Standard_NN_Part-1.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(3000, 785)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>f1</th>\n",
       "      <th>f2</th>\n",
       "      <th>f3</th>\n",
       "      <th>f4</th>\n",
       "      <th>f5</th>\n",
       "      <th>f6</th>\n",
       "      <th>f7</th>\n",
       "      <th>f8</th>\n",
       "      <th>f9</th>\n",
       "      <th>f10</th>\n",
       "      <th>...</th>\n",
       "      <th>f776</th>\n",
       "      <th>f777</th>\n",
       "      <th>f778</th>\n",
       "      <th>f779</th>\n",
       "      <th>f780</th>\n",
       "      <th>f781</th>\n",
       "      <th>f782</th>\n",
       "      <th>f783</th>\n",
       "      <th>f784</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>3000.0</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3000.000000</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>3000.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.029667</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.417667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.392875</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.907468</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>7.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>75.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 785 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           f1      f2      f3      f4      f5      f6      f7      f8      f9  \\\n",
       "count  3000.0  3000.0  3000.0  3000.0  3000.0  3000.0  3000.0  3000.0  3000.0   \n",
       "mean      0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "std       0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "min       0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "25%       0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "50%       0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "75%       0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "max       0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0     0.0   \n",
       "\n",
       "          f10  ...         f776    f777    f778    f779    f780    f781  \\\n",
       "count  3000.0  ...  3000.000000  3000.0  3000.0  3000.0  3000.0  3000.0   \n",
       "mean      0.0  ...     0.029667     0.0     0.0     0.0     0.0     0.0   \n",
       "std       0.0  ...     1.392875     0.0     0.0     0.0     0.0     0.0   \n",
       "min       0.0  ...     0.000000     0.0     0.0     0.0     0.0     0.0   \n",
       "25%       0.0  ...     0.000000     0.0     0.0     0.0     0.0     0.0   \n",
       "50%       0.0  ...     0.000000     0.0     0.0     0.0     0.0     0.0   \n",
       "75%       0.0  ...     0.000000     0.0     0.0     0.0     0.0     0.0   \n",
       "max       0.0  ...    75.000000     0.0     0.0     0.0     0.0     0.0   \n",
       "\n",
       "         f782    f783    f784        label  \n",
       "count  3000.0  3000.0  3000.0  3000.000000  \n",
       "mean      0.0     0.0     0.0     4.417667  \n",
       "std       0.0     0.0     0.0     2.907468  \n",
       "min       0.0     0.0     0.0     0.000000  \n",
       "25%       0.0     0.0     0.0     2.000000  \n",
       "50%       0.0     0.0     0.0     4.000000  \n",
       "75%       0.0     0.0     0.0     7.000000  \n",
       "max       0.0     0.0     0.0     9.000000  \n",
       "\n",
       "[8 rows x 785 columns]"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn import svm\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from matplotlib import pyplot as plt\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.metrics import f1_score\n",
    "from sklearn.metrics import classification_report, confusion_matrix\n",
    "import itertools\n",
    "from keras.utils import to_categorical\n",
    "#loading the dataset\n",
    "d = pd.read_csv(r'C:\\Users\\hp\\Downloads\\HAPT Data Set\\assignment4data.csv')\n",
    "#visualizing the datasets \n",
    "print(d.shape)\n",
    "d.head(5)\n",
    "d.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(3000, 784)\n",
      "(3000,)\n",
      "(2250, 10)\n"
     ]
    }
   ],
   "source": [
    "X=d.iloc[:,0:784]  # taking all the features\n",
    "y = d.iloc[:,784]   #labels\n",
    "print(X.shape)\n",
    "print(y.shape)\n",
    "scaler = StandardScaler()\n",
    "X = scaler.fit_transform(X)\n",
    "#splitting the dataset into train_set and test_set\n",
    "X_train, X_test, y_train, y_test = train_test_split(X,y,test_size=0.25)\n",
    "#reshaping labels\n",
    "y_train = to_categorical(y_train, 10)\n",
    "y_test = to_categorical(y_test, 10)\n",
    "print(y_train.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# A helpful function we'll be using all over the place to plot training information:\n",
    "def plot_training_history(history, model):\n",
    "    plt.plot(history.history['acc'])\n",
