzi8GlJmnZ+bazFy7atWqrlcnaUBmU6qT2ZTqZDal4YzaaN4TEbsBtD/vLVeSpDGYTalOZlOqk9mUOjJqo3kxcHz7+/HARWXKkTQmsynVyWxKdTKbUkcG+XqTzwPfBPaJiDsi4kTgT4GDI+JG4OD2b0kTZDalOplNqU5mU5qsZfPNkJlvmGPSQYVrkTQEsynVyWxKdTKb0mR1fjEgSZIkSdLSMu8RzZKeeuopHn300aHHrVixooNqJEmSJEld8IimJEmSJKkoG01JkiRJUlE2mpIkSZKkomw0JUmSJElF2WhKkiRJkoqy0ZQkSZIkFWWjKUmSJEkqykZTkiRJklSUjaYkSZIkqSgbTUmSJElSUTaakiRJkqSibDQlSZIkSUUtm+TKnnzySTZu3Dj0uBUrVnRQjSRJkiSpCx7RlCRJkiQVZaMpSZIkSSrKRlOSJEmSVNS8jWZEnBkR90bEtT33/VlEXB8R34+IL0XEzt2WKWk2synVyWxKdTKb0mQNckRzHXDIrPvWA7+Wmb8O3AC8r3Bdkua3DrMp1WgdZlOq0TrMpjQx8zaamXkl8MCs+y7LzE3tn98CVndQm6QtMJtSncymVCezKU1Wic9ongBcOtfEiDgpIqYiYur+++8vsDpJAxo4m9PT0xMsS1ryzKZUJ7MpFTRWoxkR7wc2AefONU9mnp6ZazNz7bOf/exxVidpQMNmc9WqVZMrTlrCzKZUJ7Mplbds1IERcTxwOHBQZma5kiSNw2xKdTKbUp3MptSNkRrNiDgE+GPgtzPz0bIlSRqV2ZTqZDalOplNqTuDfL3J54FvAvtExB0RcSLwP4BnAesj4pqI+EzHdUqaxWxKdTKbUp3MpjRZ8x7RzMw39Ln7jA5qkTQEsynVyWxKdTKb0mSVuOqsJEmSJEm/MPLFgEaxzTbbsMMOO0xylZIkSZKkCfOIpiRJkiSpKBtNSZIkSVJRNpqSJEmSpKJsNCVJkiRJRdloSpIkSZKKstGUJEmSJBVloylJkiRJKspGU5IkSZJUlI2mJEmSJKkoG01JkiRJUlE2mpIkSZKkomw0JUmSJElF2WhKkiRJkoqKzJzcyiKmgVvnmLwSuG9ixcytljqgnlqsY3Oj1LJXZq7qophxmc2h1FIH1FNLLXWA2VwI1rG5WmrZ2uswm+Oxjs3VUsvWXsdA2Zxoo7klETGVmWut45dqqcU6NldTLV2r5bFax+ZqqaWWOqCuWrpWy2O1js3VUot1LIxaHq91bK6WWpZKHZ46K0mSJEkqykZTkiRJklRUTY3m6QtdQKuWOqCeWqxjczXV0rVaHqt1bK6WWmqpA+qqpWu1PFbr2FwttVjHwqjl8VrH5mqpZUnUUc1nNCVJkiRJi0NNRzQlSZIkSYuAjaYkSZIkqaiJN5oRcUhE/DAiboqIU/tM3zYivthOvyoi1nRQw54R8fWI2BAR10XEyX3mOTAiHoqIa9rbB0vX0a7nloj4QbuOqT7TIyL+st0e34+Il3ZUxz49j/WaiNgYEe+eNU8n2yQizoyIeyPi2p77do2I9RFxY/tzlznGHt/Oc2NEHN9RLX8WEde32/9LEbHzHGO3+FzWrIZctusxm5uvZ8lnc6nmEszmHLWYTbO54Mxm31oWPJsLmct22WazV2ZO7AZsA/wIeCHwTOB7wL6z5nkH8Jn292OAL3ZQx27AS9vfnwXc0KeOA4GvTGCb3AKs3ML0w4BLgQBeDlw1oefpbpovY+18mwC/BbwUuLbnvk8Cp7a/nwp8os+4XYGb25+7tL/v0kEtrwWWtb9/ol8tgzyXtd5qyWW7bLM5/3O15LK5FHPZ83ybzSGfU7NpNif0fJvNIZ/TSWdz0rlsl202e26TPqL5MuCmzLw5Mx8HvgAcOWueI4Gz2t/PBw6KiChZRGbelZlXt7//FNgA7FFyHQUdCfxtNr4F7BwRu3W8zoOAH2XmrR2vB4DMvBJ4YNbdva+Ds4Cj+gx9HbA+Mx/IzJ8A64FDSteSmZdl5qb2z28Bq8dZR4WqyCWYzQEsyWwu0VyC2RyV2TSbXTObo5l0NieaSzCbs0260dwDuL3n7zvYPAy/mKfdGA8Bz+6qoPZUhv2Bq/pMfkVEfC8iLo2IF3dUQgKXRcR3IuKkPtMH2WalHQN8fo5pk9gmAM/NzLug+UcUeE6feRZi25xAszeun/mey1pVl0swm3Mwm/0txlyC2ZyL2ezPbE6O2eyvtmzWkEtYwtlcNs7gEfTbkzP7+1UGmaeIiNgBuAB4d2ZunDX5appD7Q9HxGHAl4G9OyjjlZl5Z0Q8B1gfEde3eyF+UWafMZ19J01EPBM4Anhfn8mT2iaDmvS2eT+wCTh3jlnmey5rVVUuwWz2YzbnWNHizSWYzbmYzdGZzTLMZn/VZHMryyUs0mxO+ojmHcCePX+vBu6ca56IWAbsxOaHoMcWEctpAnluZl44e3pmbszMh9vfLwGWR8TK0nVk5p3tz3uBL9GcjtFrkG1W0qHA1Zl5z+wJk9omrXtmTqdof97bZ56JbZv2Q9mHA2/MzL7BH+C5rFU1uWyXbzb7M5uzLPJcgtnsy2zOyWxOjtnso7Js1pJLWMLZnHSj+W1g74h4Qbun4Rjg4lnzXAzMXGnp94CvzbUhRtWeI38GsCEz/3yOeZ43cy59RLyMZlvdX7iO7SPiWTO/03xI99pZs10MvCkaLwcemjn83pE3MMdpBpPYJj16XwfHAxf1mecfgNdGxC7RXMHrte19RUXEIcAfA0dk5qNzzDPIc1mrKnIJZnMeZrPHEsglmM1+6zCbczObk2M2N19HbdmsJZewlLOZHV7tqd+N5opTN9Bcrev97X0fbh80wHbAecBNwD8BL+yghlfRHI7+PnBNezsMeBvwtnaedwLX0VxJ7FvAAR3U8cJ2+d9r1zWzPXrrCODT7fb6AbC2w+dmBU3Qduq5r/NtQvMPwV3AEzR7dE6k+RzD5cCN7c9d23nXAp/tGXtC+1q5CXhLR7XcRHPe/MxrZeYqcrsDl2zpudxabjXksl2P2exfz5LO5lLNZfsYzObT6zCbaTZruJnNzeqoJpsLlct22Waz5xbtQiVJkiRJKmLSp85KkiRJkhY5G01JkiRJUlE2mpIkSZKkomw0JUmSJElF2WhKkiRJkoqy0ZQkSZIkFWWjKUmSJEkq6v8DsdOBS3hrT0MAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA5oAAADvCAYAAACT8qtkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAHjtJREFUeJzt3XuUpHWd3/H3Vwa5iOOAtBfAMGpcjsox4naMF7KgCAKOgnHjgQM6irvEeIyaoC5GjdlV11U5u4mJWUOEHQRWXW+RuLBCYL1lBW0RkVkQUAdnlss0sNxBGfjmj+dpLaqrpqurflX9m+7365w6U13P7VvPU5/u+T63isxEkiRJkqRSHrXUBUiSJEmSlhcbTUmSJElSUTaakiRJkqSibDQlSZIkSUXZaEqSJEmSirLRlCRJkiQVZaO5hCLixRFxXUTcExHH9hh+QET8MCLujoi3RcSnIuL97bBDI2LL5KseTGet0o7GbEp1MptSnZZ5Ni+IiPVLXccOKTN3mAfwBuDHwH3AzcCfA2sWMf0m4GUF6xlpfsDFwNu3M/wM4M/6DDsU2DKu99bO863ADPBLYMNSb//tfCa+M6Z5PxO4BLgTuB549VK/31ofZvMRw8aaTWCXdvk3AHcDPwSOWurPQJ/PhNmsYzuYzZzY381zgJuAu4Brgd9b6s9An8+E2axjO5jNnEw2O+b9DOAB4Jyl/gz0qO0/j6su4EXA99r/N1wJHDyJ97TDHNGMiFOAjwLvAh4HvADYH7goIh69lLWNYH9g4wjDi4hGr8/CjcCHgDPHXUNtImIV8FXga8BewMnAORHxW0taWIXM5vj0yeYqYDNwCM36fj/wVxGxdtz11MBsDs5sjs92/m5+BFibmauBVwEfiojfHnc9NTCbgzOb47OdbM75JPD9cddRk4jYCzgP+DiwBvgY8H8iYs+xL3ypu/cBu/DVwD3Aa7te3wPYCpzU/rwB+FDH8ENp95AAZwMPA/e383o3sBZIml+GN9LshTylY/pFza9P7b9Ps1fvdpqNvE/7+k+7pt+la7pLgIdo9rrcA/xWZz2D1ELzi+vvgDuAHwGHdsz/G8CHgf/XTvdPt7P+P8Qij2j2qhU4pd1eNwFv7Br3U8BFNHtavgns3w6b20arumr/PZo9pw+06+ke4I6Cn7kD23lGx2sXAh9c6jzU9MBsLmk2O6a5EnjNgNtsXq2YzWX3wGwueTaBA9r189rtjddr3WE2l+0Ds7lk2QSOA/6KRR45nMtP+/wNwHeA04B/BH5Ox1lF7bgfoTl6eCfNzpe9ut9nx/ibgJcBRwK/Ah5s3/ePCn7m1gEbu167FnjTuD/vO8oRzRcBuwJf7nwxM+8BLgAOX2gGmfk64BfAKzNzj8z8WMfgl9AcSj8CODUiXjbi/ACIiJfSfNheCzyZ5lS3z7XTP71r+l92zf+lwLeBt7bDr11MLRGxL/DXNE3iXsA7gS9FxFTHpK+j+YX02La2cXoSzV67fYE3AZ/s2pNyAvBBYG/gCuDchWaYmVcDbwa+277vNb3Gi4j/ERF39Hlc2Wf20ee1Axeqa4Uxm0uczYh4Is0f7WH3FJvN5clsLlE228/1fcA1NP/ZP3+BVdOP2VyezOYSZDMiVgN/RLPzZlT/AvgJTfY+BpwREZ2f/9cDJwH7ANuATyw0w8z8G+CPgc+37/uf9RovIr62nWx+rc/sg/n5nEg2d5RGc2/g1szc1mPYTe3wUfxhZt6bmT8G/gI4fsT5zTkBODMzL29D9x7ghRM6xe1E4PzMPD8zH87Mi2iutzy6Y5wNmbkxM7dl5oNjrudB4I8y88HMPJ9mb80BHcP/OjO/1a6n99Ksp6eUWHBmviUz1/R5PKfPZNfQ7Fl8V0TsHBFH0JyquHuJmpYRs7l4xbIZETvT/OfyrMy8Zsh6zObyZDYXr0g2M/MtNP/Z/Zc0zcQve403ALO5PJnNxSuRzQ8CZ2Tm5gL13JCZ/yszHwLOomm8n9gx/OzMvCoz76W5vOW1EbFTgeWSmeu2k811fSb7O2CfiDi+zeZ64OlMIJs7SqN5K7B3e/5/tye3w0fR+aG7gWYPRAn70LFXpd1bdRvN3slx2x/41517OoCDadbXnBJhG9RtXb9U76M5TWReLe16up1y22HR2l9SxwKvoLlI/xSa0y2qvSvaEjGbi1ckm+01KGfTnGrz1hHqMZvLk9lcvGJ/NzPzocz8DrAf8G+HrMdsLk9mc/FGymZEPJfm9NQ/K1TPzXNPMvO+9mnPbNKss50ZfQfC0DLzNuAY4D8At9Ccpvt/mUA2d5RG87s0ewT/VeeLEfEY4CiaO10B3Msju/Mndc0n+8y/cw/gP6E5t32U+c25kSYcnfU+HviHBaYbRnctm2n2qHTu6XhMZv7JdqZZSr/eBhGxB82pETfSbAPovx0WfA/R3EL7nj6PvqcbZuaVmXlIZj4+M18OPI3mnHv9htlcWPFstqfonEGzB/U1Yz4jwWzumMzmwibxd3MVzZGDcTCbOyazubDS2TyU5hrWX0TEzTSn3r4mIi4vWXSH7m3wIM0OhEdsg/YoZ+fpv4Nk84LtZPOCftNl5jcz859n5l40pxkfwASyuUM0mpl5J/CHwH+LiCPbw75rgS/QdONnt6NeARwdEXtFxJOAd3TN6haaX3rd3h8Ru0fEs4E3Ap8fcX5z/hJ4Y0Q8NyJ2oTn3+rLM3LTgm1687lrOAV4ZES+PiJ0iYtdovqdov0FnGBGrImJXYCdgbh6rOoZnRBxaqP6jI+LgaO629kGa9bQ5M2dpfomd2L6Pk3jkH+1bgP1iO3dpy8w3Z3O+e6/Hs/tNFxHPad/z7hHxTpo9ZxtKvNnlwmwOpHg2aW6D/0yaa1ju7x5oNmU2B1I0mxHxhIg4LiL2aKd/Oc1pi5d0jGM2VzizOZDSfzdPp8nAc9vHp2iu+Xw5QESsbbO5tlD9J0bEsyJid5rrQr/YnmZ7LbBrRLwimktf3kfzlWVzbgHWxnbumpuZR20nm0f1my4iDmo/a6tpbmS0JTO/XuLNbs8O0WgCZHNh8n+kWTl3AZfR7OE4LH9z0fHZNHei2kRzp7PPd83mI8D7ojns/s6O179Jcxeti4HTMvPCEec3V/PFNOdmf4nmvPun09zxahweUUt7DvoxNOtslmZdvYvFbfP30dy961Sa8+Pvb1+jDfc9NN8BVcJfAh+gOfXnt2muBZjz+zS13wY8m+Zc8zmX0NwE5eaIGPV0k26vo9luW4HDgMOz6wJ3mc0BFM1mROwP/BuaP5Y3x2/2ZJ7QDjebAszmAEr/3Uya02S30NyN8jTgHZn5VTCb+g2zuaCi2czM+zLz5rkHTQ4faHfKQHME8gbKHZ09m2YHy800N356W1vHncBbgE+3y7qXR56++oX239ui/NHWd9McVd1MswPo1YXn31Nk1nT25GS1ey5+DuycvS/KVh8RcSLw7Mx8T4F5baDZs/K+kQvTsmA2h2c2NU5mc3hmU+NkNocXEe8DZjPzfxaY1zdovjrl0yMXtgz0uhBZWlBmnrPUNUiaz2xKdTKbUp0y80NLXcNytcOcOitJkiRJ2jGs6FNnJUmSJEnleURTkiRJklTURK/R3HvvvXPt2rWTXKRUjU2bNnHrrbfGUtfRi9nUSmY2pTqZTalOg2Zzoo3m2rVrmZmZmeQipWpMT08vdQl9mU2tZGZTqpPZlOo0aDY9dVaSJEmSVJSNpiRJkiSpqJEazYg4MiJ+EhHXR8SppYqSNBqzKdXJbEp1MptSeUM3mhGxE/BJ4CjgWcDxEfGsUoVJGo7ZlOpkNqU6mU1pPEY5ovl84PrM/Flm/gr4HHBMmbIkjcBsSnUym1KdzKY0BqM0mvsCmzt+3tK+9ggRcXJEzETEzOzs7AiLkzQgsynVyWxKdTKb0hiM0mj2+u6UnPdC5umZOZ2Z01NTUyMsTtKAzKZUJ7Mp1clsSmMwSqO5BXhKx8/7ATeOVo6kAsymVCezKdXJbEpjMEqj+X3gGRHx1Ih4NHAccF6ZsiSNwGxKdTKbUp3MpjQGq4adMDO3RcRbga8DOwFnZubGYpVJGorZlOpkNqU6mU1pPIZuNAEy83zg/EK1SCrEbEp1MptSncymVN4op85KkiRJkjSPjaYkSZIkqSgbTUmSJElSUTaakiRJkqSibDQlSZIkSUXZaEqSJEmSirLRlCRJkiQVZaMpSZIkSSrKRlOSJEmSVJSNpiRJkiSpKBtNSZIkSVJRNpqSJEmSpKJsNCVJkiRJRdloSpIkSZKKstGUJEmSJBVloylJkiRJKspGU5IkSZJUlI2mJEmSJKkoG01JkiRJUlFDN5oR8ZSI+NuIuDoiNkbE20sWJmk4ZlOqk9mU6mQ2pfFYNcK024BTMvPyiHgs8IOIuCgz/75QbZKGYzalOplNqU5mUxqDoY9oZuZNmXl5+/xu4Gpg31KFSRqO2ZTqZDalOplNaTyKXKMZEWuBg4DLegw7OSJmImJmdna2xOIkDchsSnUym1KdzKZUzsiNZkTsAXwJeEdm3tU9PDNPz8zpzJyempoadXGSBmQ2pTqZTalOZlMqa6RGMyJ2pgnkuZn55TIlSRqV2ZTqZDalOplNqbxR7jobwBnA1Zn5p+VKkjQKsynVyWxKdTKb0niMckTzxcDrgJdGxBXt4+hCdUkantmU6mQ2pTqZTWkMhv56k8z8DhAFa5FUgNmU6mQ2pTqZTWk8RvkeTUkrXHO20eLtsssui57mgQceGGpZkiRJmrwiX28iSZIkSdIcG01JkiRJUlE2mpIkSZKkomw0JUmSJElF2WhKkiRJkoqy0ZQkSZIkFWWjKUmSJEkqykZTkiRJklSUjaYkSZIkqSgbTUmSJElSUTaakiRJkqSibDQlSZIkSUXZaEqSJEmSilq11AVI2nFl5lKXIEmSpAp5RFOSJEmSVJSNpiRJkiSpKBtNSZIkSVJRIzeaEbFTRPwwIr5WoiBJZZhNqU5mU6qT2ZTKKnFE8+3A1QXmI6kssynVyWxKdTKbUkEjNZoRsR/wCuDTZcqRVILZlOpkNqU6mU2pvFGPaP4X4N3Aw/1GiIiTI2ImImZmZ2dHXJykAZlNqU5mU6qT2ZQKG7rRjIh1wNbM/MH2xsvM0zNzOjOnp6amhl2cpAGZTalOZlOqk9mUxmOUI5ovBl4VEZuAzwEvjYhzilQlaRRmU6qT2ZTqZDalMRi60czM92Tmfpm5FjgOuCQzTyxWmaShmE2pTmZTqpPZlMbD79GUJEmSJBW1qsRMMvMbwDdKzEtSOWZTqpPZlOpkNqVyPKIpSZIkSSrKRlOSJEmSVJSNpiRJkiSpKBtNSZIkSVJRNpqSJEmSpKJsNCVJkiRJRdloSpIkSZKKstGUJEmSJBVloylJkiRJKspGU5IkSZJUlI2mJEmSJKkoG01JkiRJUlE2mpIkSZKkomw0JUmSJElF2WhKkiRJkoqy0ZQkSZIkFWWjKUmSJEkqykZTkiRJklSUjaYkSZIkqaiRGs2IWBMRX4yIayLi6oh4YanCJA3PbEp1MptSncymVN6qEaf/r8DfZObvRsSjgd0L1CRpdGZTqpPZlOpkNqXChm40I2I18DvAGwAy81fAr8qUJWlYZlOqk9mU6mQ2pfEY5dTZpwGzwF9ExA8j4tMR8ZjukSLi5IiYiYiZ2dnZERYnaUBmU6qT2ZTqZDalMRil0VwFPA/488w8CLgXOLV7pMw8PTOnM3N6ampqhMVJGpDZlOpkNqU6mU1pDEZpNLcAWzLzsvbnL9KEVNLSMptSncymVCezKY3B0I1mZt4MbI6IA9qXDgP+vkhVkoZmNqU6mU2pTmZTGo9R7zr774Bz27tz/Qx44+glSSrAbEp1MptSncymVNhIjWZmXgFMF6pFUiFmU6qT2ZTqZDal8ka5RlOSJEmSpHlsNCVJkiRJRdloSpIkSZKKstGUJEmSJBVloylJkiRJKspGU5IkSZJUlI2mJEmSJKkoG01JkiRJUlE2mpIkSZKkomw0JUmSJElF2WhKkiRJkoqy0ZQkSZIkFWWjKUmSJEkqykZTkiRJklSUjaYkSZIkqSgbTUmSJElSUTaakiRJkqSibDQlSZIkSUWN1GhGxL+PiI0RcVVEfDYidi1VmKThmU2pTmZTqpPZlMobutGMiH2BtwHTmXkgsBNwXKnCJA3HbEp1MptSncymNB6jnjq7CtgtIlYBuwM3jl6SpALMplQnsynVyWxKhQ3daGbmPwCnAb8AbgLuzMwLu8eLiJMjYiYiZmZnZ4evVNJAzKZUJ7Mp1clsSuMxyqmzewLHAE8F9gEeExEndo+Xmadn5nRmTk9NTQ1fqaSBmE2pTmZTqpPZlMZjlFNnXwb8PDNnM/NB4MvAi8qUJWkEZlOqk9mU6mQ2pTEYpdH8BfCCiNg9IgI4DLi6TFmSRmA2pTqZTalOZlMag1Gu0bwM+CJwOfDjdl6nF6pL0pDMplQnsynVyWxK47FqlIkz8wPABwrVIqkQsynVyWxKdTKbUnmjfr2JJEmSJEmPMNIRTUmSJEn1ufvuu4eabvPmzUNN98xnPnOo6ZrLYrUceURTkiRJklSUjaYkSZIkqSgbTUmSJElSUTaakiRJkqSibDQlSZIkSUXZaEqSJEmSirLRlCRJkiQVZaMpSZIkSSrKRlOSJEmSVJSNpiRJkiSpKBtNSZIkSVJRNpqSJEmSpKJsNCVJkiRJRa1a6gIkSZKkHcm9997L9773vUVP9+1vf3uo5V166aWLnmZmZmaoZa1evXqo6T784Q8PNd26deuGmk7184imJEmSJKkoG01JkiRJUlE2mpIkSZKkohZsNCPizIjYGhFXdby2V0RcFBHXtf/uOd4yJXUzm1KdzKZUJ7MpTdYgRzQ3AEd2vXYqcHFmPgO4uP1Z0mRtwGxKNdqA2ZRqtAGzKU3Mgo1mZn4LuL3r5WOAs9rnZwHHFq5L0gLMplQnsynVyWxKkzXsNZpPzMybANp/n9BvxIg4OSJmImJmdnZ2yMVJGpDZlOpkNqU6DZXNO+64Y2IFSjuqsd8MKDNPz8zpzJyempoa9+IkDchsSnUym1KdOrO5Zs2apS5Hqt6wjeYtEfFkgPbfreVKkjQCsynVyWxKdTKb0pgM22ieB6xvn68HvlqmHEkjMptSncymVCezKY3JIF9v8lngu8ABEbElIt4E/AlweERcBxze/ixpgsymVCezKdXJbEqTtWqhETLz+D6DDitci6RFMJtSncymVCezKU3W2G8GJEmSJElaWRY8olnSbbfdxtlnn73o6Q466KChlnfggQcONZ0kSTV4+OGHuf/++xc93W677TaGauqwfv36hUfq4TOf+cyipznttNOGWtYpp5wy1HTacWzbto2tWxd/36CNGzcOtbzLLrts0dNs3rx5qGW95CUvGWq6devWDTWdli+PaEqSJEmSirLRlCRJkiQVZaMpSZIkSSrKRlOSJEmSVJSNpiRJkiSpKBtNSZIkSVJRNpqSJEmSpKJsNCVJkiRJRdloSpIkSZKKstGUJEmSJBVloylJkiRJKspGU5IkSZJU1KpJLmzTpk28/vWvX/R0Bx100FDL+8QnPrHoaQ4++OChliVJUmmPetSj2G233Za6jKq8+tWvHmq6E044YdHTHHHEEUMtS8vf6tWrOfzwwxc93R133DGGanp76KGHhppu/fr1hSvRSuURTUmSJElSUTaakiRJkqSibDQlSZIkSUUt2GhGxJkRsTUirup47eMRcU1EXBkRX4mINeMtU1I3synVyWxKdTKb0mQNckRzA3Bk12sXAQdm5nOAa4H3FK5L0sI2YDalGm3AbEo12oDZlCZmwUYzM78F3N712oWZua398VJgvzHUJmk7zKZUJ7Mp1clsSpNV4hrNk4AL+g2MiJMjYiYiZgosS9LgBs7m7OzsBMuSVjyzKdXJbEoFjdRoRsR7gW3Auf3GyczTM3M6M6dHWZakwS02m1NTU5MrTlrBzKZUJ7Mplbdq2AkjYj2wDjgsM7NcSZJGYTalOplNqU5mUxqPoRrNiDgS+APgkMy8r2xJkoZlNqU6mU2pTmZTGp9Bvt7ks8B3gQMiYktEvAn478BjgYsi4oqI+NSY65TUxWxKdTKbUp3MpjRZCx7RzMzje7x8xhhqkbQIZlOqk9mU6mQ2pckqcddZSZIkSZJ+beibAQ1jjz32YHp68TeffeUrXznU8g4++OChppMkSXU69thjl7oEiYhgl112WfR0J5544lDLO+SQQxY9za677jrUsryjrkrxiKYkSZIkqSgbTUmSJElSUTaakiRJkqSibDQlSZIkSUXZaEqSJEmSirLRlCRJkiQVZaMpSZIkSSrKRlOSJEmSVJSNpiRJkiSpKBtNSZIkSVJRNpqSJEmSpKJsNCVJkiRJRdloSpIkSZKKisyc3MIiZoEb+gzeG7h1YsX0V0sdUE8t1jHfMLXsn5lT4yhmVGZzUWqpA+qppZY6wGwuBeuYr5ZadvQ6zOZorGO+WmrZ0esYKJsTbTS3JyJmMnPaOn6jllqsY76aahm3Wt6rdcxXSy211AF11TJutbxX65ivllqsY2nU8n6tY75aalkpdXjqrCRJkiSpKBtNSZIkSVJRNTWapy91Aa1a6oB6arGO+WqqZdxqea/WMV8ttdRSB9RVy7jV8l6tY75aarGOpVHL+7WO+WqpZUXUUc01mpIkSZKk5aGmI5qSJEmSpGXARlOSJEmSVNTEG82IODIifhIR10fEqT2G7xIRn2+HXxYRa8dQw1Mi4m8j4uqI2BgRb+8xzqERcWdEXNE+/lPpOtrlbIqIH7fLmOkxPCLiE+36uDIinjemOg7oeK9XRMRdEfGOrnHGsk4i4syI2BoRV3W8tldEXBQR17X/7tln2vXtONdFxPox1fLxiLimXf9fiYg1fabd7rasWQ25bJdjNucvZ8Vnc6XmEsxmn1rMptlccmazZy1Lns2lzGU7b7PZKTMn9gB2An4KPA14NPAj4Fld47wF+FT7/Djg82Oo48nA89rnjwWu7VHHocDXJrBONgF7b2f40cAFQAAvAC6b0Ha6mebLWMe+ToDfAZ4HXNXx2seAU9vnpwIf7THdXsDP2n/3bJ/vOYZajgBWtc8/2quWQbZlrY9actnO22wuvK1WXDZXYi47trfZXOQ2NZtmc0Lb22wucptOOpuTzmU7b7PZ8Zj0Ec3nA9dn5s8y81fA54BjusY5Bjirff5F4LCIiJJFZOZNmXl5+/xu4Gpg35LLKOgY4DPZuBRYExFPHvMyDwN+mpk3jHk5AGTmt4Dbu17u/BycBRzbY9KXAxdl5u2Z+Y/ARcCRpWvJzAszc1v746XAfqMso0JV5BLM5gBWZDZXaC7BbA7LbJrNcTObw5l0NieaSzCb3SbdaO4LbO74eQvzw/DrcdqVcSfw+HEV1J7KcBBwWY/BL4yIH0XEBRHx7DGVkMCFEfGDiDi5x/BB1llpxwGf7TNsEusE4ImZeRM0v0SBJ/QYZynWzUk0e+N6WWhb1qq6XILZ7MNs9rYccwlmsx+z2ZvZnByz2Vtt2awhl7CCs7lqlImH0GtPTvf3qwwyThERsQfwJeAdmXlX1+DLaQ613xMRRwP/G3jGGMp4cWbeGBFPAC6KiGvavRC/LrPHNGP7TpqIeDTwKuA9PQZPap0MatLr5r3ANuDcPqMstC1rVVUuwWz2Yjb7LGj55hLMZj9mc3hmswyz2Vs12dzBcgnLNJuTPqK5BXhKx8/7ATf2GyciVgGPY/4h6JFFxM40gTw3M7/cPTwz78rMe9rn5wM7R8TepevIzBvbf7cCX6E5HaPTIOuspKOAyzPzlu4Bk1onrVvmTqdo/93aY5yJrZv2oux1wAmZ2TP4A2zLWlWTy3b+ZrM3s9llmecSzGZPZrMvszk5ZrOHyrJZSy5hBWdz0o3m94FnRMRT2z0NxwHndY1zHjB3p6XfBS7ptyKG1Z4jfwZwdWb+aZ9xnjR3Ln1EPJ9mXd1WuI7HRMRj557TXKR7Vddo5wGvj8YLgDvnDr+PyfH0Oc1gEuukQ+fnYD3w1R7jfB04IiL2jOYOXke0rxUVEUcCfwC8KjPv6zPOINuyVlXkEszmAsxmhxWQSzCbvZZhNvszm5NjNucvo7Zs1pJLWMnZzDHe7anXg+aOU9fS3K3rve1rf9S+aYBdgS8A1wPfA542hhoOpjkcfSVwRfs4Gngz8OZ2nLcCG2nuJHYp8KIx1PG0dv4/apc1tz466wjgk+36+jEwPcZtsztN0B7X8drY1wnNL4KbgAdp9ui8ieY6houB69p/92rHnQY+3THtSe1n5XrgjWOq5Xqa8+bnPitzd5HbBzh/e9tyR3nUkMt2OWazdz0rOpsrNZftezCbj6zDbKbZrOFhNufVUU02lyqX7bzNZscj2plKkiRJklTEpE+dlSRJkiQtczaakiRJkqSibDQlSZIkSUXZaEqSJEmSirLRlCRJkiQVZaMpSZIkSSrKRlOSJEmSVNT/B3KyZCN9SOjGAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# passing average image for each number to one of the CNN filters and plotting the output\n",
    "visualize_filter(model, X_test, y_test, 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA5oAAADvCAYAAACT8qtkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzt3Xu4HHWd5/H3dziJEUm4JQgCIV4QRXFVjq54G1bQcHFE0fUBL4OKguvjiPs44+BtXa+Ml8fZnRl3HFYUFvEy3lYfB2ZEHXV1AI0YrkFECZDhdogKCRBC9Lt/VB1tTvrknK7+VZ9Kzvv1PP2cPl31q/p2dX9y8q2qro7MRJIkSZKkUv5orguQJEmSJO1YbDQlSZIkSUXZaEqSJEmSirLRlCRJkiQVZaMpSZIkSSrKRlOSJEmSVJSN5hyKiGdExM8jYmNEvLDP9IMi4qcRsSEi3hQRn4iId9XTDo+IdaOvenZ6a5W2N2ZT6iazKXXTDp7NCyLipLmuY7uUmdvNDXgVcAVwD3Ar8PfAbgOMXwscWbCeoZYHfBs4bRvTzwL+eppphwPr2npu9TLfCKwC7gPOnuvXfxvviR+0uPwTgDXA3cAvgGfN9XPu4s1sPmBaq9kEHlSv/wZgA/BT4Oi5fg9M854wm914Hcxmjuzv5meAW4C7gGuB1871e2Ca94TZ7MbrYDZzNNnsWfaBwCbgM3P9HuhT239vsy7gNOD6OptrgEe3/Zy2myOaEfEW4EPAXwC7Ak8DDgAujIiFc1nbEA4ArhpiehFR6fdeuBl4P/Cptmvoooh4LtV77tXAYuDZwC/ntKgOMpvtmSabY8BNwB9Tbe93Af8YESvarqcrzObsmM32bOPv5hnAisxcArwAeH9EHNp2PV1hNmfHbLZnG9mc9HHgx23X0TUR8VrgZOBYYBfg+cAdra94rrv3WXbgS4CNwEunPL4LcDvwmvr3s4H390w/nHoPCXAu8Dvg3npZbwVWAAmcQtVU3QK8pWf8QMubpvbXAdcBvwK+DjysfvwXU8Y/aMq47wC/pdrrshF4dG89s6mF6h+ufwN+A1wGHN6z/O8CHwB+WI971Da2//sZ8Ihmv1qBt9Sv1y3Aq6fM+wngQqojNN8DDqinTb5GY1Nqfy3w2Hr7/LZ+3r8p/L77N+DkuX7/d/lmNuc2mz1jLgdePMvXzGzOg5vZnPtsAgfV2+el25qv37YzmzvuzWzOXTapjrb/IwMeOZzMT33/VcAPgI8Cv6Y6Qnj0lHnPAH4E3Al8Ddhj6vPsmX8tcCRwFLAZuL9+3pcVfM/9EdUO6iNG/X7fXo5oPh1YBHyl98HM3AhcADx3pgVk5iuBG4E/ycxdMvPDPZP/E9Wh9OcBp0fEkUMuD4CIeA7Vm+2lwD5Up7p9vh7/yCnj75uy/OcA/w94Yz392kFqiYh9gX+iahL3AP4c+HJELOsZ+kqqf5AW17W1aW+qvXb7Uu1R+XhE7N4z/eXA+4ClwGrgvJkWmJlrgNcDF9XPe7d+80XE/4qI30xzu3yaMTsB48CyiLguItZFxN9FxIMHeM7zgdmc42xGxEOp/mg33VNsNndMZnOOslm/r+8BrqH6z/75M2ya6ZjNHZPZnINsRsQS4L1UO2+G9R+Bn1Fl78PAWRERPdP/FHgN8DBgC/A3My0wM/8Z+CDwhfp5/4d+80XEN7aRzW9Ms/j96tvjI+KmiLg+It4zw5HfIraXRnMpcEdmbukz7ZZ6+jDek5l3Z+YVwKeBE4dc3qSXA5/KzEvr0L0NOGxEp7i9Ajg/M8/PzN9l5oVUn7c8pmeeszPzqszckpn3t1zP/cB7M/P+zDyfam/NQT3T/ykzv19vp3dQbaf9S6w4M9+QmbtNc3vCNMMeCiwAXgI8C3gi8CTgnSVq2oGYzcEVy2ZELKD6z+U5mXlNw3rM5o7JbA6uSDYz8w1U/9l9FlUzcV+/+WbBbO6YzObgSmTzfcBZmXlTgXpuyMz/nZm/Bc6harwf2jP93My8MjPvpvp4y0vrHTFDy8znbyObz59m2H71z+cBh1DtjDiRagdWq7aXRvMOYGlEjPWZtg/Dn2Pc+6a7gWoPRAkPo2evSr23aj3V3sm2HQD85949HcAzqbbXpBJhm631U/5RvYfqNJGtaqm3068o9zo0cW/9828z85bMvAP4GA/8R01ms4ki2az3RJ5LdarNG4eox2zumMzm4Ir93czM32bmD6j+g/dfGtZjNndMZnNwQ2UzIp5IdXrqXxeq59bJO5l5T323bzapttkCht+BMIzJbH44M3+TmWuBf2AE2dxeGs2LqPYIHt/7YEQ8BDia6kpXUF1FaeeeWfaespycZvm9ewCXU53bPszyJt1MFY7eevcE/n2GcU1MreUmqj0qvXs6HpKZf7WNMXPp969BROxCdWrEzVSvAUz/Osz4HKK6hPbGaW59TzfMzF9TfT6mS9uoi8zmzIpnsz5F5yyqPagvbvmMBLO5fTKbMxvF380x4JFDVTk9s7l9MpszK53Nw6k+w3pjRNxKdertiyPi0pJF95j6GtxPtQPhAa9BfZSz9/Tf2WTzgm1k84Jphv2Maqf0yLO5XTSamXkn8B7gbyPiqIhYUB+q/yLVP2rn1rOuBo6JiD0iYm/gzVMWdRvwiD6reFdE7BwRj6O6UtoXhlzepM8Cr46IJ0bEg6jOvb6k3pNQ2tRaPgP8SUSsjIidImJRVN9TtN8047cSEWMRsQjYCZhcxljP9IyIwwvVf0xEPDOqq629j2o73ZSZE1T/iL2ifh6v4YF/tG8D9ottXKUtM1+f1fnu/W6P20ZNnwb+LCL2iupzMW8Gpjv/fV4ym7NSPJtUl8F/LNVnWO6dOtFsymzOStFs1u/HEyJil3r8SqrT077TM4/ZnOfM5qyU/rt5JlUGnljfPkH1mc+VABGxos7mikL1vyIiDo6Inak+F/ql+jTba4FFEXFsVB99eSfVV5ZNug1YEdv47GRmHr2NbB49zZh7qN4Hb42IxfV2ex0jyOZ20WgCZPXB5LdTXeXpLuAS/nAFpcnPP5xLdSWqtcA3+UO4Jp0BvDOqw+5/3vP496iuovVt4KOZ+c0hlzdZ87epzs3+MtV594+kuuJVGx5QS30O+nFU22yCalv9BYO95u+kOtx+OtX58ffWj1G/STdSfQdUCZ8F3k116s+hVJ8FmPQ6qtrXA4+juurYpO9QXQTl1ogofZnm91FdAvtaqu8b+inVVc3Uw2zOqGg2I+IA4FSqP5a3xh/2ZL68nm42BZjNWSj9dzOpTpNdR3U1yo8Cb87Mr4HZ1B+YzRkVzWZm3pOZt07eqHK4qd4pA9URyBsod3T2XKqr6t5KdeGnN9V13Am8Afhkva67qf69mPTF+uf6KH+09Y1Uz/tmqqPqn2UEX18YmfP3DId6z8X1wILs/6FsTSMiXgE8LjPfVmBZZ1Nd7tkLBggwm8Mwm2qT2WzObKpNZrO5iHgnMJGZ/1BgWd+l+uqUTw5d2A6g3weRpRll5mfmugZJWzObUjeZTambMvP9c13Djmq7OXVWkiRJkrR9mNenzkqSJEmSyvOIpiRJkiSpqJF+RnPp0qW5YsWKUa5S6oy1a9dyxx13xFzX0Y/Z1HzW5WzuueeeuXz58oHHbdy4sdH6mozbsmW01x158IMf3GjcbrvtNvCYhQun/QYQjUCXs+nfTbXprrvuGnjMpk2bGq0rYvCIrV+/no0bN844cKSN5ooVK1i1atUoVyl1xvj4+FyXMC2zqfmsy9lcvnw53/ve9wYe12QMwCWXXDLwmNtuu63Rupp6/OMf32jci170ooHHNGnyVU6Xs+nfTbXpW9/61sBjrr766kbrarJD7YwzzpjVfJ46K0mSJEkqykZTkiRJklTUUI1mRBwVET+LiOsi4vRSRUkajtmUuslsSt1kNqXyGjeaEbET8HHgaOBg4MSIOLhUYZKaMZtSN5lNqZvMptSOYY5oPhW4LjN/mZmbgc8Dx5UpS9IQzKbUTWZT6iazKbVgmEZzX+Cmnt/X1Y89QEScEhGrImLVxMTEEKuTNEtmU+qmgbO5fv36kRUnzWP+3ZRaMEyj2e+7U3KrBzLPzMzxzBxftmzZEKuTNEtmU+qmgbO55557jqAsad7z76bUgmEazXXA/j2/7wfcPFw5kgowm1I3mU2pm8ym1IJhGs0fAwdGxMMjYiFwAvD1MmVJGoLZlLrJbErdZDalFow1HZiZWyLijcC/ADsBn8rMq4pVJqkRsyl1k9mUuslsSu1o3GgCZOb5wPmFapFUiNmUuslsSt1kNqXyhjl1VpIkSZKkrQx1RFOSJLVnp512YsmSJQOPe/SjH91ofTvvvPPAYzZs2NBoXQsXLmw0rulzW758eaNxktTUaaed1mjc6tWrBx7zqEc9qtG6jj322IHHLFiwYFbzeURTkiRJklSUjaYkSZIkqSgbTUmSJElSUTaakiRJkqSibDQlSZIkSUXZaEqSJEmSirLRlCRJkiQVZaMpSZIkSSrKRlOSJEmSVJSNpiRJkiSpKBtNSZIkSVJRNpqSJEmSpKJsNCVJkiRJRY2NcmVbtmzhjjvuGHjc0qVLW6hGkqQd00EHHdRo3K677jrwmI0bNzZaV2Y2Grdw4cJG4zZt2jTwmEWLFjVal6QdyzHHHNNo3EUXXdRo3CGHHDLwmJUrVzZa1/HHHz/wmA9+8IOzms8jmpIkSZKkomw0JUmSJElF2WhKkiRJkopq3GhGxP4R8a8RsSYiroqI00oWJqkZsyl1k9mUuslsSu0Y5mJAW4C3ZOalEbEY+ElEXJiZVxeqTVIzZlPqJrMpdZPZlFrQ+IhmZt6SmZfW9zcAa4B9SxUmqRmzKXWT2ZS6yWxK7SjyGc2IWAE8Cbikz7RTImJVRKxav359idVJmqXZZnNiYmLUpUnzmtmUuslsSuUM3WhGxC7Al4E3Z+ZdU6dn5pmZOZ6Z43vuueewq5M0S4Nkc9myZaMvUJqnzKbUTWZTKmuoRjMiFlAF8rzM/EqZkiQNy2xK3WQ2pW4ym1J5w1x1NoCzgDWZ+bFyJUkahtmUuslsSt1kNqV2DHNE8xnAK4HnRMTq+nZMobokNWc2pW4ym1I3mU2pBY2/3iQzfwBEwVokFWA2pW4ym1I3mU2pHcN8j+bANm/ezNq1awcet3Tp0vLFSJKkB1i8ePHAY+66a6trpszKjTfe2Ghc06t9btiwYeAxmzdvbrSuAw88sNG4lStXNhonaXaOP/74RuMuv/zyRuMOO+ywRuPe/va3Dzzmmc98ZqN1tanI15tIkiRJkjTJRlOSJEmSVJSNpiRJkiSpKBtNSZIkSVJRNpqSJEmSpKJsNCVJkiRJRdloSpIkSZKKstGUJEmSJBVloylJkiRJKspGU5IkSZJUlI2mJEmSJKkoG01JkiRJUlE2mpIkSZKkosZGubJ7772XNWvWDDxufHy8hWokSVKvO++8c+Axq1evbrSuK6+8cqTj1q5dO/CYRYsWNVrXy172skbjVq5c2WicNN+cccYZjcZt2rSp0biPfOQjjcadeOKJjcbtKDyiKUmSJEkqykZTkiRJklSUjaYkSZIkqaihG82I2CkifhoR3yhRkKQyzKbUTWZT6iazKZVV4ojmacDgV/iR1DazKXWT2ZS6yWxKBQ3VaEbEfsCxwCfLlCOpBLMpdZPZlLrJbErlDXtE838AbwV+N90MEXFKRKyKiFUbNmwYcnWSZmmgbE5MTIyuMml+M5tSN5lNqbDGjWZEPB+4PTN/sq35MvPMzBzPzPHFixc3XZ2kWWqSzWXLlo2oOmn+MptSN5lNqR3DHNF8BvCCiFgLfB54TkR8pkhVkoZhNqVuMptSN5lNqQWNG83MfFtm7peZK4ATgO9k5iuKVSapEbMpdZPZlLrJbErt8Hs0JUmSJElFjZVYSGZ+F/huiWVJKsdsSt1kNqVuMptSOR7RlCRJkiQVVeSI5mxt2rSJq6++epSrlCRJs3TNNdcMPObiiy9utK7LLrus0bhrr7220bh169YNPOaQQw5ptK7DDjus0ThpPtqyZcvAY/baa69G6zr//PMbjVMzHtGUJEmSJBVloylJkiRJKspGU5IkSZJUlI2mJEmSJKkoG01JkiRJUlE2mpIkSZKkomw0JUmSJElF2WhKkiRJkoqy0ZQkSZIkFWWjKUmSJEkqykZTkiRJklSUjaYkSZIkqSgbTUmSJElSUWOjXNndd9/Nj370o1GuUpIkzdL1118/8Jjbb7+90bp+97vfNRq36667Nhp33333DTzm4IMPbrSuQw89tNE4aT4aGxu8HTn55JNbqESleURTkiRJklSUjaYkSZIkqSgbTUmSJElSUUM1mhGxW0R8KSKuiYg1EXFYqcIkNWc2pW4ym1I3mU2pvGEvBvQ/gX/OzJdExEJg5wI1SRqe2ZS6yWxK3WQ2pcIaN5oRsQR4NvAqgMzcDGwuU5akpsym1E1mU+omsym1Y5hTZx8BTACfjoifRsQnI+IhU2eKiFMiYlVErNq82cxKIzBwNicmJkZfpTT/mE2pm8ym1IJhGs0x4MnA32fmk4C7gdOnzpSZZ2bmeGaOL1y4cIjVSZqlgbO5bNmyUdcozUdmU+omsym1YJhGcx2wLjMvqX//ElVIJc0tsyl1k9mUuslsSi1o3Ghm5q3ATRFxUP3QEcDVRaqS1JjZlLrJbErdZDaldgx71dk/A86rr871S+DVw5ckqQCzKXWT2ZS6yWxKhQ3VaGbmamC8UC2SCjGbUjeZTambzKZU3rBHNAdy7733csUVVww8btOmTY3Wt2jRokbjJEnanm3ZsmVk61q+fPlIx913332Nxm3cuHHgMYceemijdUmShrsYkCRJkiRJW7HRlCRJkiQVZaMpSZIkSSrKRlOSJEmSVJSNpiRJkiSpKBtNSZIkSVJRNpqSJEmSpKJsNCVJkiRJRdloSpIkSZKKstGUJEmSJBVloylJkiRJKspGU5IkSZJUlI2mJEmSJKmosVGubMuWLUxMTAw87pxzzmm0vlNPPbXROEmStmeZ2WjcE57whIHHPPaxj220rv3337/RuCVLljQad//99w88ZunSpY3WJUnyiKYkSZIkqTAbTUmSJElSUTaakiRJkqSihmo0I+K/RsRVEXFlRHwuIhaVKkxSc2ZT6iazKXWT2ZTKa9xoRsS+wJuA8cx8PLATcEKpwiQ1YzalbjKbUjeZTakdw546OwY8OCLGgJ2Bm4cvSVIBZlPqJrMpdZPZlApr3Ghm5r8DHwVuBG4B7szMb06dLyJOiYhVEbGqeZmSZqtJNpt87ZCkwZhNqZvMptSOYU6d3R04Dng48DDgIRHxiqnzZeaZmTmemePNy5Q0W02yuWzZslGXKc07ZlPqJrMptWOYU2ePBK7PzInMvB/4CvD0MmVJGoLZlLrJbErdZDalFgzTaN4IPC0ido6IAI4A1pQpS9IQzKbUTWZT6iazKbVgmM9oXgJ8CbgUuKJe1pmF6pLUkNmUuslsSt1kNqV2jA0zODPfDby7UC2SCjGbUjeZTambzKZU3rBfbyJJkiRJ0gMMdURzUAsWLGDvvfceeNwPf/jDRus79dRTG42TJG2fqo9XacGCBY3GPeUpTylciSRpvvKIpiRJkiSpKBtNSZIkSVJRNpqSJEmSpKJsNCVJkiRJRdloSpIkSZKKstGUJEmSJBVloylJkiRJKspGU5IkSZJUlI2mJEmSJKkoG01JkiRJUlE2mpIkSZKkomw0JUmSJElF2WhKkiRJkooaG+XKlixZwpFHHjnwuLGxZmXecMMNA4854IADGq1LkjT3HvOYxww8Zu3ateULkSRpnvOIpiRJkiSpKBtNSZIkSVJRNpqSJEmSpKJmbDQj4lMRcXtEXNnz2B4RcWFE/Lz+uXu7ZUqaymxK3WQ2pW4ym9JozeaI5tnAUVMeOx34dmYeCHy7/l3SaJ2N2ZS66GzMptRFZ2M2pZGZsdHMzO8Dv5ry8HHAOfX9c4AXFq5L0gzMptRNZlPqJrMpjVbTz2g+NDNvAah/7jXdjBFxSkSsiohVmzZtarg6SbPUKJsTExMjK1Cap8ym1E1mU2pJ6xcDyswzM3M8M8cXLVrU9uokzVJvNpctWzbX5UiqmU2pm8ymNJimjeZtEbEPQP3z9nIlSRqC2ZS6yWxK3WQ2pZY0bTS/DpxU3z8J+FqZciQNyWxK3WQ2pW4ym1JLZvP1Jp8DLgIOioh1EXEy8FfAcyPi58Bz698ljZDZlLrJbErdZDal0RqbaYbMPHGaSUcUrkXSAMym1E1mU+omsymNVusXA5IkSZIkzS8zHtEsad999+UDH/jAwONuvvnmRus74IADGo2TJG2f1qxZM/CY8fHxFiqRJGl+84imJEmSJKkoG01JkiRJUlE2mpIkSZKkomw0JUmSJElF2WhKkiRJkoqy0ZQkSZIkFWWjKUmSJEkqykZTkiRJklSUjaYkSZIkqSgbTUmSJElSUTaakiRJkqSibDQlSZIkSUWNjXJlCxYsYJ999hl4XJMxkiRJkqS54RFNSZIkSVJRNpqSJEmSpKJsNCVJkiRJRc3YaEbEpyLi9oi4suexj0TENRFxeUR8NSJ2a7dMSVOZTambzKbUTWZTGq3ZHNE8GzhqymMXAo/PzCcA1wJvK1yXpJmdjdmUuuhszKbURWdjNqWRmbHRzMzvA7+a8tg3M3NL/evFwH4t1CZpG8ym1E1mU+omsymNVonPaL4GuGC6iRFxSkSsiohVExMTBVYnaZbMptRNZlPqJrMpFTRUoxkR7wC2AOdNN09mnpmZ45k5vmzZsmFWJ2mWzKbUTWZT6iazKZU31nRgRJwEPB84IjOzXEmShmE2pW4ym1I3mU2pHY0azYg4CvhL4I8z856yJUlqymxK3WQ2pW4ym1J7ZvP1Jp8DLgIOioh1EXEy8HfAYuDCiFgdEZ9ouU5JU5hNqZvMptRNZlMarRmPaGbmiX0ePquFWiQNwGxK3WQ2pW4ym9JolbjqrCRJkiRJv9f4YkCSJEmSpO1XRLS2bI9oSpIkSZKKstGUJEmSJBVloylJkiRJKspGU5IkSZJUlI2mJEmSJKkoG01JkiRJUlE2mpIkSZKkomw0JUmSJElF2WhKkiRJkoqy0ZQkSZIkFWWjKUmSJEkqykZTkiRJklSUjaYkSZIkqajIzNGtLGICuGGayUuBO0ZWzPS6Ugd0pxbr2FqTWg7IzGVtFDMsszmQrtQB3amlK3WA2ZwL1rG1rtSyvddhNodjHVvrSi3bex2zyuZIG81tiYhVmTluHX/QlVqsY2tdqqVtXXmu1rG1rtTSlTqgW7W0rSvP1Tq21pVarGNudOX5WsfWulLLfKnDU2clSZIkSUXZaEqSJEmSiupSo3nmXBdQ60od0J1arGNrXaqlbV15rtaxta7U0pU6oFu1tK0rz9U6ttaVWqxjbnTl+VrH1rpSy7yoozOf0ZQkSZIk7Ri6dERTkiRJkrQDsNGUJEmSJBU18kYzIo6KiJ9FxHURcXqf6Q+KiC/U0y+JiBUt1LB/RPxrRKyJiKsi4rQ+8xweEXdGxOr69t9K11GvZ21EXFGvY1Wf6RERf1Nvj8sj4skt1XFQz3NdHRF3RcSbp8zTyjaJiE9FxO0RcWXPY3tExIUR8fP65+7TjD2pnufnEXFSS7V8JCKuqbf/VyNit2nGbvO17LIu5LJej9ncej3zPpvzNZdgNqepxWyazTlnNvvWMufZnMtc1ss2m70yc2Q3YCfgF8AjgIXAZcDBU+Z5A/CJ+v4JwBdaqGMf4Mn1/cXAtX3qOBz4xgi2yVpg6TamHwNcAATwNOCSEb1Ot1J9GWvr2wR4NvBk4Mqexz4MnF7fPx34UJ9xewC/rH/uXt/fvYVangeM1fc/1K+W2byWXb11JZf1ss3mzK/VvMvmfMxlz+ttNgd8Tc2m2RzR6202B3xNR53NUeeyXrbZ7LmN+ojmU4HrMvOXmbkZ+Dxw3JR5jgPOqe9/CTgiIqJkEZl5S2ZeWt/fAKwB9i25joKOA/5PVi4GdouIfVpe5xHALzLzhpbXA0Bmfh/41ZSHe98H5wAv7DN0JXBhZv4qM38NXAgcVbqWzPxmZm6pf70Y2G+YdXRQJ3IJZnMW5mU252kuwWw2ZTbNZtvMZjOjzuZIcwlmc6pRN5r7Ajf1/L6OrcPw+3nqjXEnsGdbBdWnMjwJuKTP5MMi4rKIuCAiHtdSCQl8MyJ+EhGn9Jk+m21W2gnA56aZNoptAvDQzLwFqn9Egb36zDMX2+Y1VHvj+pnpteyqzuUSzOY0zGZ/O2IuwWxOx2z2ZzZHx2z217VsdiGXMI+zOTbM4Ab67cmZ+v0qs5mniIjYBfgy8ObMvGvK5EupDrVvjIhjgP8LHNhCGc/IzJsjYi/gwoi4pt4L8fsy+4xp7TtpImIh8ALgbX0mj2qbzNaot807gC3AedPMMtNr2VWdyiWYzX7M5jQr2nFzCWZzOmazObNZhtnsrzPZ3M5yCTtoNkd9RHMdsH/P7/sBN083T0SMAbuy9SHooUXEAqpAnpeZX5k6PTPvysyN9f3zgQURsbR0HZl5c/3zduCrVKdj9JrNNivpaODSzLxt6oRRbZPabZOnU9Q/b+8zz8i2Tf2h7OcDL8/MvsGfxWvZVZ3JZb18s9mf2ZxiB88lmM2+zOa0zObomM0+OpbNruQS5nE2R91o/hg4MCIeXu9pOAH4+pR5vg5MXmnpJcB3ptsQTdXnyJ8FrMnMj00zz96T59JHxFOpttX6wnU8JCIWT96n+pDulVNm+zrwp1F5GnDn5OH3lpzINKcZjGKb9Oh9H5wEfK3PPP8CPC8ido/qCl7Pqx8rKiKOAv4SeEFm3jPNPLN5LbuqE7kEszkDs9ljHuQSzGa/dZjN6ZnN0TGbW6+ja9nsSi5hPmczW7zaU78b1RWnrqW6Wtc76sfeWz9pgEXAF4HrgB8Bj2ihhmdSHY6+HFhd344BXg+8vp7njcBVVFcSuxh4egt1PKJe/mX1uia3R28dAXy83l5XAOMtvjY7UwVt157HWt8mVP8Q3ALcT7VH52SqzzF8G/h5/XOPet5x4JM9Y19Tv1euA17dUi3XUZ03P/lembyK3MOwDoTIAAAAfUlEQVSA87f1Wm4vty7ksl6P2exfz7zO5nzNZf0czOYD6zCbaTa7cDObW9XRmWzOVS7rZZvNnlvUC5UkSZIkqYhRnzorSZIkSdrB2WhKkiRJkoqy0ZQkSZIkFWWjKUmSJEkqykZTkiRJklSUjaYkSZIkqSgbTUmSJElSUf8fD/DvY6XQJ+cAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA5oAAADvCAYAAACT8qtkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzt3X+UXHWd5vHnWTqBAOFHSAeREKKinFEWJbSuqDPDCmhgGGHXGQ8oEsUZ1vW4qMf5ASuujrg6os64M7rDojBBBHWEuHKYMBBRhmVXkZgggiAB+RUJ0AQhIRES4LN/3NtaVFenq+793qpvd79f59RJd937vfdTt+rpyufeW7ccEQIAAAAAIJV/M+gCAAAAAADTC40mAAAAACApGk0AAAAAQFI0mgAAAACApGg0AQAAAABJ0WgCAAAAAJKi0Rwg26+3vc72k7ZP7DD9YNtrbW+2fYbt82x/tJx2pO31/a+6O621AlMN2QTyRDaBPE3zbF5le9mg65iSImLK3CS9S9JPJW2V9JCkf5C0Vw/j75V0dMJ6ai1P0rWSPrCD6RdI+tsJph0paX1Tj61c5vslrZb0tKTlg37+d/CauKGhZT/ZdntW0t8P+jHneCObz5vWaDYl7Vyu/z5JmyWtlXTsoF8DE7wmyGYezwPZjL69b35N0gZJmyTdKelPBv0amOA1QTbzeB7IZvQnmy3LfqmkpyR9bdCvgQ61fbyJuiQt6pDNkPThph