    "    plt.plot(history.history['val_acc'])\n",
    "    plt.title('model accuracy')\n",
    "    plt.ylabel('accuracy')\n",
    "    plt.xlabel('epoch')\n",
    "    plt.legend(['training', 'validation'], loc='best')\n",
    "\n",
    "    plt.show()\n",
    "\n",
    "    plt.plot(history.history['loss'])\n",
    "    plt.plot(history.history['val_loss'])\n",
    "    plt.title('model loss')\n",
    "    plt.ylabel('loss')\n",
    "    plt.xlabel('epoch')\n",
    "    plt.legend(['training', 'validation'], loc='best')\n",
    "\n",
    "    plt.show()\n",
    "    \n",
    "    loss, accuracy  = model.evaluate(X_test, y_test, verbose=False)\n",
    "    print(f'Test loss: {loss:.3}')\n",
    "    print(f'Test accuracy: {accuracy:.3}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Trying first with single hidden layer networks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_1 (Dense)              (None, 5)                 3925      \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 10)                60        \n",
      "=================================================================\n",
      "Total params: 3,985\n",
      "Trainable params: 3,985\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "from keras.utils import to_categorical\n",
    "model = Sequential()\n",
    "# starting with very simple model\n",
    "model.add(Dense(units=5, activation='sigmoid', input_shape=(784,)))  # single hidden layer with five neurons +input layer\n",
    "model.add(Dense(units=10, activation='softmax'))                    # output layer\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 1800 samples, validate on 450 samples\n",
      "Epoch 1/50\n",
      "1800/1800 [==============================] - 3s 2ms/step - loss: 2.3663 - acc: 0.1222 - val_loss: 2.3859 - val_acc: 0.1089\n",
      "Epoch 2/50\n",
      "1800/1800 [==============================] - 0s 42us/step - loss: 2.3279 - acc: 0.1361 - val_loss: 2.3492 - val_acc: 0.1289\n",
      "Epoch 3/50\n",
      "1800/1800 [==============================] - 0s 33us/step - loss: 2.2927 - acc: 0.1506 - val_loss: 2.3143 - val_acc: 0.1400\n",
      "Epoch 4/50\n",
      "1800/1800 [==============================] - 0s 33us/step - loss: 2.2601 - acc: 0.1661 - val_loss: 2.2820 - val_acc: 0.1511\n",
      "Epoch 5/50\n",
      "1800/1800 [==============================] - 0s 33us/step - loss: 2.2302 - acc: 0.1889 - val_loss: 2.2535 - val_acc: 0.1756\n",
      "Epoch 6/50\n",
      "1800/1800 [==============================] - 0s 38us/step - loss: 2.2035 - acc: 0.2122 - val_loss: 2.2274 - val_acc: 0.2044\n",
      "Epoch 7/50\n",
      "1800/1800 [==============================] - 0s 33us/step - loss: 2.1790 - acc: 0.2389 - val_loss: 2.2030 - val_acc: 0.2422\n",
      "Epoch 8/50\n",
      "1800/1800 [==============================] - 0s 34us/step - loss: 2.1568 - acc: 0.2639 - val_loss: 2.1808 - val_acc: 0.2600\n",
      "Epoch 9/50\n",
      "1800/1800 [==============================] - 0s 33us/step - loss: 2.1363 - acc: 0.2967 - val_loss: 2.1609 - val_acc: 0.2778\n",
      "Epoch 10/50\n",
      "1800/1800 [==============================] - 0s 32us/step - loss: 2.1177 - acc: 0.3200 - val_loss: 2.1421 - val_acc: 0.2956\n",
      "Epoch 11/50\n",
      "1800/1800 [==============================] - 0s 38us/step - loss: 2.1001 - acc: 0.3511 - val_loss: 2.1251 - val_acc: 0.3267\n",
      "Epoch 12/50\n",
      "1800/1800 [==============================] - 0s 33us/step - loss: 2.0842 - acc: 0.3778 - val_loss: 2.1090 - val_acc: 0.3511\n",
      "Epoch 13/50\n",
      "1800/1800 [==============================] - 0s 33us/step - loss: 2.0690 - acc: 0.4006 - val_loss: 2.0939 - val_acc: 0.3711\n",
      "Epoch 14/50\n",
      "1800/1800 [==============================] - 0s 34us/step - loss: 2.0551 - acc: 0.4167 - val_loss: 2.0803 - val_acc: 0.3889\n",
      "Epoch 15/50\n",
      "1800/1800 [==============================] - 0s 37us/step - loss: 2.0420 - acc: 0.4261 - val_loss: 2.0673 - val_acc: 0.4000\n",
      "Epoch 16/50\n",
      "1800/1800 [==============================] - 0s 36us/step - loss: 2.0296 - acc: 0.4344 - val_loss: 2.0548 - val_acc: 0.4067\n",
      "Epoch 17/50\n",
      "1800/1800 [==============================] - 0s 34us/step - loss: 2.0178 - acc: 0.4467 - val_loss: 2.0435 - val_acc: 0.4289\n",
      "Epoch 18/50\n",
      "1800/1800 [==============================] - 0s 33us/step - loss: 2.0069 - acc: 0.4600 - val_loss: 2.0326 - val_acc: 0.4400\n",
      "Epoch 19/50\n",
      "1800/1800 [==============================] - 0s 33us/step - loss: 1.9964 - acc: 0.4733 - val_loss: 2.0221 - val_acc: 0.4556\n",
      "Epoch 20/50\n",
      "1800/1800 [==============================] - 0s 36us/step - loss: 1.9862 - acc: 0.4789 - val_loss: 2.0117 - val_acc: 0.4578\n",
      "Epoch 21/50\n",
      "1800/1800 [==============================] - 0s 34us/step - loss: 1.9763 - acc: 0.4828 - val_loss: 2.0018 - val_acc: 0.4578\n",
      "Epoch 22/50\n",
      "1800/1800 [==============================] - 0s 35us/step - loss: 1.9667 - acc: 0.4894 - val_loss: 1.9927 - val_acc: 0.4533\n",
      "Epoch 23/50\n",
      "1800/1800 [==============================] - 0s 33us/step - loss: 1.9578 - acc: 0.4956 - val_loss: 1.9833 - val_acc: 0.4600\n",
      "Epoch 24/50\n",
      "1800/1800 [==============================] - 0s 33us/step - loss: 1.9488 - acc: 0.5067 - val_loss: 1.9746 - val_acc: 0.4733\n",
      "Epoch 25/50\n",
      "1800/1800 [==============================] - 0s 35us/step - loss: 1.9402 - acc: 0.5072 - val_loss: 1.9662 - val_acc: 0.4778\n",
      "Epoch 26/50\n",
      "1800/1800 [==============================] - 0s 34us/step - loss: 1.9318 - acc: 0.5144 - val_loss: 1.9580 - val_acc: 0.4756\n",
      "Epoch 27/50\n",
      "1800/1800 [==============================] - 0s 35us/step - loss: 1.9237 - acc: 0.5178 - val_loss: 1.9501 - val_acc: 0.4822\n",
      "Epoch 28/50\n",
      "1800/1800 [==============================] - 0s 34us/step - loss: 1.9158 - acc: 0.5228 - val_loss: 1.9423 - val_acc: 0.4889\n",
      "Epoch 29/50\n",
      "1800/1800 [==============================] - 0s 35us/step - loss: 1.9081 - acc: 0.5261 - val_loss: 1.9345 - val_acc: 0.4911\n",
      "Epoch 30/50\n",
      "1800/1800 [==============================] - 0s 35us/step - loss: 1.9005 - acc: 0.5278 - val_loss: 1.9272 - val_acc: 0.4956\n",
      "Epoch 31/50\n",
      "1800/1800 [==============================] - 0s 33us/step - loss: 1.8932 - acc: 0.5317 - val_loss: 1.9197 - val_acc: 0.5022\n",
      "Epoch 32/50\n",
      "1800/1800 [==============================] - 0s 34us/step - loss: 1.8859 - acc: 0.5350 - val_loss: 1.9123 - val_acc: 0.5089\n",
      "Epoch 33/50\n",
      "1800/1800 [==============================] - 0s 36us/step - loss: 1.8788 - acc: 0.5394 - val_loss: 1.9052 - val_acc: 0.5067\n",
      "Epoch 34/50\n",
      "1800/1800 [==============================] - 0s 37us/step - loss: 1.8718 - acc: 0.5400 - val_loss: 1.8987 - val_acc: 0.5067\n",
      "Epoch 35/50\n",
      "1800/1800 [==============================] - 0s 34us/step - loss: 1.8652 - acc: 0.5428 - val_loss: 1.8923 - val_acc: 0.5089\n",
      "Epoch 36/50\n",
      "1800/1800 [==============================] - 0s 35us/step - loss: 1.8587 - acc: 0.5433 - val_loss: 1.8860 - val_acc: 0.5067\n",
      "Epoch 37/50\n",
      "1800/1800 [==============================] - 0s 34us/step - loss: 1.8522 - acc: 0.5433 - val_loss: 1.8798 - val_acc: 0.5089\n",
      "Epoch 38/50\n",
      "1800/1800 [==============================] - 0s 34us/step - loss: 1.8459 - acc: 0.5444 - val_loss: 1.8738 - val_acc: 0.5044\n",
      "Epoch 39/50\n",
      "1800/1800 [==============================] - 0s 34us/step - loss: 1.8396 - acc: 0.5456 - val_loss: 1.8678 - val_acc: 0.5044\n",
      "Epoch 40/50\n",
      "1800/1800 [==============================] - 0s 34us/step - loss: 1.8335 - acc: 0.5456 - val_loss: 1.8619 - val_acc: 0.5067\n",
      "Epoch 41/50\n",
      "1800/1800 [==============================] - 0s 32us/step - loss: 1.8275 - acc: 0.5472 - val_loss: 1.8560 - val_acc: 0.5089\n",
      "Epoch 42/50\n",
      "1800/1800 [==============================] - 0s 35us/step - loss: 1.8214 - acc: 0.5517 - val_loss: 1.8503 - val_acc: 0.5133\n",
      "Epoch 43/50\n",
      "1800/1800 [==============================] - 0s 34us/step - loss: 1.8155 - acc: 0.5578 - val_loss: 1.8448 - val_acc: 0.5156\n",
      "Epoch 44/50\n",
      "1800/1800 [==============================] - 0s 36us/step - loss: 1.8097 - acc: 0.5589 - val_loss: 1.8394 - val_acc: 0.5156\n",
      "Epoch 45/50\n",
      "1800/1800 [==============================] - 0s 32us/step - loss: 1.8039 - acc: 0.5606 - val_loss: 1.8341 - val_acc: 0.5133\n",
      "Epoch 46/50\n",
      "1800/1800 [==============================] - 0s 34us/step - loss: 1.7983 - acc: 0.5639 - val_loss: 1.8288 - val_acc: 0.5178\n",
      "Epoch 47/50\n",
      "1800/1800 [==============================] - 0s 33us/step - loss: 1.7927 - acc: 0.5656 - val_loss: 1.8234 - val_acc: 0.5178\n",
      "Epoch 48/50\n",
      "1800/1800 [==============================] - 0s 34us/step - loss: 1.7870 - acc: 0.5667 - val_loss: 1.8184 - val_acc: 0.5200\n",
      "Epoch 49/50\n",
      "1800/1800 [==============================] - 0s 34us/step - loss: 1.7815 - acc: 0.5667 - val_loss: 1.8134 - val_acc: 0.5222\n",
      "Epoch 50/50\n",
      "1800/1800 [==============================] - 0s 35us/step - loss: 1.7760 - acc: 0.5689 - val_loss: 1.8083 - val_acc: 0.5222\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test loss: 1.78\n",
      "Test accuracy: 0.568\n"
     ]
    }
   ],
   "source": [