/TlDmiafvDkj4j6c8l7SnptZIOlLTK9uxB1lbDgZJuqzE9CRc6vRYelPRJSRc2XUOOImL3sZukfSX9WtK3BlxWdshmcybI5pCkByT9vort/VFJ/2R7cdP15IJsdodsNmcH75uflrQ4IvaQ9BZJn7R9eNP15IJsdodsNmcH2RzzJUk3NV1HTiLi/rZs/ltJz0m6vB8rz/4maQ8V3ffb2u7fXdIjkk4rf18u6ZMt049UuYdE0sXlRv11uay/kLRYRUd/uoqmaoNauvtelzdB7X8q6S5Jj0m6QtILy/vvbhu/c9u476nYE/hUOf1lrfV0U4uKP1z/T9Ljkn4i6ciW5V8n6b9L+r/luIN2sP0/qR6PaHaqVdKHy+drg6R3t817nqRVKo7Q/KukA8tpY8/RUFvtfyLpd8rt82z5uB9v8DW4TNIvJHnQecjpRjYHm82WMbdIemuXzxnZnAE3sjn4bEo6uNw+b9vRfJ22ncjmtL2JbA4sm5JOkvRP6vHI4Vh+yp/fJekGSZ+T9CtJ96jlrKJy3k9L+pGkJyR9R9K89sfZMv+9ko6WtFTSNknby8f9kwZfgx+T9P1+vN6nyhHN10naRdKK1jsj4klJV0k6ZrIFRMQ7Jd0v6Q+j6OjPbZn871UcSn+TpDNtH11zeZIk229U8WJ7m6T9VJzq9o1y/Evaxj/dtvw3Svo/kt5fTr+zl1ps7y/pn1U0ifMk/Zmky20Ptwx9p4o/SHPL2pr0AhV77faX9B5JX7K9d8v0d0g6R9J8STdLumSyBUbE7ZLeK+kH5ePeq9N8tv+n7ccnuN3SZf3LJH01yoTiN8jmgLNpe18Vb9pV9xSTzemJbA4om+XrequkO1T8Z3/lJJtmImRzeiKbA8im7T0kfULFzpu6/p2kn6vI3rmSLrDtlumnSjpN0gslPSPp7yZbYET8i6RPSfpm+bhf2Wk+21fuIJtXdln/qZIu6nLeWqZKozlf0qMR8UyHaRvK6XX8VURsiYifSvpHSSfXXN6Yd0i6MCLWlKE7S9IRfTrF7RRJKyNiZUQ8FxGrVHze8riWeZZHxG0R8UxEbG+4nu2SPhER2yNipYq9NQe3TP/niLi+3E4fUbGdDkix4oh4X0TsNcHt0MnG216k4jTFvoRyiiGbvUuWTduzVPzn8qKIuKNiPWRzeiKbvUuSzYh4n4r/7P6uimbi6U7zdYFsTk9ks3cpsnmOpAsi4oEE9dwXEV+OiGdVvMb3U3Gq+JiLI+LWiNii4uMtb7O9U4L1KiKO30E2j59svO3fLWu9LEU9k5kqjeajkubbHuowbb9yeh2tL7r7VOyBSOGFatmrUu6t2qhi72TTDpT0x617OiS9QcX2GpMibN3a2PZHdauK00TG1VJup8eU7nmo61QVF064Z9CFZIhs9i5JNsvPoFys4lSb99eoh2xOT2Szd8neNyPi2Yi4QdJCSf+5Yj1kc3oim72rlU3br1JxeurfJqrnobEfImJr+WPHbKrYZrNUfwdCKsskXV4+f42bKo3mD1TsEfyPrXfa3k3SsSqudCVJWyTt2jLLC9qWM9HpG617ABepOLe9zvLGPKgiHK317iPpl5OMq6K9lgdU7FFp3dOxW0T89Q7GDNJvngPbu6s4NeJBFc+BNPHzMOljKC+h/eQEt25ON+zbKQZTENmcXPJslqfoXKBir+RbGz4jgWxOTWRzcv143xyS9JJaVU6MbE5NZHNyqbN5pIrPsN5v+yEVp96+1faalEW3aH8OtqvYgfC856A8ytl6+m832bxqB9m8apKxcyT9sfqYzSnRaEbEE5L+StLf215qe1Z5qP5bKj4sf3E5682SjrM9z/YLJH2wbVEPS3pxh1V81Pautl8h6d2SvllzeWMulfRu26+yvbOKc69vjIh7J33QvWuv5WuS/tD2m23vZHsXF99TtLDbBdoesr2LpJ0kjS1jqGV62D4yUf3H2X6Di6utnaNiOz0QEaMq/oidUj6O0/T8N+2HJS30Dq7SFhHvjZarbbXdXrGjomy/TsXeOq6a1wHZ7ErybKq4DP7vqPgMy6/bJ5JNkM2uJM2m7QW2T7K9ezn+zSpOW/xeyzxkc4Yjm11J/b55vooMvKq8nafiM59vliTbi8tsLk5U/ym2X257VxWfC72sPM32Tkm72P4DFx99OVvFV5aNeVjSYu/gqrkRcewOsnnsJHX9BxUXU/p+rUfXgynRaEpSFB9M/q8qrvK0SdKNKvZwHBW//dDxxSquRHWvpGv023CN+bSks10cdv+zlvv/VcVVtK6V9LmIuKbm8sZqvlbFudmXqzjv/iUqrnjVhOfVUp6DfoKKbTaqYlv9uXp7zs9WcfWuM1WcH//r8j6V4X5SxXdApXCpiqtgPSbpcBWfBRjzpypq3yjpFSquOjbmeyougvKQ7bqnm3SyTNKKiNjcwLKnBbI5qaTZtH2gpP+k4s3yoZY9me8op5NNSCKbXUj9vhkqTpNdr+JqlJ+T9MGI+I5ENvFbZHNSSbMZEVsj4qGxm4ocPlXulJGKI5D3Kd3R2YtVXFX3IRUXfjqjrOMJSe+T9JVyXVtU/L0YM7ZzZqObOdra9wt0uY/ryk655+IeSbOi84eyMQHbp0h6RUSclWBZy1Vc7vns2oVhWiCb1ZFNNIlsVkc20SSyWZ3tsyWNRsT/SrCs61R8dcpXahc2DXT6IDIwqYj42qBrADAe2QTyRDaBPEXEJwddw3Q1ZU6dBQAAAABMDTP61FkAAAAAQHoc0QQAAAAAJNXXz2jOmzcvFi7s5Qr+hdmzJ7wCNzBl3HvvvXr00Uc96Do6mT9/fixevHjQZTRiy5Ytk8/UwaZNmxJXMrF58+ZVGrfzzjtPPhMmlXM2586dG/Pn9/49388991yl9fXzLCe72ibv57hnn3220rqefvrpyWfKYNy2bdt6HlP1tVVl+0eEIiLLbE7n901gMt2+b/a10Vy4cKGuvPLKnsctWrSogWqA/hoZGRl0CRNavHixVq9ePegyGvGjH/2o0rirr7660rgq/1F/+9vfXmldBx10UKVxeL6cszl//nx9/OMf73nc1q1bK61v+/btPY+p2vgNDVX7L0jVHSyzZs3qeczjjz9eaV333HNPpXHr1q2rNO7uu++uNO6Xv+z92xyq7oTbZZddeh7z1FNPVVpXP0zn901gMt2+b3LqLAAAAAAgKRpNAAAAAEBStRpN20tt/9z2XbbPTFUUgHrIJpAnsgnkiWwC6VVuNG3vJOlLko6V9HJJJ9t+earCAFRDNoE8kU0gT2QTaEadI5qvkXRXRPwiIrZJ+oakE9KUBaAGsgnkiWwCeSKbQAPqNJr7S3qg5ff15X3PY/t026ttr37sscdqrA5Al3rO5ujoaN+KA2awnrO5efPmvhUHzGC8bwINqNNodrqe+bjr+kfE+RExEhEjVb8rDkBPes7m8PBwH8oCZryeszl37tw+lAXMeLxvAg2o02iul3RAy+8LJT1YrxwACZBNIE9kE8gT2QQaUKfRvEnSS22/yPZsSSdJuiJNWQBqIJtAnsgmkCeyCTRgqOrAiHjG9vslXS1pJ0kXRsRtySoDUAnZBPJENoE8kU2gGZUbTUmKiJWSViaqBUAiZBPIE9kE8kQ2gfTqnDoLAAAAAMA4tY5o9mrbtm1av359z+MWLVrUQDUzz5e//OVK4z71qU9VGnf55ZdXGrdkyZJK4zB1PP300z2PufTSSyutq+q4+++/v9K4Y445pucxBx10UKV1YfqbPXu2DjjggMlnbDNnzpxK69trr716HrNgwYJK69pnn30qjeunbdu2VRpX9e9H1XEPPljtujW/+tWveh5jd7pA6+SqvE7OOuusSusCkAeOaAIAAAAAkqLRBAAAAAAkRaMJAAAAAEiKRhMAAAAAkBSNJgAAAAAgKRpNAAAAAEBSNJoAAAAAgKRoNAEAAAAASdFoAgAAAACSotEEAAAAACRFowkAAAAASIpGEwAAAACQFI0mAAAAACCpoX6u7LnnntOmTZv6uUq0OOKIIyqNe9nLXlZp3JIlSyqNw/S3cuXKnsdceumlldb13e9+t9K4RYsWVRp39NFHVxoHdDJnzhwdcsghPY+bN29epfUNDfX1vwXZmz17dqVxBx10UF/HTVfnnnvuoEsAUANHNAEAAAAASdFoAgAAAACSotEEAAAAACRVudG0fYDt79u+3fZttj+QsjAA1ZBNIE9kE8gT2QSaUedT/89I+nBErLE9V9KPba+KiJ8lqg1ANWQTyBPZBPJENoEGVD6iGREbImJN+fNmSbdL2j9VYQCqIZtAnsgmkCeyCTQjyWc0bS+WdJikGztMO932aturn3jiiRSrA9ClbrM5Ojra79KAGa3bbG7cuLHfpQEzGu+bQDq1G03bu0u6XNIHI2Lcl2RGxPkRMRIRI3vuuWfd1QHoUi/ZHB4e7n+BwAzVSzb32Wef/hcIzFC8bwJp1Wo0bc9SEchLImJFmpIA1EU2gTyRTSBPZBNIr85VZy3pAkm3R8TfpCsJQB1kE8gT2QTyRDaBZtQ5ovl6Se+U9EbbN5e34xLVBaA6sgnkiWwCeSKbQAMqf71JRNwgyQlrAZAA2QTyRDaBPJFNoBl1vkezZ3vssYeWLl3az1WixSGHHFJp3NVXX524Esx0ixYt6nlM1QsvvPKVr6w07tRTT6007sQTT6w0DuhkaGhICxYsGHQZAAD0LMnXmwAAAAAAMIZGEwAAAACQFI0mAAAAACApGk0AAAAAQFI0mgAAAACApGg0AQAAAABJ0WgCAAAAAJKi0QQAAAAAJEWjCQAAAABIikYTAAAAAJAUjSYAAAAAICkaTQAAAABAUjSaAAAAAICkhgZdAICZ5/DDD+95zOc///lK69p9990rjZs7d26lcRjvs5/9bM9j1q1bV2ldixYt6nnMhg0bKq0LAABMjCOaAAAAAICkaDQBAAAAAEnRaAIAAAAAkqrdaNreyfZa21emKAhAGmQTyBPZBPJENoG0UhzR/ICk2xMsB0BaZBPIE9kE8kQ2gYRqNZq2F0r6A0lfSVMOgBTIJpAnsgnkiWwC6dU9ovkFSX8h6bmJZrB9uu3VtlePjo7WXB2ALpFNIE9kE8gT2QQSq9xo2j5e0iMR8eMdzRcR50fESESMDA8PV10dgC6RTSBPZBPIE9kEmlHniObrJb3F9r2SviHpjba/lqQqAHWQTSBPZBPIE9kEGlC50YyIsyJiYUQslnSSpO9FxCnJKgNQCdkE8kQ2gTyRTaAZfI8mAAAAACCpoRQLiYjrJF2XYlkA0iGbQJ7IJpAnsgmkwxFNAAAAAEBSSY5oAkDT9ttvv0GXgIp23nnnnsdcf/31ldZ1zz339Dxm+/btldYFAAAmxhFNAAAAAEBSNJoAAAAAgKRoNAEAAAAASdFoAgAAAACSotEEAAAAACRFowkAAAAASIpGEwAAAACQFI0mAAAAACApGk0AAAAAQFI0mgAAAACApGg0AQAAAABJ0WgCAAAAAJKi0QQAAAAAJDU06AIAANPbGWec0ZcxknTLLbf0PObkk0+utC4AAKa66667rucxmzdv7mo+jmgCAAAAAJKi0QQAAAAAJEWjCQAAAABIqlajaXsv25fZvsP27baPSFUYgOrIJpAnsgnkiWwC6dW9GND/kPQvEfFHtmdL2jVBTQDqI5tAnsgmkCeyCSRWudG0vYek35P0LkmKiG2StqUpC0BVZBPIE9kE8kQ2gWbUOXX2xZJGJf2j7bW2v2J7t/aZbJ9ue7Xt1aOjozVWB6BLZBPIE9kE8kQ2gQbUaTSHJC2R9A8RcZikLZLObJ8pIs6PiJGIGBkeHq6xOgBdIptAnsgmkCeyCTSgTqO5XtL6iLix/P0yFSEFMFhkE8gT2QTyRDaBBlRuNCPiIUkP2D64vOsoST9LUhWAysgmkCeyCeSJbALNqHvV2f8i6ZLy6ly/kPTu+iUBSIBsAnkim0CeyCaQWK1GMyJuljSSqBYAiZBNIE9kE8gT2QTSq3tEEwOyZs2ansd84QtfqLSur371q5XGAUC/HXrooT2PmTNnTgOVAADQP48++milcTfccEPPY5588smu5qtzMSAAAAAAAMah0QQAAAAAJEWjCQAAAABIikYTAAAAAJAUjSYAAAAAICkaTQAAAABAUjSaAAAAAICkaDQBAAAAAEnRaAIAAAAAkqLRBAAAAAAkRaMJAAAAAEiKRhMAAAAAkBSNJgAAAAAgqaFBFzDTbdy4sdK4ww8/vOcxH/rQhyqtCwAAAEC+Lr744krj1q5d2/OYrVu3djUfRzQBAAAAAEnRaAIAAAAAkqLRBAAAAAAkVavRtP0h27fZvtX2123vkqowANWRTSBPZBPIE9kE0qvcaNreX9IZkkYi4hBJO0k6KVVhAKohm0CeyCaQJ7IJNKPuqbNDkubYHpK0q6QH65cEIAGyCeSJbAJ5IptAYpUbzYj4paTPSbpf0gZJT0TENe3z2T7d9mrbq0dHR6tXCqArZBPIE9kE8kQ2gWbUOXV2b0knSHqRpBdK2s32Ke3zRcT5ETESESPDw8PVKwXQFbIJ5IlsAnkim0Az6pw6e7SkeyJiNCK2S1oh6XVpygJQA9kE8kQ2gTyRTaABdRrN+yW91vauti3pKEm3pykLQA1kE8gT2QTyRDaBBtT5jOaNki6TtEbST8tlnZ+oLgAVkU0gT2QTyBPZBJoxVGdwRHxM0scS1QIgEbIJ5IlsAnkim0B6db/eBAAAAACA56l1RBP13XTTTZXGnXPOOT2POfvssyutCwAAAEC+VqxYUWnc9u3bGxvDEU0AAAAAQFI0mgAAAACApGg0AQAAAABJ0WgCAAAAAJKi0QQAAAAAJEWjCQAAAABIikYTAAAAAJAUjSYAAAAAICkaTQAAAABAUjSaAAAAAICkaDQBAAAAAEnRaAIAAAAAkqLRBAAAAAAkNTToAma6pUuX9nUcAAAAgHx98Ytf7HnM2rVrK63rsMMO63mM7a7m44gmAAAAACApGk0AAAAAQFI0mgAAAACApCZtNG1faPsR27e23DfP9irb68p/9262TADtyCaQJ7IJ5IlsAv3VzRHN5ZLarzxzpqRrI+Klkq4tfwfQX8tFNoEcLRfZBHK0XGQT6JtJG82IuF7SY213nyDpovLniySdmLguAJMgm0CeyCaQJ7IJ9FfVz2juGxEbJKn8d8FEM9o+3fZq26tHR0crrg5Al8gmkCeyCeSJbAINafxiQBFxfkSMRMTI8PBw06sD0CWyCeSJbAJ5IptAb6o2mg/b3k+Syn8fSVcSgBrIJpAnsgnkiWwCDanaaF4haVn58zJJ30lTDoCayCaQJ7IJ5IlsAg3p5utNvi7pB5IOtr3e9nsk/bWkY2yvk3RM+TuAPiKbQJ7IJpAnsgn019BkM0TEyRNMOipxLQB6QDaBPJFNIE9kE+ivxi8GBAAAAACYWSY9ogkAAACgvkceqXatoQULJvzWFUxDF154Yc9jtmzZUmldBx54YM9jfvazn3U1H0c0AQAAAABJ0WgCAAAAAJKi0QQAAAAAJEWjCQAAAABIikYTAAAAAJAUjSYAAAAAICkaTQAAAABAUjSaAAAAAICkaDQBAAAAAEnRaAIAAAAAkqLRBAAAAAAkRaMJAAAAAEhqaNAFAAAAADPBggULBl0C+mjFihWVxq1du7bnMfvuu2+ldR166KE9j7nuuuu6mo8jmgAAAACApGg0AQAAAABJ0WgCAAAAAJKatNG0faHtR2zf2nLfZ23fYfsW29+2vVezZQJoRzaBPJFNIE9kE+ivbo5oLpe0tO2+VZIOiYhDJd0p6azEdQGY3HKRTSBHy0U2gRwtF9kE+mbSRjMirpf0WNt910TEM+WvP5S0sIHaAOwA2QTyRDaBPJFNoL9SfEbzNElXTTTR9um2V9tePTo6mmB1ALpENoE8kU0gT2QTSKhWo2n7I5KekXTJRPNExPkRMRIRI8PDw3VWB6BLZBPIE9kE8kQ2gfSGqg60vUzS8ZKOiohIVxKAOsgmkCeyCeSJbALNqNRo2l4q6S8l/X5EbE1bEoCqyCaQJ7IJ5IlsAs3p5utNvi7pB5IOtr3e9nskfVHSXEmrbN9s+7yG6wTQhmwCeSKbQJ7IJtBfkx7RjIiTO9x9QQO1AOgB2QTyRDaBPJFNoL9SXHUWAAAAAIDfqHwxIAAAAABAZxdddFGlcbNmzep5zJIlSyqt69WvfnXPY3bbbbeu5uOIJgAAAAAgKRpNAAAAAEBSNJoAAAAAgKRoNAEAAAAASdFoAgAAAACSotEEAAAAACRFowkAAAAASIpGEwAAAACQFI0mAAAAACApGk0AAAAAQFI0mgAAAACApGg0AQAAAABJ0WgCAAAAAJJyRPRvZfaopPsmmDxf0qN9K2ZiudQh5VMLdYxXpZYDI2K4iWLqIps9yaUOKZ9acqlDIpuDQB3j5VLLVK+DbNZDHePlUstUr6OrbPa10dwR26sjYoQ6fiuXWqhjvJxqaVouj5U6xsulllzqkPKqpWm5PFbqGC+XWqhjMHJ5vNQxXi61zJQ6OHUWAAAAAJAUjSYAAAAAIKmcGs3zB11AKZc6pHxqoY7xcqqlabk8VuoYL5dacqlDyquWpuXyWKljvFxqoY7ByOXxUsd4udQyI+rI5jOaAAAAAIDpIacjmgAAAACAaYBGEwAAAACQVN8bTdtLbf/c9l22z+wwfWfb3yyn32h7cQM1HGD7+7Zvt32b7Q90mOdI20/Yvrm8/bfUdZTrudf2T8t1rO4w3bb/rtwet9he0lAdB7c81pttb7L9wbZ5Gtkmti+0/YjtW1vum2d7le115b97TzB2WTnPOtvLGqrls7bvKLf/t23vNcHYHT6XOcshl+V6yOb49cz4bM7UXEpkc4JayCbZHDiy2bGWgWdzkLksl002W0VE326SdpJ0t6QXS5ot6SeSXt42z/sknVf+fJKkbzZQx36SlpQ/z5V0Z4c6jpR0ZR+2yb2S5u9g+nGSrpJkSa+VdGOfnqeHVHwZa+PbRNLvSVoi6daW+86VdGb585mSPtNh3DxJvyj/3bv8ee8GanmTpKHy5890qqWb5zLXWy65LJdNNid/rmZcNmdiLlueb7LZ43NKNslmn55vstnjc9rvbPY7l+WyyWbLrd9HNF8j6a6I+EVEbJP0DUkntM1zgqSLyp8vk3SUbacsIiI2RMSa8ufNkm6XtH/KdSR0gqSvRuGHkvayvV/D6zxK0t0RcV/D65EkRcT1kh5ru7v1dXCRpBM7DH2zpFUR8VhE/ErSKklLU9cSEddExDPlrz+UtLDOOjKURS4lstmFGZnNGZpLiWxWRTbJZtPIZjX9zmZfcymRzXb9bjT3l/RAy+/rNT4Mv5mn3BhPSNqnqYLKUxkOk3Rjh8lH2P6J7atsv6KhEkLSNbZ/bPv0DtO72WapnSTp6xNM68c2kaR9I2KDVPwRlbSgwzyD2Danqdgb18lkz2WussulRDYnQDY7m465lMjmRMhmZ2Szf8hmZ7llM4dcSjM4m0N1BlfQaU9O+/erdDNPErZ3l3S5pA9GxKa2yWtUHGp/0vZxkv63pJc2UMbrI+JB2wskrbJ9R7kX4jdldhjT2HfS2J4t6S2SzuowuV/bpFv93jYfkfSMpEsmmGWy5zJXWeVSIpudkM0JVjR9cymRzYmQzerIZhpks7NssjnFcilN02z2+4jmekkHtPy+UNKDE81je0jSnhp/CLo227NUBPKSiFjRPj0iNkXEk+XPKyXNsj0/dR0R8WD57yOSvq3idIxW3WyzlI6VtCYiHm6f0K9tUnp47HSK8t9HOszTt21Tfij7eEnviIiOwe/iucxVNrksl082OyObbaZ5LiWy2RHZnBDZ7B+y2UFm2cwll9IMzma/G82bJL3U9ovKPQ0nSbqibZ4rJI1daemPJH1vog1RVXmO/AWSbo+Iv5lgnheMnUtv+zUqttXGxHXsZnvu2M8qPqR7a9tsV0g61YXXSnpi7PB7Q07WBKcZ9GObtGh9HSyT9J0O81wt6U2293ZxBa83lfclZXuppL+U9JaI2DrBPN08l7nKIpcS2ZwE2WwxA3Ipkc1O6yCbEyOb/UM2x68jt2zmkktpJmczGrzaU6ebiitO3anial0fKe/7RPmgJWkXSd+SdJekH0l6cQM1vEHF4ehbJN1c3o6T9F5J7y3neb+k21RcSeyHkl7XQB0vLpf/k3JdY9ujtQ5L+lK5vX4qaaTB52ZXFUHbs+W+xreJij8EGyRtV7FH5z0qPsdwraR15b/zynlHJH2lZexp5WvlLknvbqiWu1ScNz/2Whm7itwLJa3c0XM5VW455LJcD9nsXM+MzuZMzWX5GMjm8+sgm0E2c7iRzXF1ZJPXHC9AAAAAU0lEQVTNQeWyXDbZbLm5XCgAAAAAAEn0+9RZAAAAAMA0R6MJAAAAAEiKRhMAAAAAkBSNJgAAAAAgKRpNAAAAAEBSNJoAAAAAgKRoNAEAAAAASf1/syjsniJsv44AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA5oAAADvCAYAAACT8qtkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzt3X2UZHV95/H3VxoZmGEYxpkBBlhHs0JET9aYNutDDCwQBCLCia7iSoIPCWs8btRjHmA1ZjVm4wMnTxsTw4qCD1ETNNGjEGExxnUjJBNEhAxBICCTmZFGcEbkceS7f9w7UvRUT1fd+lX1r6ffr3PqdHfd+7v3W7fq093fe2/disxEkiRJkqRSHrfQBUiSJEmS9i42mpIkSZKkomw0JUmSJElF2WhKkiRJkoqy0ZQkSZIkFWWjKUmSJEkqykZzAUXE8yLimxFxb0Sc0Wf60RHxtYj4XkT8SkS8PyJ+s512XERsnnzVg+mtVVpszKZUJ7Mp1Wkvz+ZlEXH2QtexKGXmorkBrwS+AdwHbAP+FFg1xPjbgBML1jPS8oArgTfsYfqFwO/PMe04YPO4Hlu7zNcDG4EHgYsW+vnfw2viK2Na9gbgUuCe9vX2x8DUQj/mGm9m8zHTxppNYL92/bcD3wO+Bpyy0K+BOV4TZrOO58Fs5sT+bn4U2ArsAG4CfnGhXwNzvCbMZh3Pg9nMyWSzZ9lPAR4APrrQr4E+tf2PcdUFPAP4v8B2YDPwtkk8pkVzRDMi3gy8G/g14CDg2cATgSsi4vELWdsIngjcMML0IqLR77WwBXgn8MFx11CpPwHuBA6jCeixwOsWtKIKmc3xmSObU8AdNK/Hg4DfBP4iIjaMu56KmM0BmM3x2cPfzd8FNmTmSuBFwDsj4ifGXU9FzOYAzOb47CGbu7wP+Mdx11GhPwe+DKymyeUvR8SLxr7Whe7eB+zCVwL3Ai+ddf8Kml9or25/vgh4Z8/042j3kAAfAR4B7m+X9es0e94SOIemqdoKvLln/FDLm6P2XwJuBu4GPgusb++/Zdb4/WaN+yLwA5q9LvcCR/XWM0gtNL+4/h74LvB14Lie5X8J+B3g/7Xj/v0etv87GfKIZr9agTe3z9dW4FWz5n0/cAXNEZq/A57YTtv1HE3Nqv0Xgae22+cH7eP+buHX3Sbg1J6f3wv82ULnoaYbZnNBs9kz5jrgxQM+Z7vVitnc626YzQXPJnB0u31euqf5+m07zOZee8NsLlg2gTOBv2DII4e78tN+/0rgK8D5NEfu/5Wes4raeX8X+Aeao4efAVbPfpw9898GnAicDDwEPNw+7q8Xft3dBxzT8/NfAueN+/W+WI5oPhdYBny6987MvBe4DPiZ+RaQmT8PfAs4LTNXZOZ7eib/J5pD6ScB50bEiSMuD4CIOJ7mxfZSmr17twOfaMf/yKzxD85a/vE0h7hf306/aZhaIuJw4PM0TeJq4FeBT0XE2p6hP0/zC+nAtrZxOpRmr93hwGuA90XEwT3TXwH8NrAGuBb42HwLzMxNwGuBr7aPe1W/+SLiTyLiu3PcrtvDKv4QODMiDmi35ynA3wzwWJcSs7nA2YyIQ2j+aHfdU2w2905mc4Gy2b6u7wNupPln/9J5Ns1czObeyWwuQDYjYiXwDpqdN6P6j8C/0GTvPcCFERE9038BeDWwHtgJ/NF8C8zMvwH+J/DJ9nH/h37zRcTn9pDNz+1hFX8A/EJE7BsRRwPPAf7PAI91JIul0VwD3JWZO/tM29pOH8XbM/P7mfkN4EPAy0dc3i6vAD6Ymde0oTsPeM6ETnE7C7g0My/NzEcy8wqa91ue2jPPRZl5Q2buzMyHx1zPw8A7MvPhzLyUZm/N0T3TP5+ZX26301tottORJVacma/LzFVz3H5sD0P/DngazXttNtNsv78uUdNexGwOr1g2I2Jfmn8uL87MGzvWYzb3TmZzeEWymZmvo/ln9/k0zcSD/eYbgNncO5nN4ZXI5m8DF2bmHQXquT0z/3dm/gC4mKbxPqRn+kcy8/rM/D7N21teGhH7FFgvmfnCPWTzhXsY+jngJTRHfG+k2RZjP4V4sTSadwFrImKqz7TD2umj6H3R3U6zB6KE9fTsVWn3Vn2HZu/kuD0R+M+9ezqAn6LZXruUCNugvjPrl+p9NKeJ7FZLu53uptzzMLT2/P4v0PyTsJzmF//BNO+p0KPM5vCKZLN9jX6E5lSb149Qj9ncO5nN4RX7u5mZP8jMrwBHAL/csR6zuXcym8MbKZsR8Qya01N/v1A923Z9k5n3td/2zSbNNtuX0XcgdBYRq2nOLHgHzdH0I4EXRMTY3z+9WBrNr9LsEfy53jsjYjnNaRlXtnd9HzigZ5ZDZy0n51h+7x7Af0dzbvsoy9tlC004eut9AvBv84zrYnYtd9DsUend07E8M9+1hzEL6YfPQUSsoDk1YgvNcwBzPw/zPoZoLqF97xy3uU43XN3W9MeZ+WBmfodmz+Cpc8y/VJnN+RXPZnuKzoU0e1BfPOYzEszm4mQ25zeJv5tTwI+MVOXczObiZDbnVzqbx9G8h/VbEbGN5tTbF0fENSWL7jH7OXiYZgfCY56D9ihn7+m/g2Tzsj1k87I5hj0Z+EFmfrg94ruZ5rTnsWdzUTSambkdeDvwvyLi5Pb84g00b2TdTLNXH5r3KJwaEasj4lDgjbMW9W2ajT3bb7bvJ3ga8CrgkyMub5c/B14VEc+IiP1ozr2+OjNvm/dBD292LR8FTouIF0TEPhGxLJrPKTpi0AVGxFRELAP2AXYtY6pnekbEcYXqPzUifiqaq639Ns12uiMzZ2h+iZ3VPo5X89g/2t8Gjog9XKUtM1+bzfnu/W5Pm2PMXTRv8P7ldjusAs6meQO6WmZzIMWzSXMZ/KfSvIfl/tkTzabM5kCKZjMi1kXEmRGxoh3/AprTFr/YM4/ZXOLM5kBK/928gCYDz2hv76d5z+cLACJiQ5vNDYXqPysijomIA2iOIl7SnmZ7E7AsIn42mre+vJXmI8t2+TawIfZw1dzMPGUP2TxljmE30eyj/i8R8bj2+X8ZE8jmomg0AbJ5Y/J/p7nK0w7gapo9HCfko286/gjNRrsNuJxHw7XL7wJvjeaw+6/23P93NFfRuhI4PzMvH3F5u2q+kubc7E/RnHf/IzRXvBqHx9TSnoN+Os02m6HZVr/GcM/5W2nO5T6X5vz4+9v7aMN9L81nQJXw58Bv0Zz68xM07wXY5Zdoav8OzXs//r5n2hdpLoKyLSJGPd1ktp+juQrYDM3rYyfwpsLrWPTM5ryKZjMingj8V5o/ltvi0T2Zr2inm00BZnMApf9uJs1psptprkZ5PvDGzPwMmE09ymzOq2g2M/O+zNy260aTwwfanTLQHIG8nXJHZz9Cc1XdbTSnqv5KW8d2mo/7+UC7ru/T/L7Y5S/br9+JgkdbM3MHTTbfRPO76Vrgepor9Y5VZNZ09uRktXsu/hXYN/u/KVtziIizgKdl5nkFlnURzeWe3zpyYdormM3uzKbGyWx2ZzY1Tmazu4h4KzCTmX9WYFlfovnolA+MXNheoN8bkaV5ZeZHF7oGSbszm1KdzKZUp8x850LXsLdaNKfOSpIkSZIWhyV96qwkSZIkqTyPaEqSJEmSiproezSXLVuWy5cvH3rcQw891Gl969cP/xm1Bx54YKd1SfO57bbbuOuuu2Kh6+hnzZo1uWHDhoUuYyweeOCBTuOmprr9euw6Tgun5myuXLky161bN7H1PfLIIxMZA9D1jKqu63v44eE/bvbBBx+cf6ZC6wKI6PYyXLFixfwz9fGEJzxh6DHLli3rtK4uas7mqlWr8rDDDht6XJf/g6XaDJrNif5HtHz5ck45Za6PeJnb5s2b55+pj7e//e1Djzn22GM7rUuaz/T09EKXMKcNGzawcePGhS5jLDZt2tRp3Nq1a+efqY81a9Z0GqeFU3M2161bx/nnnz/0uK7N2H333Tf0mK47c7o2Y13Xt3Xr1qHH3HLLLZ3WtWXLlk7j9ttvv/ln6uO5z31up3GvfOUrhx5z1FFHdVpXFzVn87DDDuPDH/7w0OOe9axnjaEaabIGzaanzkqSJEmSirLRlCRJkiQVNVKjGREnR8S/RMTNEXFuqaIkjcZsSnUym1KdzKZUXudGMyL2Ad4HnAIcA7w8Io4pVZikbsymVCezKdXJbErjMcoRzZ8Ebs7MWzPzIeATwOllypI0ArMp1clsSnUym9IYjNJoHg7c0fPz5va+x4iIcyJiY0Rs7HqlOElDGTqbMzMzEytOWsKGzuaOHTsmVpy0hA2dze9+97sTK05arEZpNPt9dspuH4qVmRdk5nRmTk/ys5ekJWzobHb9KA9JQxk6mytXrpxAWdKSN3Q2V61aNYGypMVtlEZzM3Bkz89HAN0+OEpSSWZTqpPZlOpkNqUxGKXR/EfgKRHxpIh4PHAm8NkyZUkagdmU6mQ2pTqZTWkMproOzMydEfF64AvAPsAHM/OGYpVJ6sRsSnUym1KdzKY0Hp0bTYDMvBS4tFAtkgoxm1KdzKZUJ7MplTfKqbOSJEmSJO1mpCOaw3rwwQe55ZZbhh531VVXdVrf5z//+aHHHHvssZ3WJalOT33qUxe6BKmzAw88kOOPP37ocYvharXbt2/vNO66667rNO4LX/jC0GNuvfXWTuu69tprO43renX+Jz3pSZ3GHXHEEZ3GCZYvX86znvWshS5DqppHNCVJkiRJRdloSpIkSZKKstGUJEmSJBVloylJkiRJKspGU5IkSZJUlI2mJEmSJKkoG01JkiRJUlE2mpIkSZKkomw0JUmSJElF2WhKkiRJkoqy0ZQkSZIkFWWjKUmSJEkqykZTkiRJklTU1CRXtv/++/P0pz996HH33HNPp/UdddRRncZJkhbWy172sk7jLrnkkqHHPPLII53WNQn77LMPK1euXOgyxuKggw7qNO75z39+p3EzMzNDj7n++us7rWvLli2dxh1wwAGdxm3YsGGi65OkQXhEU5IkSZJUlI2mJEmSJKkoG01JkiRJUlGdG82IODIi/jYiNkXEDRHxhpKFSerGbEp1MptSncymNB6jXAxoJ/DmzLwmIg4E/ikirsjMfy5Um6RuzKZUJ7Mp1clsSmPQ+YhmZm7NzGva778HbAIOL1WYpG7MplQnsynVyWxK41HkPZoRsQH4ceDqPtPOiYiNEbHxgQceKLE6SQMaNJtdLvsvqTuzKdXJbErljNxoRsQK4FPAGzNzx+zpmXlBZk5n5vSyZctGXZ2kAQ2TzbVr106+QGmJMptSncymVNZIjWZE7EsTyI9l5qfLlCRpVGZTqpPZlOpkNqXyRrnqbAAXApsy8/fKlSRpFGZTqpPZlOpkNqXxGOWI5vOAnweOj4hr29upheqS1J3ZlOpkNqU6mU1pDDp/vElmfgWIgrVIKsBsSnUym1KdzKY0HqN8jubQDjnkEN70pjcNPe7YY4/ttL6zzjqr0zhJ0sLasGFDp3GPPPJI2UK01zjjjDOGHrNixYpO6zrxxBM7jVu9enWncaeddlqncZI0TkU+3kSSJEmSpF1sNCVJkiRJRdloSpIkSZKKstGUJEmSJBVloylJkiRJKspGU5IkSZJUlI2mJEmSJKkoG01JkiRJUlE2mpIkSZKkomw0JUmSJElF2WhKkiRJkoqy0ZQkSZIkFWWjKUmSJEkqamqSK1u2bBnHHHPM0OO6jJEkLV7vfve7JzZuenq607q0uDzuccPvWz/ppJM6ravrOEnam3hEU5IkSZJUlI2mJEmSJKkoG01JkiRJUlEjN5oRsU9EfC0iPleiIEllmE2pTmZTqpPZlMoqcUTzDcCmAsuRVJbZlOpkNqU6mU2poJEazYg4AvhZ4ANlypFUgtmU6mQ2pTqZTam8UY9o/gHw68Ajc80QEedExMaI2DgzMzPi6iQNyGxKdTKbUp3MplRY50YzIl4I3JmZ/7Sn+TLzgsyczszptWvXdl2dpAGZTalOZlOqk9mUxmOUI5rPA14UEbcBnwCOj4iPFqlK0ijMplQnsynVyWxKY9C50czM8zLziMzcAJwJfDEzzypWmaROzKZUJ7Mp1clsSuPh52hKkiRJkoqaKrGQzPwS8KUSy5JUjtmU6mQ2pTqZTakcj2hKkiRJkoqy0ZQkSZIkFWWjKUmSJEkqykZTkiRJklSUjaYkSZIkqSgbTUmSJElSUTaakiRJkqSibDQlSZIkSUXZaEqSJEmSirLRlCRJkiQVZaMpSZIkSSrKRlOSJEmSVJSNpiRJkiSpKBtNSZIkSVJRNpqSJEmSpKKmFroASZJUh5tuumnoMXfddVendR111FGdxq1Zs6bTOEnS7q677rqhx9x///0DzecRTUmSJElSUTaakiRJkqSibDQlSZIkSUWN1GhGxKqIuCQiboyITRHxnFKFSerObEp1MptSncymVN6oFwP6Q+BvMvMlEfF44IACNUkandmU6mQ2pTqZTamwzo1mRKwEfhp4JUBmPgQ8VKYsSV2ZTalOZlOqk9mUxmOUU2efDMwAH4qIr0XEByJi+eyZIuKciNgYERtnZmZGWJ2kAZlNqU5mU6qT2ZTGYJRGcwp4JvCnmfnjwPeBc2fPlJkXZOZ0Zk6vXbt2hNVJGpDZlOpkNqU6mU1pDEZpNDcDmzPz6vbnS2hCKmlhmU2pTmZTqpPZlMagc6OZmduAOyLi6PauE4B/LlKVpM7MplQnsynVyWxK4zHqVWf/G/Cx9upctwKvGr0kSQWYTalOZlOqk9mUChup0czMa4HpQrVIKsRsSnUym1KdzKZU3qhHNCVJUmXuv//+TuO2bds29JgdO3Z0WlfXGiVJ5Zx44olDj7nnnnsGmm+UiwFJkiRJkrQbG01JkiRJUlE2mpIkSZKkomw0JUmSJElF2WhKkiRJkoqy0ZQkSZIkFWWjKUmSJEkqykZTkiRJklSUjaYkSZIkqSgbTUmSJElSUTaakiRJkqSibDQlSZIkSUXZaEqSJEmSippa6AIkSVJZ+++/f6dxP/qjPzr0mO3bt3da19q1azuNkyTt7m1ve1uncTMzM4UreZRHNCVJkiRJRdloSpIkSZKKstGUJEmSJBU1UqMZEW+KiBsi4vqI+HhELCtVmKTuzKZUJ7Mp1clsSuV1bjQj4nDgV4DpzHw6sA9wZqnCJHVjNqU6mU2pTmZTGo9RT52dAvaPiCngAGDL6CVJKsBsSnUym1KdzKZUWOdGMzP/DTgf+BawFdiemZfPni8izomIjRGxcZyXz5XUMJtSncymVCezKY3HKKfOHgycDjwJWA8sj4izZs+XmRdk5nRmTvuZWdL4mU2pTmZTqpPZlMZjlFNnTwT+NTNnMvNh4NPAc8uUJWkEZlOqk9mU6mQ2pTEYpdH8FvDsiDggIgI4AdhUpixJIzCbUp3MplQnsymNwSjv0bwauAS4BvhGu6wLCtUlqSOzKdXJbEp1MpvSeEyNMjgzfwv4rUK1SCrEbEp1MptSncymVN6oH28iSZIkSdJjjHREU5Ik7T3WrVs3kTGSpLIeeuihTuPOO++8ocd86EMfGmg+j2hKkiRJkoqy0ZQkSZIkFWWjKUmSJEkqykZTkiRJklSUjaYkSZIkqSgbTUmSJElSUTaakiRJkqSibDQlSZIkSUXZaEqSJEmSirLRlCRJkiQVZaMpSZIkSSrKRlOSJEmSVJSNpiRJkiSpqKmFLmAQ27Zt6zTu0EMPLVyJJEmSJNXlXe9618TWdfnllw80n0c0JUmSJElF2WhKkiRJkoqy0ZQkSZIkFTVvoxkRH4yIOyPi+p77VkfEFRHxzfbrweMtU9JsZlOqk9mU6mQ2pcka5IjmRcDJs+47F7gyM58CXNn+LGmyLsJsSjW6CLMp1egizKY0MfM2mpn5ZeDuWXefDlzcfn8xcEbhuiTNw2xKdTKbUp3MpjRZXd+jeUhmbgVov66ba8aIOCciNkbExpmZmY6rkzQgsynVyWxKdTKb0piM/WJAmXlBZk5n5vTatWvHvTpJAzKbUp3MplQnsykNp2uj+e2IOAyg/XpnuZIkjcBsSnUym1KdzKY0Jl0bzc8CZ7ffnw18pkw5kkZkNqU6mU2pTmZTGpNBPt7k48BXgaMjYnNEvAZ4F/AzEfFN4GfanyVNkNmU6mQ2pTqZTWmypuabITNfPsekEwrXImkIZlOqk9mU6mQ2pcka+8WAJEmSJElLy7xHNEt64IEH2LRp09Djbr755k7rO+200zqNkyRJdbr33ns7jdu5c+fQY1atWtVpXZIkj2hKkiRJkgqz0ZQkSZIkFWWjKUmSJEkqykZTkiRJklSUjaYkSZIkqSgbTUmSJElSUTaakiRJkqSibDQlSZIkSUXZaEqSJEmSirLRlCRJkiQVZaMpSZIkSSrKRlOSJEmSVNTUpFeYmRMZI0lavLZs2dJp3Pr16wtXotqsWLFioUuQJA3AI5qSJEmSpKJsNCVJkiRJRdloSpIkSZKKmrfRjIgPRsSdEXF9z33vjYgbI+K6iPiriFg13jIlzWY2pTqZTalOZlOarEGOaF4EnDzrviuAp2fmjwE3AecVrkvS/C7CbEo1ugizKdXoIsymNDHzNpqZ+WXg7ln3XZ6ZO9sfrwKOGENtkvbAbEp1MptSncymNFkl3qP5auCyuSZGxDkRsTEiNt59991zzSapvIGzOTMzM8GypCXPbEp1MptSQSM1mhHxFmAn8LG55snMCzJzOjOnV69ePcrqJA1o2GyuXbt2csVJS5jZlOpkNqXyproOjIizgRcCJ2RmlitJ0ijMplQnsynVyWxK49Gp0YyIk4HfAI7NzPvKliSpK7Mp1clsSnUym9L4DPLxJh8HvgocHRGbI+I1wB8DBwJXRMS1EfH+MdcpaRazKdXJbEp1MpvSZM17RDMzX97n7gvHUIukIZhNqU5mU6qT2ZQmq8RVZyVJkiRJ+qHOFwPqYtmyZRxzzDFDjzvyyCPHUI0kqVbr169f6BIkSdIIPKIpSZIkSSrKRlOSJEmSVJSNpiRJkiSpKBtNSZIkSVJRNpqSJEmSpKJsNCVJkiRJRdloSpIkSZKKstGUJEmSJBVloylJkiRJKspGU5IkSZJUlI2mJEmSJKkoG01JkiRJUlE2mpIkSZKkoiIzJ7eyiBng9jkmrwHumlgxc6ulDqinFuvYXZdanpiZa8dRzKjM5lBqqQPqqaWWOsBsLgTr2F0ttSz2OszmaKxjd7XUstjrGCibE2009yQiNmbmtHU8qpZarGN3NdUybrU8VuvYXS211FIH1FXLuNXyWK1jd7XUYh0Lo5bHax27q6WWpVKHp85KkiRJkoqy0ZQkSZIkFVVTo3nBQhfQqqUOqKcW69hdTbWMWy2P1Tp2V0sttdQBddUybrU8VuvYXS21WMfCqOXxWsfuaqllSdRRzXs0JUmSJEl7h5qOaEqSJEmS9gI2mpIkSZKkoibeaEbEyRHxLxFxc0Sc22f6fhHxyXb61RGxYQw1HBkRfxsRmyLihoh4Q595jouI7RFxbXt7W+k62vXcFhHfaNexsc/0iIg/arfHdRHxzDHVcXTPY702InZExBtnzTOWbRIRH4yIOyPi+p77VkfEFRHxzfbrwXOMPbud55sRcfaYanlvRNzYbv+/iohVc4zd43NZsxpy2a7HbO6+niWfzaWaSzCbc9RiNs3mgjObfWtZ8GwuZC7bZZvNXpk5sRuwD3AL8GTg8cDXgWNmzfM64P3t92cCnxxDHYcBz2y/PxC4qU8dxwGfm8A2uQ1Ys4fppwKXAQE8G7h6Qs/TNpoPYx37NgF+GngmcH3Pfe8Bzm2/Pxd4d59xq4Fb268Ht98fPIZaTgKm2u/f3a+WQZ7LWm+15LJdttmc/7lactlcirnseb7N5pDPqdk0mxN6vs3mkM/ppLM56Vy2yzabPbdJH9H8SeDmzLw1Mx8CPgGcPmue04GL2+8vAU6IiChZRGZuzcxr2u+/B2wCDi+5joJOBz6cjauAVRFx2JjXeQJwS2bePub1AJCZXwbunnV37+vgYuCMPkNfAFyRmXdn5j3AFcDJpWvJzMszc2f741XAEaOso0JV5BLM5gCWZDaXaC7BbHZlNs3muJnNbiadzYnmEszmbJNuNA8H7uj5eTO7h+GH87QbYzvwhHEV1J7K8OPA1X0mPycivh4Rl0XE08ZUQgKXR8Q/RcQ5faYPss1KOxP4+BzTJrFNAA7JzK3Q/BIF1vWZZyG2zatp9sb1M99zWavqcglmcw5ms7+9MZdgNudiNvszm5NjNvurLZs15BKWcDanRhncQb89ObM/X2WQeYqIiBXAp4A3ZuaOWZOvoTnUfm9EnAr8NfCUMZTxvMzcEhHrgCsi4sZ2L8QPy+wzZmyfSRMRjwdeBJzXZ/KktsmgJr1t3gLsBD42xyzzPZe1qiqXYDb7MZtzrGjvzSWYzbmYze7MZhlms79qsrnIcgl7aTYnfURzM3Bkz89HAFvmmicipoCD2P0Q9MgiYl+aQH4sMz89e3pm7sjMe9vvLwX2jYg1pevIzC3t1zuBv6I5HaPXINuspFOAazLz27MnTGqbtL6963SK9uudfeaZ2LZp35T9QuAVmdk3+AM8l7WqJpft8s1mf2Zzlr08l2A2+zKbczKbk2M2+6gsm7XkEpZwNifdaP4j8JSIeFK7p+FM4LOz5vkssOtKSy8BvjjXhuiqPUf+QmBTZv7eHPMcuutc+oj4SZpt9Z3CdSyPiAN3fU/zJt3rZ832WeAXovFsYPuuw+9j8nLmOM1gEtukR+/r4GzgM33m+QJwUkQcHM0VvE5q7ysqIk4GfgN4UWbeN8c8gzyXtaoil2A252E2eyyBXILZ7LcOszk3szk5ZnP3ddSWzVpyCUs5mznGqz31u9Fcceommqt1vaW97x3tgwZYBvwlcDPwD8CTx1DDT9Ecjr4OuLa9nQq8FnhtO8/rgRtoriR2FfDcMdTx5Hb5X2/XtWt79NYRwPva7fUNYHqMz80BNEE7qOe+sW8Tml8EW4GHafbovIbmfQxXAt9sv65u550GPtAz9tXta+Vm4FVjquVmmvPmd71Wdl1Fbj1w6Z6ey8VyqyGX7XrMZv96lnQ2l2ou28fPwCNNAAAAZElEQVRgNh9bh9lMs1nDzWzuVkc12VyoXLbLNps9t2gXKkmSJElSEZM+dVaSJEmStJez0ZQkSZIkFWWjKUmSJEkqykZTkiRJklSUjaYkSZIkqSgbTUmSJElSUTaakiRJkqSi/j90/80ampORgQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# passing average image for each number to one of the CNN filters and plotting the output\n",
    "visualize_filter(model, X_test, y_test, 2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA5oAAADvCAYAAACT8qtkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAHfNJREFUeJzt3X+0ZWdZH/Dv00wAgWAIGQIkKWMosACXAo4WxUoKCAkisdWyQkER0BRZVOhCJRQsFWNFZUlrS6UpwVBAQPlRWBCQLBCBCtExJEAMhIAJhCRkSCQkRIEkb//Ye8jJnXtn7j3nPffumfl81jprzj17v3s/Z5/zvXeevffZp1prAQAAgF7+yVYXAAAAwMFFowkAAEBXGk0AAAC60mgCAADQlUYTAACArjSaAAAAdKXR3EJV9ciq+lxV3VhVP7XK9AdW1Seq6oaq+uWqenVV/fo47cSqumLzq16f2VrhQCObME2yCdN0kGfzvVX19K2u44DUWjtgbkl+PsmnktyU5Ookf5jkyA2MvyzJYzvWs9DyknwgyfP2Mf2sJK9cY9qJSa5Y1nMbl/ncJLuSfDPJ2Vv9+u/jPfHRJS37qCTvSPKNJJcn+bdb/XynepPN201bajaT3HFc/+VJbkjyiSQnb/V7YI33hGxO43WQzbZpfzffkOSqJF9PckmSX9jq98Aa7wnZnMbrIJttc7I5s+z7J/nHJG/Y6vfAKrX952XVlWRHkj8f32+fWdb2XXk7YI5oVtULkvxOkl9N8t1JHpHkvknOrao7bGVtC7hvkosWmN5FDVZ7L1yZ5Iwkr112DRP1qiTfSnJMkqcm+cOqesjWljQ9srk8a2RzW5IvJXlUhu3960n+pKp2LLueCZHNdZDN5dnH383fTrKjtXa3JE9KckZV/cCy65kQ2VwH2VyefWRzj1cl+etl1zFBb8qwY/oeSV6c5K1VtX3pa93q7n2dXfjdktyY5MkrHr9rkmuSPHP8+ewkZ8xMPzHjHpIkr09ya5J/GJf1axm6+5bktAxN1VVJXjAzfkPLW6P2X0xyaZLrkrwryX3Gxz+/YvwdV4z7YJJbMux1uTHJA2brWU8tGX5x/WWSryW5MMmJM8v/UJLfSvL/xnH/bB/b/4xs8IjmarUmecH4el2V5Bkr5n11knMzHKH5iyT3ndkD05JsW1H7LyR50Lh9bhmf99c6vufukuGP5QNmHnt9kpdvdR6mdItsbmk2Z8Z8MslPr/M126vWyOZBd4tsbnk2kzxw3D5P3td8q227yOZBe4tsblk2k5ya5E+ywSOHe/Iz3v/5JB9N8ookf5/k7zJzVtE4728n+ask1yd5Z5KjVj7PmfkvS/LYJCeN+fn2+Lwv7Piee0CGsxOPmHnsI0mevez3+4FyRPNHktwpydtnH2yt3ZjkvUl+fH8LaK39bJIvJvnJ1tpdW2u/OzP5X2Y4lP64JKdX1WMXXF6SpKoeneHN9uQk985wGsmbx/H3WzH+myuW/+gMb4LnjtMv2UgtVXVskvdkaBKPSvIrSd62Yu/Fz2b4hXTEWNsy3SvDXrtjkzwryauq6u4z05+a5DeTHJ3kgiRv3N8CW2sXJ3l2ko+Nz/vI1earqv9ZVV9b4/bJNRb/gCS3rNjuFyaxZ/b2ZHOLs1lVx2R4v867p1g2D06yuUXZHN/Xe05PuyrJOfvZNGuRzYOTbG5BNqvqbklelmHnzaL+eZLPZsje7yY5q6pqZvrPJXlmkvskuTnJH+xvga219yX5L0neMj7v719tvqp69z6y+e41Fv+QJF9ord0w89imZPNAaTSPTvLV1trNq0y7apy+iN9orX2jtfapJH+U5CkLLm+PpyZ5bWvt/DF0L0ryw5t0itvTkpzTWjuntXZra+3cDJ+3fMLMPGe31i5qrd3cWvv2kuv5dpKXtda+3Vo7J8PemgfOTH9Pa+3D43Z6cYbtdHyPFbfWntNaO3KN2/etMeyuGfZEzbo+wy8wbiObG9ctm1V1eIb/XL6utfaZOeuRzYOTbG5cl2y21p6T4f34LzI0E99cbb51kM2Dk2xuXI9s/maSs1prX+pQz+Wttf/dWrslyesyNN7HzEx/fWvt0621b2T4eMuTq+qwDutNa+2J+8jmE9cYtmXZPFAaza8mObqqtq0y7d7j9EXMvukuz7AHoof7ZGavyri36toMeyeX7b5J/s3sno4kP5phe+3RI2zrde2KX6o3ZXjj71XLuJ2uS7/XYR43Zji9ZdbdMpyixG1kc+O6ZHP8DMrrM5xq89wF6pHNg5Nsbly3v5uttVtaax9NclySX5qzHtk8OMnmxi2Uzap6aIbTU1/ZqZ6r99xprd003l01mxm22eFZfAfCIrYsmwdKo/mxDHsE//Xsg1V1lyQnZ7jSVTJc5ezOM7Pca8Vy2hrLn90D+E8znNu+yPL2uDJDOGbrvUeSL+9n3DxW1vKlDHtUZvd03KW19vJ9jNlK33kNququGU6NuDLDa5Cs/Trs9zmMl9C+cY3bWqcbXpJkW1Xdf+ax788mfJD9ACOb+9c9m+MpOmdl2IP600s+I0E2D0yyuX+b8XdzW5L7LVTl2mTzwCSb+9c7mydm+AzrF6vq6gyn3v50VZ3fs+gZK1+Db2fYgXC712A8yjl7+u96svnefWTzvWsMuyjJCVU1ewRzU7J5QDSarbXrk/xGkv9eVSdV1eHjofo/zfBh+dePs16Q5AlVdVRV3SvJ81cs6itJTlhlFb9eVXeu4cpoz0jylgWXt8cfJ3lGVT20qu6Y4dzr81prl+33SW/cylrekOQnq+rxVXVYVd2phu8pOm69C6yqbVV1pySHJdmzjG0z01tVndip/idU1Y/WcLW138ywnb7UWtud4ZfY08bn8czc/o/2V5IcV/u4Sltr7dltON99tduq56ePpzu8PcnLquouVfXIJKfktvcakc116p7NDJfBf1CGz7D8w8qJsolsrkvXbFbVPavq1Kq66zj+8RlOW/zgzDyyeYiTzXXp/XfzzAwZeOh4e3WGz3w+PkmqaseYzR2d6n9aVT24qu6c4XOhbx1Ps70kyZ2q6idq+OjLSzJ8ZdkeX0myo/Zx1dzW2sn7yObJa4y5JMPr/9Jx2/2rJN+X5G19nu7aDohGM0na8MHk/5jhKk9fT3Jehj0cj2m3fej49Rk+3HpZkvfntnDt8dtJXlLDYfdfmXn8LzJcResDSV7RWnv/gsvbU/MHMpyb/bYM593fL8MVr5bhdrWM56CfkmGb7c6wrX41G3vNX5Lh6l2nZzg//h/GxzKG+8YM3wHVwx8neWmGU39+IMNnAfb4xQy1X5vhg8t/OTPtgxn2yFxdVYuebrLSc5J8V4arwL0pyS+11uyZXUE296trNqvqvkn+XYY/llfXbXsynzpOl02SyOY69P672TKcJntFhqtRviLJ81tr70xkk9vI5n51zWZr7abW2tV7bhly+I/jTplkOAJ5efodnX19hqvqXp3hwk+/PNZxfYaMvGZc1zcy/L7Y40/Hf6+t/kdbT02yM8Pvppcn+ZmZ57801dqUzp7cXOOei79Lcnhb/UPZrKGqnpbkIa21F3VY1tkZLvf8koUL46Agm/OTTZZJNucnmyyTbM6vql6SZHdr7X91WNaHMnx1ymsWLuwgsNoHkWG/Wmtv2OoagL3JJkyTbMI0tdbO2OoaDlYHzKmzAAAAHBgO6VNnAQAA6M8RTQAAALra1M9oHn300W3Hjh2buUoOQLfeeutc4+Y5On/YYYfNta55XHbZZfnqV79am7bCDZBNDhZf/vLGLxp4/fXX56abbpJNmBh/N2Ga1pvNTW00d+zYkV27dm3mKjkA3XDDDXONu+WWWzY85sgjj5xrXfPYuXPnpq1ro2STg8VLXrLxi3CeddZZS6ikD9nkUObvJkzTerPp1