    "model.compile(optimizer=\"sgd\", loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "history = model.fit(X_train, y_train, batch_size=128, epochs=50, verbose=True, validation_split=.2)\n",
    "plot_training_history(history, model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 2025 samples, validate on 225 samples\n",
      "Epoch 1/20\n",
      "2025/2025 [==============================] - 1s 437us/step - loss: 2.3665 - acc: 0.1472 - val_loss: 2.2373 - val_acc: 0.2444\n",
      "Epoch 2/20\n",
      "2025/2025 [==============================] - 0s 70us/step - loss: 2.1551 - acc: 0.2973 - val_loss: 2.0827 - val_acc: 0.3778\n",
      "Epoch 3/20\n",
      "2025/2025 [==============================] - 0s 68us/step - loss: 2.0169 - acc: 0.4380 - val_loss: 1.9693 - val_acc: 0.5156\n",
      "Epoch 4/20\n",
      "2025/2025 [==============================] - 0s 111us/step - loss: 1.9067 - acc: 0.5175 - val_loss: 1.8703 - val_acc: 0.5778\n",
      "Epoch 5/20\n",
      "2025/2025 [==============================] - 0s 71us/step - loss: 1.8090 - acc: 0.5788 - val_loss: 1.7832 - val_acc: 0.6133\n",
      "Epoch 6/20\n",
      "2025/2025 [==============================] - 0s 64us/step - loss: 1.7199 - acc: 0.6232 - val_loss: 1.7014 - val_acc: 0.6533\n",
      "Epoch 7/20\n",
      "2025/2025 [==============================] - 0s 60us/step - loss: 1.6385 - acc: 0.6543 - val_loss: 1.6275 - val_acc: 0.6578\n",
      "Epoch 8/20\n",
      "2025/2025 [==============================] - 0s 58us/step - loss: 1.5643 - acc: 0.6736 - val_loss: 1.5599 - val_acc: 0.6800\n",
      "Epoch 9/20\n",
      "2025/2025 [==============================] - 0s 92us/step - loss: 1.4957 - acc: 0.6978 - val_loss: 1.4986 - val_acc: 0.7111\n",
      "Epoch 10/20\n",
      "2025/2025 [==============================] - 0s 88us/step - loss: 1.4327 - acc: 0.7195 - val_loss: 1.4410 - val_acc: 0.7156\n",
      "Epoch 11/20\n",
      "2025/2025 [==============================] - 0s 65us/step - loss: 1.3752 - acc: 0.7378 - val_loss: 1.3860 - val_acc: 0.7422\n",
      "Epoch 12/20\n",
      "2025/2025 [==============================] - 0s 62us/step - loss: 1.3226 - acc: 0.7531 - val_loss: 1.3386 - val_acc: 0.7467\n",
      "Epoch 13/20\n",
      "2025/2025 [==============================] - 0s 61us/step - loss: 1.2734 - acc: 0.7635 - val_loss: 1.2935 - val_acc: 0.7644\n",
      "Epoch 14/20\n",
      "2025/2025 [==============================] - 0s 61us/step - loss: 1.2286 - acc: 0.7733 - val_loss: 1.2523 - val_acc: 0.7689\n",
      "Epoch 15/20\n",
      "2025/2025 [==============================] - 0s 74us/step - loss: 1.1864 - acc: 0.7807 - val_loss: 1.2147 - val_acc: 0.7689\n",
      "Epoch 16/20\n",
      "2025/2025 [==============================] - 0s 58us/step - loss: 1.1480 - acc: 0.7906 - val_loss: 1.1785 - val_acc: 0.7689\n",
      "Epoch 17/20\n",
      "2025/2025 [==============================] - 0s 83us/step - loss: 1.1115 - acc: 0.7990 - val_loss: 1.1452 - val_acc: 0.7822\n",
      "Epoch 18/20\n",
      "2025/2025 [==============================] - 0s 65us/step - loss: 1.0782 - acc: 0.8015 - val_loss: 1.1140 - val_acc: 0.7867\n",
      "Epoch 19/20\n",
      "2025/2025 [==============================] - 0s 60us/step - loss: 1.0469 - acc: 0.8079 - val_loss: 1.0859 - val_acc: 0.7911\n",
      "Epoch 20/20\n",
      "2025/2025 [==============================] - 0s 60us/step - loss: 1.0178 - acc: 0.8114 - val_loss: 1.0588 - val_acc: 0.8000\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test loss: 1.0\n",
      "Test accuracy: 0.816\n"
     ]
    }
   ],
   "source": [
    "model_2 = Sequential()\n",
    "\n",
    "model_2.add(Dense(units=256, activation='sigmoid', input_shape=(784,)))  # single hidden layer with 256 neurons\n",
    "model_2.add(Dense(units=10, activation='softmax'))\n",
    "\n",
    "model_2.compile(optimizer=\"sgd\", loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "history_2 = model_2.fit(X_train, y_train, batch_size=128, epochs=20, verbose=True, validation_split=.1)\n",
    "plot_training_history(history_2, model_2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "best accuracy we got is around 80% near 5-7 epoch after which model starts overfitting. Conlusion increasing the number of neurons in hidden layer increses accuracy... but starts overfitting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 2025 samples, validate on 225 samples\n",
      "Epoch 1/20\n",
      "2025/2025 [==============================] - 7s 3ms/step - loss: 9.5448 - acc: 0.0864 - val_loss: 11.5168 - val_acc: 0.1022\n",
      "Epoch 2/20\n",
      "2025/2025 [==============================] - 4s 2ms/step - loss: 12.3207 - acc: 0.0904 - val_loss: 12.2789 - val_acc: 0.1022\n",
      "Epoch 3/20\n",
      "2025/2025 [==============================] - 3s 2ms/step - loss: 11.7439 - acc: 0.1249 - val_loss: 9.9697 - val_acc: 0.2356\n",
      "Epoch 4/20\n",
      "2025/2025 [==============================] - 4s 2ms/step - loss: 11.4090 - acc: 0.1348 - val_loss: 11.9979 - val_acc: 0.1244\n",
      "Epoch 5/20\n",
      "2025/2025 [==============================] - 4s 2ms/step - loss: 11.9492 - acc: 0.1111 - val_loss: 11.2108 - val_acc: 0.1200\n",
      "Epoch 6/20\n",
      "2025/2025 [==============================] - 3s 2ms/step - loss: 11.7092 - acc: 0.1215 - val_loss: 10.8195 - val_acc: 0.1200\n",
      "Epoch 7/20\n",
      "2025/2025 [==============================] - 4s 2ms/step - loss: 11.4661 - acc: 0.1427 - val_loss: 10.6893 - val_acc: 0.1511\n",
      "Epoch 8/20\n",
      "2025/2025 [==============================] - 4s 2ms/step - loss: 11.0957 - acc: 0.1709 - val_loss: 10.4043 - val_acc: 0.1422\n",
      "Epoch 9/20\n",
      "2025/2025 [==============================] - 3s 2ms/step - loss: 11.0527 - acc: 0.1664 - val_loss: 10.4899 - val_acc: 0.1422\n",
      "Epoch 10/20\n",
      "2025/2025 [==============================] - 3s 2ms/step - loss: 10.9976 - acc: 0.1714 - val_loss: 10.1015 - val_acc: 0.1956\n",
      "Epoch 11/20\n",
      "2025/2025 [==============================] - 3s 2ms/step - loss: 10.7695 - acc: 0.1936 - val_loss: 10.3187 - val_acc: 0.1600\n",
      "Epoch 12/20\n",
      "2025/2025 [==============================] - 4s 2ms/step - loss: 10.8538 - acc: 0.1906 - val_loss: 9.7555 - val_acc: 0.3378\n",
      "Epoch 13/20\n",
      "2025/2025 [==============================] - 4s 2ms/step - loss: 10.7282 - acc: 0.2114 - val_loss: 10.1780 - val_acc: 0.1956\n",
      "Epoch 14/20\n",
      "2025/2025 [==============================] - 4s 2ms/step - loss: 10.5877 - acc: 0.2296 - val_loss: 10.3924 - val_acc: 0.1333\n",
      "Epoch 15/20\n",
      "2025/2025 [==============================] - 4s 2ms/step - loss: 10.5971 - acc: 0.2464 - val_loss: 9.7203 - val_acc: 0.3644\n",
      "Epoch 16/20\n",
      "2025/2025 [==============================] - 4s 2ms/step - loss: 10.6307 - acc: 0.2415 - val_loss: 10.1405 - val_acc: 0.2489\n",
      "Epoch 17/20\n",
      "2025/2025 [==============================] - 4s 2ms/step - loss: 10.4988 - acc: 0.2573 - val_loss: 9.9923 - val_acc: 0.2444\n",
      "Epoch 18/20\n",
      "2025/2025 [==============================] - 4s 2ms/step - loss: 10.4377 - acc: 0.2765 - val_loss: 9.7672 - val_acc: 0.3156\n",
      "Epoch 19/20\n",
      "2025/2025 [==============================] - 4s 2ms/step - loss: 10.3386 - acc: 0.2998 - val_loss: 9.7188 - val_acc: 0.3600\n",
      "Epoch 20/20\n",
      "2025/2025 [==============================] - 3s 2ms/step - loss: 10.3928 - acc: 0.2894 - val_loss: 10.1503 - val_acc: 0.2000\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test loss: 10.6\n",
      "Test accuracy: 0.203\n"
     ]
    }
   ],
   "source": [
    "model_3 = Sequential()\n",
    "\n",
    "model_3.add(Dense(units=20048, activation='sigmoid', input_shape=(784,)))  # with 2048 neurons\n",
    "model_3.add(Dense(units=10, activation='softmax'))\n",
    "\n",
    "model_3.compile(optimizer=\"sgd\", loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "history_3 = model_3.fit(X_train, y_train, batch_size=128, epochs=20, verbose=True, validation_split=.1)\n",
    "plot_training_history(history_3, model_3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Increasing more neurons model did better..but it become sluggish as complexity of network increases.\n",
    "\n",
    "This network took longer to train. It appears to be overfit as early as the second training epoch, and indeed when we run the test data it's worse than the training performance. Not only that, but it actually performs significantly WORSE than the simpler networks above. Five nodes performed about as well ten thousand nodes, but 256 nodes performed better than either of these"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Multilayer perceptrons"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_43 (Dense)             (None, 256)               200960    \n",
      "_________________________________________________________________\n",
      "dense_44 (Dense)             (None, 256)               65792     \n",
      "_________________________________________________________________\n",
      "dense_45 (Dense)             (None, 256)               65792     \n",
      "_________________________________________________________________\n",
      "dense_46 (Dense)             (None, 10)                2570      \n",
      "=================================================================\n",
      "Total params: 335,114\n",
      "Trainable params: 335,114\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# 9000 total nodes in hidden layers\n",
    "long_model = Sequential()\n",
    "long_model.add(Dense(units=256, activation='sigmoid', input_shape=(784,)))\n",
    "long_model.add(Dense(units=256, activation='sigmoid'))\n",