FkAAAC60mgCAADQ1UKNZlWdVFWfrapLq+r0XkUBi5FNmCbZhGmSTehv7kazqg5L8qokJyd5cJKnVNWDexUGzEc2YZpkE6ZJNmE5Fjmi+UNJLm2tfaG19q0kb05ySp+ygAXIJkyTbMI0ySYswSKN5rFJvjTz8xXjY7dTVadV1a6q2rV79+4FVgesk2zCNMkmTJNswhIs0miu9t0pe32RYWvtzNbaztbazu3bty+wOmCdZBOmSTZhmmQTlmCRRvOKJMfP/HxckisXKwfoQDZhmmQTpkk2YQkWaTT/Osn9q+p7quoOSU5N8q4+ZQELkE2YJtmEaZJNWIJt8w5srd1cVc9N8mdJDkvy2tbaRd0qA+YimzBNsgnTJJuwHHM3mknSWjsnyTmdagE6kU2YJtmEaZJN6G+RU2cBAABgLwsd0YRlOOKII7a6BOAAdcYZZ2x4zPve974lVAIAhzZHNAEAAOhKowkAAEBXGk0AAAC60mgCAADQlUYTAACArjSaAAAAdKXRBAAAoCuNJgAAAF1pNAEAAOhKowkAAEBXGk0AAAC60mgCAADQlUYTAACArrZtdQEAwIHrmmuumWvcN7/5zbnGHX/88XONA2BzOaIJAABAVxpNAAAAutJoAgAA0NXcjWZVHV9Vf15VF1fVRVX1vJ6FAfORTZgm2YRpkk1YjkUuBnRzkhe01s6vqiOS/E1Vndta+9tOtQHzkU2YJtmEaZJNWIK5j2i21q5qrZ0/3r8hycVJju1VGDAf2YRpkk2YJtmE5ejyGc2q2pHkYUnOW2XaaVW1q6p27d69u8fqgHWSTZgm2YRpkk3oZ+FGs6rumuRtSZ7fWvv6yumttTNbaztbazu3b9++6OqAdZJNmCbZhGmSTehroUazqg7PEMg3ttbe3qckYFGyCdMkmzBNsgn9LXLV2UpyVpKLW2u/368kYBGyCdMkmzBNsgnLscgRzUcm+dkkj66qC8bbEzrVBcxPNmGaZBOmSTZhCeb+epPW2keTVMdagA5kE6ZJNmGaZBOWY5Hv0QQADiK33HLLhsdceeWVc63rqKOOmmscAAeGLl9vAgAAAHtoNAEAAOhKowkAAEBXGk0AAAC60mgCAADQlUYTAACArjSaAAAAdKXRBAAAoCuNJgAAAF1pNAEAAOhKowkAAEBXGk0AAAC60mgCAADQ1batLgAAmIbDDjtsw2Me+tCHLqESAA50jmgCAADQlUYTAACArjSaAAAAdLVwo1lVh1XVJ6rq3T0KAvqQTZgm2YRpkk3oq8cRzeclubjDcoC+ZBOmSTZhmmQTOlqo0ayq45L8RJLX9CkH6EE2YZpkE6ZJNqG/RY9o/tckv5bk1rVmqKrTqmpXVe3avXv3gqsD1kk2YZpkE6ZJNqGzuRvNqnpikmtaa3+zr/laa2e21na21nZu37593tUB6ySbME2yCdMkm7AcixzRfGSSJ1XVZUnenOTRVfWGLlUBi5BNmCbZhGmSTViCuRvN1tqLWmvHtdZ2JDk1yQdba0/rVhkwF9mEaZJNmCbZhOXwPZoAAAB0ta3HQlprH0ryoR7LAvqRTZgm2YRpkk3oxxFNAAAAutJoAgAA0JVGEwAAgK40mgAAAHSl0QQAAKArjSYAAABdaTQBAADoSqMJAABAVxpNAAAAutJoAgAA0JVGEwAAgK40mgAAAHSl0QQAAKArjSYAAABdaTQBAADoSqMJAABAVxpNAAAAutJoAgAA0JVGEwAAgK4WajSr6siqemtVfaaqLq6qH+5VGDA/2YRpkk2YJtmE/rYtOP6/JXlfa+1nquoOSe7coSZgcbIJ0ySbME2yCZ3N3WhW1d2S/FiSn0+S1tq3knyrT1nAvGQTpkk2YZpkE5ZjkVNnT0iyO8kfVdUnquo1VXWXlTNV1WlVtauqdu3evXuB1QHrJJswTbIJ0ySbsASLNJrbkjw8yR+21h6W5BtJTl85U2vtzNbaztbazu3bty+wOmCdZBOmSTZhmmQTlmCRRvOKJFe01s4bf35rhpACW0s2YZpkE6ZJNmEJ5m40W2tXJ/lSVT1wfOgxSf62S1XA3GQTpkk2YZpkE5Zj0avO/vskbxyvzvWFJM9YvCSgA9mEaZJNmCbZhM4WajRbaxck2dmpFqAT2YRpkk2YJtmE/hb5jCYAAADsRaMJAABAVxpNAAAAutJoAgAA0JVGEwAAgK40mgAAAHSl0QQAAKArjSYAAABdaTQBAADoSqMJAABAVxpNAAAAutJoAgAA0JVGEwAAgK40mgAAAHSl0QQAAKArjSYAAABdaTQBAADoSqMJAABAVws1mlX1H6rqoqr6dFW9qaru1KswYH6yCdMkmzBNsgn9zd1oVtWxSX45yc7W2vcmOSzJqb0KA+YjmzBNsgnTJJuwHIueOrstyXdV1bYkd05y5eIlAR3IJkyTbMI0ySZ0Nnej2Vr7cpJXJPlikquSXN9ae//K+arqtKraVVW7du/ePX+lwLrIJkyTbMI0ySYsxyKnzt49ySlJvifJfZLcpaqetnK+1tqZrbWdrbWd27dvn79SYF1kE6ZJNmGaZBOWY5FTZx+b5O9aa7tba99O8vYkP9KnLGABsgnTJJswTbIJS7BIo/nFJI+oqjtXVSV5TJKL+5QFLEA2YZpkE6ZJNmEJFvmM5nlJ3prk/CSfGpd1Zqe6gDnJJkyTbMI0ySYsx7ZFBrfWXprkpZ1qATqRTZgm2YRpkk3ob9GvNwEAAIDb0WgCAADQlUYTAACArjSaAAAAdKXRBAAAoCuNJgAAAF1pNAEAAOhKowkAAEBXGk0AAAC60mgCAADQlUYTAACArjSaAAAAdKXRBAAAoCuNJgAAAF1pNAEAAOhKowkAAEBXGk0AAAC60mgCAADQlUYTAACArvbbaFbVa6vqmqr69MxjR1XVuVX1ufHfuy+3TGAl2YRpkk2YJtmEzbWeI5pnJzlpxWOnJ/lAa+3+ST4w/gxsrrMjmzBFZ0c2YYrOjmzCptlvo9la+3CS61Y8fEqS1433X5fkpzrXBeyHbMI0ySZMk2zC5pr3M5rHtNauSpLx33uuNWNVnVZVu6pq1+7du+dcHbBOsgnTJJswTbIJS7L0iwG11s5sre1sre3cvn37slcHrJNswjTJJkyTbMLGzNtofqWq7p0k47/X9CsJWIBswjTJJkyTbMKSzNtovivJ08f7T0/yzj7lAAuSTZgm2YRpkk1YkvV8vcmbknwsyQOr6oqqelaSlyf58ar6XJIfH38GNpFswjTJJkyTbMLm2ra/GVprT1lj0mM61wJsgGzCNMkmTJNswuZa+sWAAAAAOLTs94gmwBRU1VzjWmudKwEAYH8c0QQAAKArjSYAAABdaTQBAADoSqMJAABAVxpNAAAAutJoAgAA0JVGEwAAgK40mgAAAHSl0QQAAKArjSYAAABdaTQBAADoSqMJAABAV9u2uoBluu666zY85sILL5xrXZ/97GfnGnf55ZfPNe5rX/vahsccfvjhc63r2GOPnWvcQx7ykLnG/eAP/uBc44455pi5xnFg+PznP7/VJQAs1cc+9rG5xl177bVzjbvf/e634TEPetCD5loXcOhxRBMAAICuNJoAAAB0pdEEAACgq/02mlX12qq6pqo+PfPY71XVZ6rqk1X1jqo6crllAivJJkyTbMI0ySZsrvUc0Tw7yUkrHjs3yfe21r4vySVJXtS5LmD/zo5swhSdHdmEKTo7sgmbZr+NZmvtw0muW/HY+1trN48/fjzJcUuoDdgH2YRpkk2YJtmEzdXjM5rPTPLetSZW1WlVtauqdu3evbvD6oB1kk2YJtmEaZJN6GihRrOqXpzk5iRvXGue1tqZrbWdrbWd27dvX2R1wDrJJkyTbMI0ySb0t23egVX19CRPTPKY1lrrVxKwCNmEaZJNmCbZhOWYq9GsqpOSvDDJo1prN/UtCZiXbMI0ySZMk2zC8qzn603elORjSR5YVVdU1bOS/I8kRyQ5t6ouqKpXL7lOYAXZhGmSTZgm2YTNtd8jmq21p6zy8FlLqAXYANmEaZJNmCbZhM3V46qzAAAA8B1zXwzoQPCRj3xkw2Pe8573bNq6kuSSSy6Za9ytt9664TH3ute95lrXwx72sLnGzeuYY46Za9w97nGPDY/Ztu2gjsBB5YQTTtjqEgDW5ZWvfOVc417+8pfPNe6aa66Za9yjHvWoDY954QtfONe6Tj755LnGAQcuRzQBAADoSqMJAABAVxpNAAAAutJoAgAA0JVGEwAAgK40mgAAAHSl0QQAAKArjSYAAABdaTQBAADoSqMJAABAVxpNAAAAutJoAgAA0JVGEwAAgK6qtbZ5K6vaneTyNSYfneSrm1bM2qZSRzKdWtSxt3lquW9rbfsyilmUbG7IVOpIplPLVOpIZHMrqGNvU6nlQK9DNhejjr1NpZYDvY51ZXNTG819qapdrbWd6rjNVGpRx96mVMuyTeW5qmNvU6llKnUk06pl2abyXNWxt6nUoo6tMZXnq469TaWWQ6UOp84CAADQlUYTAACArqbUaJ651QWMplJHMp1a1LG3KdWybFN5rurY21RqmUodybRqWbapPFd17G0qtahja0zl+apjb1Op5ZCoYzKf0QQAAODgMKUjmgAAABwENJoAAAB0temNZlWdVFWfrapLq+r0VabfsareMk4/r6p2LKGG46vqz6vq4qq6qKqet8o8J1bV9VV1wXj7T73rGNdzWVV9alzHrlWmV1X9wbg9PllVD19SHQ+cea4XVNXXq+r5K+ZZyjapqtdW1TVV9emZx46qqnOr6nPjv3dfY+zTx3k+V1VPX1Itv1dVnxm3/zuq6sg1xu7ztZyyKeRyXI9s7r2eQz6bh2ouE9lcoxbZlM0tJ5ur1rLl2dzKXI7Lls1ZrbVNuyU5LMnnk5yQ5A5JLkzy4BXzPCfJq8f7pyZ5yxLquHeSh4/3j0hyySp1nJjk3ZuwTS5LcvQ+pj8hyXuTVJJHJDlvk16nqzN8GevSt0mSH0vy8CSfnnnsd5OcPt4/PcnvrDLuqCRfGP+9+3j/7kuo5XFJto33f2e1WtbzWk71NpVcjsuWzf2/VodcNg/FXM683rK5wddUNmVzk15v2dzga7rZ2dzsXI7Lls2Z22Yf0fyhJJe21r7QWvtWkjcnOWXFPKcked14/61JHlNV1bOI1tpVrbXzx/s3JLk4ybE919HRKUn+Txt8PMmRVXXvJa/zMUk+31q7fMnrSZK01j6c5LoVD8++D16X5KdWGfr4JOe21q5rrf19knOTnNS7ltba+1trN48/fjzJcYusY4ImkctENtfhkMzmIZrLRDbnJZuyuWyyOZ/Nzuam5jKRzZU2u9E8NsmXZn6+InuH4TvzjBvj+iT3WFZB46kMD0ty3iqTf7iqLqyq91bVQ5ZUQkvy/qr6m6o6bZXp69lmvZ2a5E1rTNuMbZIkx7TWrkqGX6JJ7rnKPFuxbZ6ZYW/cavb3Wk7V5HKZyOYaZHN1B2MuE9lci2yuTjY3j2yubmrZnEIuk0M4m9sWGTyH1fbkrPx+lfXM00VV3TXJ25I8v7X29RWTz89wqP3GqnpCkv+b5P5LKOORrbUrq+qeSc6tqs+MeyG+U+YqY5b2nTRVdYckT0ryolUmb9Y2Wa/N3jYvTnJzkjeuMcv+XsupmlQuE9lcjWyusaKDN5eJbK5FNucnm33I5uomk80DLJfJQZrNzT6ieUWS42d+Pi7JlWvNU1Xbknx39j4EvbCqOjxDIN/YWnv7yumtta+31m4c75+T5PCqOrp3Ha21K8d/r0nyjgynY8xazzbr6eQk57fWvrJywmZtk9FX9pxOMf57zSrzbNq2GT+U/cQkT22trRr8dbyWUzWZXI7Ll83VyeYKB3kuE9lclWyuSTY3j2yuYmLZnEouk0M4m5vdaP51kvtX1feMexpOTfKuFfO8K8meKy39TJIPrrUh5jWeI39Wkotba7+/xjz32nMufVX9UIZtdW3nOu5SVUfsuZ/hQ7qfXjHbu5L8XA0ekeT6PYffl+QpWeM0g83YJjNm3wdPT/LOVeb5sySPq6q713AFr8eNj3VVVScleWGSJ7XWblpjnvW8llM1iVwmsrkfsjnjEMhlIpurrUM21yabm0c2917H1LI5lVwmh3I22xKv9rTaLcMVpy7JcLWuF4+PvWx80klypyR/muTSJH+V5IQl1PCjGQ5HfzLJBePtCUmeneTZ4zzPTXJRhiuJfTzJjyyhjhPG5V84rmvP9pito5K8atxen0qyc4mvzZ0zBO27Zx5b+jbJ8IvgqiTfzrBH51kZPsfwgSSfG/89apx3Z5LXzIx95vheuTTJM5ZUy6UZzpvf817ZcxW5+yQ5Z1+v5YFym0Iux/XI5ur1HNLZPFRzOT4H2bx9HbLZZHMKN9ncq47JZHOrcjkuWzZnbjUuFAAAALrY7FNnAQAAOMhpNAEAAOhKowkAAEBXGk0AAAC60mgCAADQlUYTAACArjSaAAAAdPX/AdJlR9+837nUAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA5oAAADvCAYAAACT8qtkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAHhNJREFUeJzt3X20ZXdZH/Dv00xIIAkkwwwveWlGKLAKKgFG5U1NASEgEle1rLBAEdCUslKhC19CwVqVii8sbW2pNAUMDQgooFAMSFYAKS1ExyQEYhAC5s0kZJJIQhIwJPz6x9kDN3funXvvOb9z7p6Zz2ets+65Z+/f3s/Z53znzrP3PvtUay0AAADQyz/Z7AIAAAA4sGg0AQAA6EqjCQAAQFcaTQAAALrSaAIAANCVRhMAAICuNJqbqKqeVFVfqKrbqupHV5j+iKq6qKq+WlU/W1VvrKpfGqadXFXXLL7q9VlaK+xvZBPGSTZhnA7wbH6wql642XXsl1pr+80tyU8l+UySO5Jcn+T3kxy9gfFXJHlax3pmWl6S85O8fB/T35zkd1eZdnKSa+b13IZlnpFkV5J/THL2Zr/++3hPfGIOyz1s2P5XJvlqkouSPHOzn+9Yb7J5j2lzzeb+8t6UzXHcZPMe0xbxd/NtSa5LcmuSzyf56c1+D6zynpDNcbwOstkWk80ly35Ykq8nedtmvwdWqO0/zqOuJA9I8o4k1ya5Jcn/TfJ9i3hO+80Rzap6ZZLfTPLzSe6X5PFJTkxyXlXdazNrm8GJSS6dYXoXNbHSe+HaJK9N8pZ51zBCW5JcneQHM3m//VKSP6qqHZtY0yjJ5vysks2D/b15sD//dZPN+dnH383XJdnRWrtvkuckeW1VPW7e9YyEbK6TbM7PPrK5xxuS/NW86xiZIzN5zo9LsjXJW5P8WVUdOfc1b3b3vs5O/L5Jbkvy3GWPH5nkhiQvHn4/O8lrl0w/OcMekiTnJPlmkq8Ny/qFJDuStCSnZ9JUXZfklUvGb2h5q9T+M0kuT3JzkvcnOXZ4/IvLxh+2bNxHktydyV6X25I8fGk966klk3+4/l+SryT5dJKTlyz/Y0n+UyZ7Nb6W5J/tY/u/Nhs8orlSrUleObxe1yV50bJ535jkvEz2gv5FkhOHaXteoy3Lav/pJP982D53D8/7K3N+H16S5Mc2Ow9jusnm5mZzmvembB4cN9nc/GwmecSwfZ67r/lW2nayeeDeZHPzspnktCR/lA0eOdyTn+H+TyX5RJLXJ/mHJH+XJUfuh3lfl+QvMzl6+L4kW5c/zyXzX5HkaUlOSXJnkm8Mz/vTc34f3prkcfN+v+8vRzSfmOTwJO9d+mBr7bYkH0zyQ2stoLX2E0muSvIjrbUjW2u/tWTyv8jkUPrTk5xZVU+bcXlJkqp6SiZvtucmeXAmp5O8cxj/0GXj/3HZ8p+S5P8kOWOY/vmN1FJVxyX5s0yaxK1Jfi7Je6pq+5KhP5HJP0hHDbXN04My2Wt3XJKXJHlDVR2zZPrzk/xakm1JLk7y9rUW2Fq7LMlLk3xyeN5HrzRfVf33qvrKKrdL1lN8VT0wk38Y5743bj8jm5uczQ7vTdk8MMnmJmVzeF/fkeRzmfxn/9w1Ns1qZPPAJJubkM2qum+SX81k582svi/J32aSvd9K8uaqqiXTfzLJi5Mcm+SuJL+31gJbax9K8utJ3jU870evNF9VfWAf2fzAeoqvqpOS3CuTnQZztb80mtuS3Nhau2uFadcN02fxK62121trn0nyB0meN+Py9nh+kre01i4cQveqJE9Y0GkkL0hybmvt3NbaN1tr52XyectnLZnn7Nbapa21u1pr35hzPd9I8quttW+01s7NZG/NI5ZM/7PW2seH7fTqTLbTCT1W3Fp7WWvt6FVu373W+Ko6NJM/4G9trX2uR00HENncuG7Z7PTelM0Dk2xuXJdsttZelsl/dr8/k2biH1eabx1k88AkmxvXI5u/luTNrbWrO9RzZWvtf7bW7s7kNNQHJ3ngkunntNY+21q7PZNTyJ9bVYd0WG9aa8/eRzafvdb4oeE+J5P3yS09atqX/aXRvDHJtqrassK0Bw/TZ7H0TXdlJnsgejg2S/aqDHurbspk7+S8nZjkXy3d05HkyZlsrz16hG29blr2j+odmZwmslctw3a6Of1eh6kN5/mfk8npDGdscjljJJsb1yWbHd+bsnlgks2N6/Z3s7V2d2vtE0mOT/JvpqxHNg9MsrlxM2VzOIL3tCS/26me6/fcaa3dMdxdMZuZbLNDM/sOhJlV1b2T/O8kn2qtvW4R69xfGs1PZrJH8F8ufbCqjkjyzEyudJUktye5z5JZHrRsOW2V5S/dA/hPMzm3fZbl7XFtJuFYWu/9k/z9GuOmsbyWqzPZo7J0T8cRrbXf2MeYzfSt12D4cPLWTLbf7cPDq70Oaz6HmlxC+7ZVbque0jOcBvHmTPZS/dgCjvruj2Rzbd2zueD3pmzun2RzbYv4u7klyUNnqnJ1srl/ks219c7myZl8hvWqqro+k1Nvf6yqLuxZ9BLLX4NvZLID4R6vwXCUc+npv+vJ5gf3kc0P7mPcYUn+NJPX619v7OlMb79oNIdDu7+S5L9W1SlVdehwqP6PM/mw/DnDrBcneVZVba2qByV5xbJFfTnJQ1ZYxS9V1X2q6lFJXpTkXTMub48/TPKiqjppeIF/PckFrbUr1nzSG7e8lrcl+ZGqekZVHVJVh9fke4qOX+8Cq2pLVR2e5JAke5axZcn0VlUnd6r/WVX15Jpcbe3XMtlOV7fWdmcSihcMz+PFuecf7S8nOb72cZW21tpL2+R895Vuj9pHTb+fyYUTfqS19rXZn+KBRzbXpXs2s8Z7UzaRzXXpms2qekBVnVZVRw7jn5HJaYsfWTKPbB7kZHNdev/dPCuTDJw03N6YyWc+n5EkVbVjyOaOTvW/oKoeWVX3yeRzoe8eTrP9fJLDq+qHa3J6+Wsy+VqgPb6cZEft46q5rbVn7iObz1xpzLCud2dyoaSfbK19s9PzXNN+0WgmSZt8MPnfZ3KVp1uTXJDJHo6ntm9/6PicTK5EdUWSD+fb4drjdUleU5PD7j+35PG/yOQDsecneX1r7cMzLm9Pzedncm72ezI57/6hmVzxah7uUctwDvqpmWyz3Zlsq5/Pxl7z12Typjwzk/PjvzY8liHct2XyHVA9/GGSX87k1J/HZfJZgD1+JpPab0ryqEyuOrbHRzK50MD1VTXr6SbfUlUnZrLH56Rh2Xv2Fj1/jaEHHdlcU9dsrvXelE32kM019f672TI5TfaaTK5G+fokr2itvS+RTb5NNtfUNZuttTtaa9fvuWWSw68PO2WSyRHIK9Pv6Ow5mVxV9/pMLvz0s0MdtyR5WZI3Deu6PZN/L/b44+HnTdX3aOsTkzw7kwtEfWVJNr+/4zpWVK2N6ezJxRr2XPxdkkPbyh/KZhVV9YIkj2qtvarDss7O5HLPr5m5MA4Isjk92WSeZHN6ssk8yeb0quo1SXa31v5Hh2V9LJOvTnnTzIUdAFb6IDKsqbX2ts2uAdibbMI4ySaMU2vttZtdw4Fqvzl1FgAAgP3DQX3qLAAAAP05ogkAAEBXC/2M5tatW9sJJ5yw9ozLbNkyXZlVNdU4mIcrrrgiN9544yjflNu2bWs7duzY7DJgU8gmjJNswjitN5sLbTRPOOGEfOhDH9rwuG3btk21vkMPPXSqcTAPO3fu3OwSVrVjx47s2rVrs8uATSGbME6yCeO03mw6dRYAAICuNJoAAAB0NVOjWVWnVNXfVtXlVXVmr6KA2cgmjJNswjjJJvQ3daNZVYckeUOSZyZ5ZJLnVdUjexUGTEc2YZxkE8ZJNmE+Zjmi+b1JLm+tfam1dmeSdyY5tU9ZwAxkE8ZJNmGcZBPmYJZG87gkVy/5/ZrhsXuoqtOraldV7brppptmWB2wThvO5u7duxdWHBzEZBPGSTZhDmZpNFf67pS21wOtndVa29la23n/+99/htUB67ThbG7fvn0BZcFBTzZhnGQT5mCWRvOaJCcs+f34JNfOVg7QgWzCOMkmjJNswhzM0mj+VZKHVdV3VNW9kpyW5P19ygJmIJswTrIJ4ySbMAdbph3YWrurqs5I8udJDknyltbapd0qA6YimzBOsgnjJJswH1M3mknSWjs3ybmdagE6kU0YJ9mEcZJN6G+WU2cBAABgLzMd0dyoqsphhx224XGHHnroHKoBAABgHhzRBAAAoCuNJgAAAF1pNAEAAOhKowkAAEBXGk0AAAC60mgCAADQlUYTAACArjSaAAAAdKXRBAAAoCuNJgAAAF1pNAEAAOhKowkAAEBXGk0AAAC62rLQlW3Zkq1bty5ylQAAACyYI5oAAAB0pdEEAACgK40mAAAAXU3daFbVCVX10aq6rKouraqX9ywMmI5swjjJJoyTbMJ8zHIxoLuSvLK1dmFVHZXkr6vqvNba33SqDZiObMI4ySaMk2zCHEx9RLO1dl1r7cLh/leTXJbkuF6FAdORTRgn2YRxkk2Yjy6f0ayqHUkek+SCFaadXlW7qmrX7t27e6wOWCfZhHGSTRgn2YR+Zm40q+rIJO9J8orW2q3Lp7fWzmqt7Wyt7dy+ffusqwPWSTZhnGQTxkk2oa+ZGs2qOjSTQL69tfbePiUBs5JNGCfZhHGSTehvlqvOVpI3J7mstfY7/UoCZiGbME6yCeMkmzAfsxzRfFKSn0jylKq6eLg9q1NdwPRkE8ZJNmGcZBPmYOqvN2mtfSJJdawF6EA2YZxkE8ZJNmE+ulx1FgAAAPbQaAIAANCVRhMAAICuNJoAAAB0pdEEAACgK40mAAAAXWk0AQAA6EqjCQAAQFcaTQAAALrSaAIAANCVRhMAAICuNJoAAAB0pdEEAACgK40mAAAAXWk0AQAA6EqjCQAAQFcaTQAAALrSaAIAANDVzI1mVR1SVRdV1Qd6FAT0IZswTrIJ4ySb0FePI5ovT3JZh+UAfckmjJNswjjJJnQ0U6NZVccn+eEkb+pTDtCDbMI4ySaMk2xCf7Me0fzPSX4hyTdXm6GqTq+qXVW1a/fu3TOuDlgn2YRxkk0YJ9mEzqZuNKvq2UluaK399b7ma62d1Vrb2VrbuX379mlXB6yTbMI4ySaMk2zCfMxyRPNJSZ5TVVckeWeSp1TV27pUBcxCNmGcZBPGSTZhDqZuNFtrr2qtHd9a25HktCQfaa29oFtlwFRkE8ZJNmGcZBPmw/doAgAA0NWWHgtprX0sycd6LAvoRzZhnGQTxkk2oR9HNAEAAOhKowkAAEBXGk0AAAC60mgCAADQlUYTAACArjSaAAAAdKXRBAAAoCuNJgAAAF1pNAEAAOhKowkAAEBXGk0AAAC60mgCAADQlUYTAACArjSaAAAAdKXRBAAAoCuNJgAAAF1pNAEAAOhKowkAAEBXGk0AAAC6mqnRrKqjq+rdVfW5qrqsqp7QqzBgerIJ4ySbME6yCf1tmXH8f0nyodbaj1fVvZLcp0NNwOxkE8ZJNmGcZBM6m7rRrKr7JvmBJD+VJK21O5Pc2acsYFqyCeMkmzBOsgnzMcupsw9JsjvJH1TVRVX1pqo6YvlMVXV6Ve2qql27d++eYXXAOskmjJNswjjJJszBLI3mliSPTfL7rbXHJLk9yZnLZ2qtndVa29la27l9+/YZVgesk2zCOMkmjJNswhzM0mhek+Sa1toFw+/vziSkwOaSTRgn2YRxkk2Yg6kbzdba9UmurqpHDA89NcnfdKkKmJpswjjJJoyTbMJ8zHrV2X+b5O3D1bm+lORFs5cEdCCbME6yCeMkm9DZTI1ma+3iJDs71QJ0IpswTrIJ4ySb0N8sn9EEAACAvWg0AQAA6EqjCQAAQFcaTQAAALrSaAIAANCVRhMAAICuNJoAAAB0pdEEAACgK40mAAAAXWk0AQAA6EqjCQAAQFcaTQAAALrSaAIAANCVRhMAAICuNJoAAAB0pdEEAACgK40mAAAAXWk0AQAA6GqmRrOq/l1VXVpVn62qd1TV4b0KA6YnmzBOsgnjJJvQ39SNZlUdl+Rnk+xsrX1nkkOSnNarMGA6sgnjJJswTrIJ8zHrqbNbkty7qrYkuU+Sa2cvCehANmGcZBPGSTahs6kbzdba3yd5fZKrklyX5JbW2oeXz1dVp1fVrqratXv37ukrBdZFNmGcZBPGSTZhPmY5dfaYJKcm+Y4kxyY5oqpesHy+1tpZrbWdrbWd27dvn75SYF1kE8ZJNmGcZBPmY5ZTZ5+W5O9aa7tba99I8t4kT+xTFjAD2YRxkk0YJ9mEOZil0bwqyeOr6j5VVUmemuSyPmUBM5BNGCfZhHGSTZiDWT6jeUGSdye5MMlnhmWd1akuYEqyCeMkmzBOsgnzsWWWwa21X07yy51qATqRTRgn2YRxkk3ob9avNwEAAIB70GgCAADQlUYTAACArjSaAAAAdKXRBAAAoCuNJgAAAF1pNAEAAOhKowkAAEBXGk0AAAC60mgCAADQlUYTAACArjSaAAAAdKXRBAAAoCuNJgAAAF1pNAEAAOhKowkAAEBXGk0AAAC60mgCAADQlUYTAACArtZsNKvqLVV1Q1V9dsljW6vqvKr6wvDzmPmWCSwnmzBOsgnjJJuwWOs5onl2klOWPXZmkvNbaw9Lcv7wO7BYZ0c2YYzOjmzCGJ0d2YSFWbPRbK19PMnNyx4+Nclbh/tvTfKjnesC1iCbME6yCeMkm7BY035G84GtteuSZPj5gNVmrKrTq2pXVe3avXv3lKsD1kk2YZxkE8ZJNmFO5n4xoNbaWa21na21ndu3b5/36oB1kk0YJ9mEcZJN2JhpG80vV9WDk2T4eUO/koAZyCaMk2zCOMkmzMm0jeb7k7xwuP/CJO/rUw4wI9mEcZJNGCfZhDlZz9ebvCPJJ5M8oqquqaqXJPmNJD9UVV9I8kPD78ACySaMk2zCOMkmLNaWtWZorT1vlUlP7VwLsAGyCeMkmzBOsgmLNfeLAQEAAHBwWfOIZk933313vvKVr2x43LXXXjvV+u68884Nj3nQgx401bqmHcfepnndkuSjH/3ohsdccsklU63ruuuu2/CYq6++eqp1AQDA/sYRTQAAALrSaAIAANCVRhMAAICuNJoAAAB0pdEEAACgK40mAAAAXWk0AQAA6EqjCQAAQFcaTQAAALrSaAIAANCVRhMAAICuNJoAAAB0tWWRK7vqqqtyxhlnbHjcJz7xianWd+WVV254zGGHHTbVuk466aSpxp188slTjXvyk5+84THf9V3fNdW6jjnmmKnG3XbbbVONu+qqq6Yad/PNN294TFVNta5pxk27LuDg9c1vfjO33377hscdccQRc6imr2n+RifJpZdeOtW4W2+9dcNjTjzxxKnW9YQnPGGqcQAHEkc0AQAA6EqjCQAAQFcaTQAAALpas9GsqrdU1Q1V9dklj/12VX2uqi6pqj+pqqPnWyawnGzCOMkmjJNswmKt54jm2UlOWfbYeUm+s7X23Uk+n+RVnesC1nZ2ZBPG6OzIJozR2ZFNWJg1G83W2seT3LzssQ+31u4afv1UkuPnUBuwD7IJ4ySbME6yCYvV4zOaL07ywdUmVtXpVbWrqnZ9/etf77A6YJ3Wnc3du3cvsCw46K07mzfeeOMCy4KDnr+b0NFMjWZVvTrJXUnevto8rbWzWms7W2s7Dz/88FlWB6zTRrO5ffv2xRUHB7GNZnPbtm2LKw4OYv5uQn9bph1YVS9M8uwkT22ttX4lAbOQTRgn2YRxkk2Yj6kazao6JckvJvnB1todfUsCpiWbME6yCeMkmzA/6/l6k3ck+WSSR1TVNVX1kiT/LclRSc6rqour6o1zrhNYRjZhnGQTxkk2YbHWPKLZWnveCg+/eQ61ABsgmzBOsgnjJJuwWD2uOgsAAADfMvXFgKZx73vfO49+9KM3PG7LlunKvOGGGzY85qijjppqXQ996EOnGvc93/M9U417zGMes+Exxx133FTrmtZ973vfqcYde+yxU417/OMfP9W4aUzzlQPnn3/+HCoBDmQXXXRRjjzyyA2P27Fjx1TrO/74jX+F4NFHHz3Vuqb92z7tuGm24zTbI0m++MUvTjXu4Q9/+FTjpn29jznmmA2Pqaqp1nX33XdveIzr8sD+zRFNAAAAutJoAgAA0JVGEwAAgK40mgAAAHSl0QQAAKArjSYAAABdaTQBAADoSqMJAABAVxpNAAAAutJoAgAA0JVGEwAAgK40mgAAAHSl0QQAAKCraq0tbmVVu5NcucrkbUluXFgxqxtLHcl4alHH3qap5cTW2vZ5FDMr2dyQsdSRjKeWsdSRyOZmUMfexlLL/l6HbM5GHXsbSy37ex3ryuZCG819qapdrbWd6vi2sdSijr2NqZZ5G8tzVcfexlLLWOpIxlXLvI3luapjb2OpRR2bYyzPVx17G0stB0sdTp0FAACgK40mAAAAXY2p0TxrswsYjKWOZDy1qGNvY6pl3sbyXNWxt7HUMpY6knHVMm9jea7q2NtYalHH5hjL81XH3sZSy0FRx2g+owkAAMCBYUxHNAEAADgAaDQBAADoauGNZlWdUlV/W1WXV9WZK0w/rKreNUy/oKp2zKGGE6rqo1V1WVVdWlUvX2Gek6vqlqq6eLj9h951DOu5oqo+M6xj1wrTq6p+b9gel1TVY+dUxyOWPNeLq+rWqnrFsnnmsk2q6i1VdUNVfXbJY1ur6ryq+sLw85hVxr5wmOcLVfXCOdXy21X1uWH7/0lVHb3K2H2+lmM2hlwO65HNvddz0GfzYM1lIpur1CKbsrnpZHPFWjY9m5uZy2HZsrlUa21htySHJPlikockuVeSTyd55LJ5XpbkjcP905K8aw51PDjJY4f7RyX5/Ap1nJzkAwvYJlck2baP6c9K8sEkleTxSS5Y0Ot0fSZfxjr3bZLkB5I8Nslnlzz2W0nOHO6fmeQ3Vxi3NcmXhp/HDPePmUMtT0+yZbj/myvVsp7Xcqy3seRyWLZsrv1aHXTZPBhzueT1ls0NvqayKZsLer1lc4Ov6aKzuehcDsuWzSW3RR/R/N4kl7fWvtRauzPJO5OcumyeU5O8dbj/7iRPrarqWURr7brW2oXD/a8muSzJcT3X0dGpSf5Xm/hUkqOr6sFzXudTk3yxtXblnNeTJGmtfTzJzcseXvo+eGuSH11h6DOSnNdau7m19g9JzktySu9aWmsfbq3dNfz6qSTHz7KOERpFLhPZXIeDMpsHaS4T2ZyWbMrmvMnmdBadzYXmMpHN5RbdaB6X5Oolv1+TvcPwrXmGjXFLkvvPq6DhVIbHJLlghclPqKpPV9UHq+pRcyqhJflwVf11VZ2+wvT1bLPeTkvyjlWmLWKbJMkDW2vXJZN/RJM8YIV5NmPbvDiTvXErWeu1HKvR5TKRzVXI5soOxFwmsrka2VyZbC6ObK5sbNkcQy6TgzibW2YZPIWV9uQs/36V9czTRVUdmeQ9SV7RWrt12eQLMznUfltVPSvJnyZ52BzKeFJr7dqqekCS86rqc8NeiG+VucKYuX0nTVXdK8lzkrxqhcmL2ibrteht8+okdyV5+yqzrPVajtWocpnI5kpkc5UVHbi5TGRzNbI5PdnsQzZXNpps7me5TA7QbC76iOY1SU5Y8vvxSa5dbZ6q2pLkftn7EPTMqurQTAL59tbae5dPb63d2lq7bbh/bpJDq2pb7zpaa9cOP29I8ieZnI6x1Hq2WU/PTHJha+3LyycsapsMvrzndIrh5w0rzLOwbTN8KPvZSZ7fWlsx+Ot4LcdqNLkcli+bK5PNZQ7wXCayuSLZXJVsLo5srmBk2RxLLpODOJuLbjT/KsnDquo7hj0NpyV5/7J53p9kz5WWfjzJR1bbENMazpF/c5LLWmu/s8o8D9pzLn1VfW8m2+qmznUcUVVH7bmfyYd0P7tstvcn+cmaeHySW/Ycfp+T52WV0wwWsU2WWPo+eGGS960wz58neXpVHVOTK3g9fXisq6o6JckvJnlOa+2OVeZZz2s5VqPIZSKba5DNJQ6CXCayudI6ZHN1srk4srn3OsaWzbHkMjmYs9nmeLWnlW6ZXHHq85lcrevVw2O/OjzpJDk8yR8nuTzJXyZ5yBxqeHImh6MvSXLxcHtWkpcmeekwzxlJLs3kSmKfSvLEOdTxkGH5nx7WtWd7LK2jkrxh2F6fSbJzjq/NfTIJ2v2WPDb3bZLJPwTXJflGJnt0XpLJ5xjOT/KF4efWYd6dSd60ZOyLh/fK5UleNKdaLs/kvPk975U9V5E7Nsm5+3ot95fbGHI5rEc2V67noM7mwZrL4TnI5j3rkM0mm2O4yeZedYwmm5uVy2HZsrnkVsNCAQAAoItFnzoLAADAAU6jCQAAQFcaTQAAALrSaAIAANCVRhMAAICuNJoAAAB0pdEEAACgq/8PYyhVwf5EMncAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# passing average image for each number to one of the CNN filters and plotting the output\n",
    "visualize_filter(model, X_test, y_test, 3)"
   ]
  },
  {
   "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
}