    "long_model.add(Dense(units=256, activation='sigmoid'))\n",
    "long_model.add(Dense(units=10, activation='softmax'))\n",
    "long_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 2025 samples, validate on 225 samples\n",
      "Epoch 1/10\n",
      "2025/2025 [==============================] - 2s 1ms/step - loss: 2.3648 - acc: 0.1111 - val_loss: 2.3326 - val_acc: 0.0933\n",
      "Epoch 2/10\n",
      "2025/2025 [==============================] - 0s 130us/step - loss: 2.3052 - acc: 0.1072 - val_loss: 2.3098 - val_acc: 0.0933\n",
      "Epoch 3/10\n",
      "2025/2025 [==============================] - 0s 121us/step - loss: 2.2976 - acc: 0.1042 - val_loss: 2.3068 - val_acc: 0.0756\n",
      "Epoch 4/10\n",
      "2025/2025 [==============================] - 0s 129us/step - loss: 2.2965 - acc: 0.1047 - val_loss: 2.3050 - val_acc: 0.0800\n",
      "Epoch 5/10\n",
      "2025/2025 [==============================] - 0s 129us/step - loss: 2.2957 - acc: 0.1086 - val_loss: 2.3037 - val_acc: 0.0933\n",
      "Epoch 6/10\n",
      "2025/2025 [==============================] - 0s 118us/step - loss: 2.2949 - acc: 0.1141 - val_loss: 2.3039 - val_acc: 0.0889\n",
      "Epoch 7/10\n",
      "2025/2025 [==============================] - 0s 120us/step - loss: 2.2935 - acc: 0.1101 - val_loss: 2.3038 - val_acc: 0.0844\n",
      "Epoch 8/10\n",
      "2025/2025 [==============================] - 0s 114us/step - loss: 2.2928 - acc: 0.1146 - val_loss: 2.3025 - val_acc: 0.0800\n",
      "Epoch 9/10\n",
      "2025/2025 [==============================] - 0s 101us/step - loss: 2.2919 - acc: 0.1180 - val_loss: 2.3034 - val_acc: 0.1156\n",
      "Epoch 10/10\n",
      "2025/2025 [==============================] - 0s 110us/step - loss: 2.2911 - acc: 0.1195 - val_loss: 2.2991 - val_acc: 0.1600\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test loss: 2.29\n",
      "Test accuracy: 0.167\n"
     ]
    }
   ],
   "source": [
    "long_model.compile(optimizer=\"sgd\", loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "history_long = long_model.fit(X_train, y_train, batch_size=128, epochs=10, verbose=True, validation_split=.1)\n",
    "plot_training_history(history_long, long_model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_11 (Dense)             (None, 2048)              1607680   \n",
      "_________________________________________________________________\n",
      "dense_12 (Dense)             (None, 1024)              2098176   \n",
      "_________________________________________________________________\n",
      "dense_13 (Dense)             (None, 512)               524800    \n",
      "_________________________________________________________________\n",
      "dense_14 (Dense)             (None, 256)               131328    \n",
      "_________________________________________________________________\n",
      "dense_15 (Dense)             (None, 128)               32896     \n",
      "_________________________________________________________________\n",
      "dense_16 (Dense)             (None, 10)                1290      \n",
      "=================================================================\n",
      "Total params: 4,396,170\n",
      "Trainable params: 4,396,170\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Train on 2025 samples, validate on 225 samples\n",
      "Epoch 1/50\n",
      "2025/2025 [==============================] - 4s 2ms/step - loss: 0.9743 - acc: 0.6795 - val_loss: 0.6179 - val_acc: 0.8044\n",
      "Epoch 2/50\n",
      "2025/2025 [==============================] - 2s 914us/step - loss: 0.3105 - acc: 0.9007 - val_loss: 0.2947 - val_acc: 0.9022\n",
      "Epoch 3/50\n",
      "2025/2025 [==============================] - 2s 902us/step - loss: 0.1066 - acc: 0.9625 - val_loss: 0.2517 - val_acc: 0.9289\n",
      "Epoch 4/50\n",
      "2025/2025 [==============================] - 2s 883us/step - loss: 0.0356 - acc: 0.9881 - val_loss: 0.2982 - val_acc: 0.9244\n",
      "Epoch 5/50\n",
      "2025/2025 [==============================] - 2s 1ms/step - loss: 0.0299 - acc: 0.9906 - val_loss: 0.4205 - val_acc: 0.9022\n",
      "Epoch 6/50\n",
      "2025/2025 [==============================] - 2s 982us/step - loss: 0.0143 - acc: 0.9946 - val_loss: 0.3403 - val_acc: 0.9200\n",
      "Epoch 7/50\n",
      "2025/2025 [==============================] - 2s 909us/step - loss: 0.0181 - acc: 0.9946 - val_loss: 0.3793 - val_acc: 0.9289\n",
      "Epoch 8/50\n",
      "2025/2025 [==============================] - 2s 962us/step - loss: 0.0314 - acc: 0.9886 - val_loss: 0.5161 - val_acc: 0.9067\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test loss: 0.277\n",
      "Test accuracy: 0.92\n"
     ]
    }
   ],
   "source": [
    "from keras.layers import Dropout\n",
    "from keras.callbacks import EarlyStopping\n",
    "long_model_two = Sequential()\n",
    "long_model_two.add(Dense(units=2048, activation='relu', input_shape=(784,)))\n",
    "long_model_two.add(Dense(units=1024, activation='relu'))\n",
    "#long_model.add(Dropout(rate=0.1))\n",
    "long_model_two.add(Dense(units=512, activation='relu'))\n",
    "long_model.add(Dropout(rate=0.2))\n",
    "long_model_two.add(Dense(units=256, activation='relu'))\n",
    "long_model.add(Dropout(rate=0.2))\n",
    "long_model_two.add(Dense(units=128, activation='relu'))\n",
    "long_model.add(Dropout(rate=0.25))\n",
    "long_model_two.add(Dense(units=10, activation='softmax'))\n",
    "long_model_two.summary()\n",
    "\n",
    "long_model_two.compile(optimizer=\"adam\", loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "# There are lots of parameters you can set here, including the amount that counts as a change\n",
    "# what metric to watch, and more. See the documentation for more details. We'll keep it\n",
    "# simple for now, and use the defaults for most of these. \n",
    "early_stopper = EarlyStopping(monitor='val_loss', patience=5, restore_best_weights=True)\n",
    "\n",
    "history = long_model_two.fit(X_train, y_train, batch_size=128, epochs=50, verbose=True, validation_split=.1,callbacks=[early_stopper])\n",
    "plot_training_history(history, long_model_two)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "dropout_rate: 0.8\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test loss: 1.78\n",
      "Test accuracy: 0.568\n",
      "\n",
      "dropout_rate: 0.5\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test loss: 1.78\n",
      "Test accuracy: 0.568\n",
      "\n",
      "dropout_rate: 0.3\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEWCAYAAABxMXBSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8li6FKAAAfbklEQVR4nO3de5gV1Z3u8e8rFxFFaS5GBBWckKhc5NIgM95A1GBUNAYVbxEnkQzGMXpyEXNmgjLjecwTY3iSqBGNxiR4IRiVOAoBhKhHY2iQIKAe0KA0eAG5iApqw+/8UdVk02ygiu5NQ/f7eZ799N5VtVat1TT9dtWqWqWIwMzMLKt96rsBZma2d3FwmJlZLg4OMzPLxcFhZma5ODjMzCwXB4eZmeXi4DDbAUm/lvTfGbddKunUUrfJrL45OMzMLBcHh1kjIKlpfbfBGg4Hh+310lNE35M0X9JHkn4l6XOSnpK0XtJ0SWUF2w+VtFDSWkmzJB1dsK63pLlpuYeBFjX2dZakeWnZ5yX1zNjGMyW9JOkDScsk3Vhj/QlpfWvT9SPS5ftJ+omkNyWtk/RcumygpMoi34dT0/c3Spok6XeSPgBGSOov6YV0H29L+oWk5gXlu0maJmm1pHcl/UDSIZI+ltS2YLu+klZKapal79bwODisofgqcBrwBeBs4CngB0A7kp/zawAkfQF4ELgWaA88CfxRUvP0l+hjwG+BNsDv03pJy/YB7gW+CbQF7gImS9o3Q/s+Ar4GtAbOBEZJOjet9/C0vT9P29QLmJeWuxXoC/xL2qbvA5szfk/OASal+5wAbAKuS78n/wwMBq5K29AKmA5MAQ4FPg/MiIh3gFnABQX1Xgo8FBGfZWyHNTAODmsofh4R70bEcuBZ4MWIeCkiPgEeBXqn210I/E9ETEt/8d0K7Efyi3kA0AwYFxGfRcQkYHbBPq4E7oqIFyNiU0TcD3ySltuhiJgVES9HxOaImE8SXienqy8BpkfEg+l+34+IeZL2Af4V+HZELE/3+XzapyxeiIjH0n1uiIg5EfGXiKiKiKUkwVfdhrOAdyLiJxGxMSLWR8SL6br7ScICSU2Ai0jC1RopB4c1FO8WvN9Q5PMB6ftDgTerV0TEZmAZ0DFdtzy2nvnzzYL3RwDfSU/1rJW0FjgsLbdDko6TNDM9xbMO+DeSv/xJ63i9SLF2JKfKiq3LYlmNNnxB0hOS3klPX/2fDG0AeBw4RtKRJEd16yLir7vYJmsAHBzW2KwgCQAAJInkl+Zy4G2gY7qs2uEF75cBN0dE64JXy4h4MMN+HwAmA4dFxEHAL4Hq/SwD/qlImVXAxu2s+whoWdCPJiSnuQrVnPr6TuBVoGtEHEhyKm9nbSAiNgITSY6MLsNHG42eg8Mam4nAmZIGp4O73yE53fQ88AJQBVwjqamk84D+BWXvBv4tPXqQpP3TQe9WGfbbClgdERsl9QcuLlg3AThV0gXpfttK6pUeDd0L3CbpUElNJP1zOqby/4AW6f6bAf8B7GyspRXwAfChpKOAUQXrngAOkXStpH0ltZJ0XMH63wAjgKHA7zL01xowB4c1KhHxGsn5+p+T/EV/NnB2RHwaEZ8C55H8glxDMh7yh4KyFSTjHL9I1y9Jt83iKmCspPXAD0kCrLret4Avk4TYapKB8WPT1d8FXiYZa1kN/AjYJyLWpXXeQ3K09BGw1VVWRXyXJLDWk4TgwwVtWE9yGups4B1gMTCoYP3/JRmUn5uOj1gjJj/IycyykPQ08EBE3FPfbbH65eAws52S1A+YRjJGs76+22P1q6SnqiQNkfSapCWSRhdZf1J6s1WVpGE11k1Jr1x5osbyX0v6e3oT1jxJvUrZB7PGTtL9JPd4XOvQMICSTUOQXuVxO8l500pgtqTJEbGoYLO3SM4Rf7dIFT8muWrkm0XWfS+9xt7MSiwiLq/vNtiepZRHHP2BJRHxRjro+BDJnaxbRMTS9Gaobe6EjYgZJIN4Zma2BynlxGcd2foGpErguO1sm9fNkn4IzABG7+xO2nbt2kXnzp3raNdmZo3DnDlzVkVEzfuDShocKrKsLkbibyC5XLA5MB64Hhi7zc6lkcBIgMMPP5yKioo62LWZWeMh6c1iy0t5qqqS5I7cap1I7tqtlYh4OxKfAPex9Q1ahduNj4jyiChv336bwDQzs11UyuCYDXSV1CWddXQ4yZQLtSKpQ/pVwLnAgtrWaWZm2ZUsOCKiCrgamAq8AkyMiIWSxkoaCsm14ekzBc4H7pK0sLq8pGdJprUeLKlS0pfSVRMkvUxyN207INNjPc3MrG40ihsAy8vLw2McZmb5SJoTEeU1l3uuKjMzy8XBYWZmuTg4zMwsl1Lex7HXu+mPC1m04oP6boaZ2S455tADGXN2tzqv10ccZmaWi484dqAUSW1mtrfzEYeZmeXi4DAzs1wcHGZmlouDw8zMcnFwmJlZLg4OMzPLxcFhZma5ODjMzCwXB4eZmeXi4DAzs1wcHGZmlouDw8zMcnFwmJlZLg4OMzPLxcFhZma5ODjMzCwXB4eZmeXi4DAzs1wcHGZmlouDw8zMcnFwmJlZLg4OMzPLxcFhZma5ODjMzCwXB4eZmeXi4DAzs1wcHGZmlouDw8zMcnFwmJlZLiUNDklDJL0maYmk0UXWnyRprqQqScNqrJsiaa2kJ2os7yLpRUmLJT0sqXkp+2BmZlsrWXBIagLcDpwBHANcJOmYGpu9BYwAHihSxY+By4os/xHw04joCqwBvl5XbTYzs50r5RFHf2BJRLwREZ8CDwHnFG4QEUsjYj6wuWbhiJgBrC9cJknAKcCkdNH9wLklaLuZmW1HKYOjI7Cs4HNluqw22gJrI6JqZ3VKGimpQlLFypUra7lbMzOrVsrgUJFlsbvqjIjxEVEeEeXt27ev5W7NzKxaKYOjEjis4HMnYEUt61wFtJbUtA7rNDOzHEoZHLOBrulVUM2B4cDk2lQYEQHMBKqvwLoceLxWrTQzs1xKFhzpOMTVwFTgFWBiRCyUNFbSUABJ/SRVAucDd0laWF1e0rPA74HBkiolfSlddT3wvyQtIRnz+FWp+mBmZttS8kd8w1ZeXh4VFRX13Qwzs72KpDkRUV5zue8cNzOzXBwcZmaWi4PDzMxycXCYmVkuDg4zM8vFwWFmZrk4OMzMLBcHh5mZ5eLgMDOzXBwcZmaWi4PDzMxycXCYmVkuDg4zM8vFwWFmZrk4OMzMLBcHh5mZ5eLgMDOzXBwcZmaWi4PDzMxycXCYmVkuDg4zM8vFwWFmZrk4OMzMLBcHh5mZ5eLgMDOzXBwcZmaWi4PDzMxycXCYmVkuDg4zM8vFwWFmZrk4OMzMLBcHh5mZ5eLgMDOzXBwcZmaWS0mDQ9IQSa9JWiJpdJH1J0maK6lK0rAa6y6XtDh9XV6wfFZa57z0dXAp+2BmZltrWqqKJTUBbgdOAyqB2ZImR8Sigs3eAkYA361Rtg0wBigHApiTll2TbnJJRFSUqu1mZrZ9pTzi6A8siYg3IuJT4CHgnMINImJpRMwHNtco+yVgWkSsTsNiGjCkhG01M7OMShkcHYFlBZ8r02V1Ufa+9DTVf0pSsQokjZRUIali5cqVedptZmY7UMrgKPYLPeqg7CUR0QM4MX1dVqyCiBgfEeURUd6+ffuMuzUzs53JFBySHpF0pqQ8QVMJHFbwuROworZlI2J5+nU98ADJKTEzM9tNsgbBncDFwGJJt0g6KkOZ2UBXSV0kNQeGA5Mz7m8qcLqkMkllwOnAVElNJbUDkNQMOAtYkLFOMzOrA5mCIyKmR8QlQB9gKTBN0vOSrkh/gRcrUwVcTRICrwATI2KhpLGShgJI6iepEjgfuEvSwrTsauC/SMJnNjA2XbYvSYDMB+YBy4G7d7HvZma2CxSRbdhBUlvgUpIxhRXABOAEoEdEDCxVA+tCeXl5VFT46l0zszwkzYmI8prLM93HIekPwFHAb4GzI+LtdNXDkvwb2cx2m88++4zKyko2btxY301pMFq0aEGnTp1o1qzoCaRtZL0B8BcR8XSxFcXSyMysVCorK2nVqhWdO3dmO1fjWw4Rwfvvv09lZSVdunTJVCbr4PjRklpXf0gHra/alUaamdXGxo0badu2rUOjjkiibdu2uY7gsgbHlRGxtvpDejf3lTnbZ2ZWJxwadSvv9zNrcOxTeId2Og9V81x7MjNrANauXcsdd9yRu9yXv/xl1q5du8NtfvjDHzJ9+vRdbdpukzU4pgITJQ2WdArwIDCldM0yM9szbS84Nm3atMNyTz75JK1bt97hNmPHjuXUU0+tVft2h6zBcT3wNDAK+BYwA/h+qRplZranGj16NK+//jq9evWiX79+DBo0iIsvvpgePXoAcO6559K3b1+6devG+PHjt5Tr3Lkzq1atYunSpRx99NFceeWVdOvWjdNPP50NGzYAMGLECCZNmrRl+zFjxtCnTx969OjBq6++CsDKlSs57bTT6NOnD9/85jc54ogjWLVq1W79HmS6qioiNpPcPX5naZtjZpbdTX9cyKIVH9RpnccceiBjzu623fW33HILCxYsYN68ecyaNYszzzyTBQsWbLki6d5776VNmzZs2LCBfv368dWvfpW2bdtuVcfixYt58MEHufvuu7ngggt45JFHuPTSS7fZV7t27Zg7dy533HEHt956K/fccw833XQTp5xyCjfccANTpkzZKpx2l6xzVXWVNEnSIklvVL9K3Tgzsz1d//79t7qM9Wc/+xnHHnssAwYMYNmyZSxevHibMl26dKFXr14A9O3bl6VLlxat+7zzzttmm+eee47hw4cDMGTIEMrKyuqwN9lkvY/jPpIHK/0UGARcQfEZbM3MdpsdHRnsLvvvv/+W97NmzWL69Om88MILtGzZkoEDBxa9zHXffffd8r5JkyZbTlVtb7smTZpQVVUFJPdd1LesYxz7RcQMkilK3oyIG4FTStcsM7M9U6tWrVi/fn3RdevWraOsrIyWLVvy6quv8pe//KXO93/CCScwceJEAP70pz+xZs2anZSoe1mPODamU6ovlnQ1yeSCfta3mTU6bdu25fjjj6d79+7st99+fO5zn9uybsiQIfzyl7+kZ8+efPGLX2TAgAF1vv8xY8Zw0UUX8fDDD3PyySfToUMHWrVqVef72ZFMkxxK6kcyw21rkllrDwR+HBF1H6cl4EkOzRqOV155haOPPrq+m1FvPvnkE5o0aULTpk154YUXGDVqFPPmzat1vcW+r7s8yWF6s98FEfE94EOS8Q0zM6sHb731FhdccAGbN2+mefPm3H337n+yxE6DIyI2SeorSbEnjMqYmTViXbt25aWXXqrXNmQd43gJeFzS74GPqhdGxB9K0iozM9tjZQ2ONsD7bH0lVQAODjOzRibrneMe1zAzMyD7EwDvIznC2EpE/Gudt8jMzPZoWW8AfAL4n/Q1g+Ry3A9L1Sgzs4bkgAMOAGDFihUMGzas6DYDBw5kZ7cNjBs3jo8//njL5yxTtZdC1lNVjxR+lvQgsOdPGm9mtgc59NBDt8x+uyvGjRvHpZdeSsuWLYFkqvb6kPWIo6auwOF12RAzs73F9ddfv9UzOW688UZuuukmBg8evGUa9Mcff3ybckuXLqV79+4AbNiwgeHDh9OzZ08uvPDCrearGjVqFOXl5XTr1o0xY8YAyeSJK1asYNCgQQwaNAj4x1TtALfddhvdu3ene/fujBs3bsv+tjeFe21kHeNYz9ZjHO+QPKPDzKz+PDUa3nm5bus8pAecccsONxk+fDjXXnstV111FQATJ05kypQpXHfddRx44IGsWrWKAQMGMHTo0O0+lvXOO++kZcuWzJ8/n/nz59OnT58t626++WbatGnDpk2bGDx4MPPnz+eaa67htttuY+bMmbRr126ruubMmcN9993Hiy++SERw3HHHcfLJJ1NWVpZ5Cvc8sp6q2r0ToZiZ7cF69+7Ne++9x4oVK1i5ciVlZWV06NCB6667jmeeeYZ99tmH5cuX8+6773LIIYcUreOZZ57hmmuuAaBnz5707Nlzy7qJEycyfvx4qqqqePvtt1m0aNFW62t67rnn+MpXvrJlpt7zzjuPZ599lqFDh2aewj2PrEccXwGejoh16efWwMCIeKzWLTAz21U7OTIopWHDhjFp0iTeeecdhg8fzoQJE1i5ciVz5syhWbNmdO7cueiU6oWKHY38/e9/59Zbb2X27NmUlZUxYsSIndazo0k9sk7hnkfWMY4x1aEBEBFrSZ7PYWbWKA0fPpyHHnqISZMmMWzYMNatW8fBBx9Ms2bNmDlzJm+++eYOy5900klMmDABgAULFjB//nwAPvjgA/bff38OOugg3n33XZ566qktZbY3pftJJ53EY489xscff8xHH33Eo48+yoknnliHvd1a1jvHiwVM1rJmZg1Ot27dWL9+PR07dqRDhw5ccsklnH322ZSXl9OrVy+OOuqoHZYfNWoUV1xxBT179qRXr170798fgGOPPZbevXvTrVs3jjzySI4//vgtZUaOHMkZZ5xBhw4dmDlz5pblffr0YcSIEVvq+MY3vkHv3r3r5LRUMVmnVb8XWAvcTjJI/u9AWUSMKEmr6pinVTdrOBr7tOqlkmda9aynqv4d+BR4GJgIbAC+Vct2mpnZXijrVVUfAaNL3BYzM9sLZDrikDQtvZKq+nOZpKmla5aZme2psp6qapdeSQVARKzBzxw3s3riZ8rVrbzfz6zBsVnSlilGJHWmyGy5Zmal1qJFC95//32HRx2JCN5//31atGiRuUzWS2r/N/CcpD+nn08CRuZsn5lZrXXq1InKykpWrlxZ301pMFq0aEGnTp0yb591cHyKpHKSsJgHPE5yZZWZ2W7VrFkzunTpUt/NaNSyDo5/g+Q5HN9JX78FbsxQboik1yQtkbTNVVmSTpI0V1KVpGE11l0uaXH6urxgeV9JL6d1/kzbm0HMzMxKIusYx7eBfsCbETEI6A3s8DhRUhOSGwbPAI4BLpJ0TI3N3gJGAA/UKNuGZEqT44D+wBhJZenqO0mOfLqmryEZ+2BmZnUga3BsjIiNAJL2jYhXgS/upEx/YElEvBERnwIPAecUbhARSyNiPrC5RtkvAdMiYnV6Bdc0YIikDsCBEfFCJCNjvwHOzdgHMzOrA1kHxyvT+zgeA6ZJWgOs2EmZjsCywjpIjiCyKFa2Y/qqLLJ8G5JGkg7gH364nzllZlZXsg6OfyV9e6OkmcBBwJSdFCs29pD1+rntlc1cZ0SMB8ZDMldVxv2amdlO5J7hNiL+vPOtgORo4LCCz53Y+VFKYdmBNcrOSpd3qrE8a51mZlYHdvWZ41nMBrpK6iKpOTAcmJyx7FTg9HRqkzLgdGBqRLwNrJc0IL2a6msklwabmdluUrLgiIgq4GqSEHgFmBgRCyWNlTQUQFI/SZXA+cBdkhamZVcD/0USPrOBsekygFHAPcAS4HXgKczMbLfJ9DyOvZ2fx2Fmll9tn8dhZmYGODjMzCwnB4eZmeXi4DAzs1wcHGZmlouDw8zMcnFwmJlZLg4OMzPLxcFhZma5ODjMzCwXB4eZmeXi4DAzs1wcHGZmlouDw8zMcnFwmJlZLg4OMzPLxcFhZma5ODjMzCwXB4eZmeXi4DAzs1wcHGZmlouDw8zMcnFwmJlZLg4OMzPLxcFhZma5ODjMzCwXB4eZmeXi4DAzs1wcHGZmlouDw8zMcnFwmJlZLg4OMzPLxcFhZma5ODjMzCyXkgaHpCGSXpO0RNLoIuv3lfRwuv5FSZ3T5c0l3SfpZUl/kzSwoMystM556evgUvbBzMy21rRUFUtqAtwOnAZUArMlTY6IRQWbfR1YExGflzQc+BFwIXAlQET0SIPhKUn9ImJzWu6SiKgoVdvNzGz7SnnE0R9YEhFvRMSnwEPAOTW2OQe4P30/CRgsScAxwAyAiHgPWAuUl7CtZmaWUSmDoyOwrOBzZbqs6DYRUQWsA9oCfwPOkdRUUhegL3BYQbn70tNU/5kGzTYkjZRUIali5cqVddMjMzMraXAU+4UeGbe5lyRoKoBxwPNAVbr+kojoAZyYvi4rtvOIGB8R5RFR3r59+11ovpmZFVPK4Khk66OETsCK7W0jqSlwELA6Iqoi4rqI6BUR5wCtgcUAEbE8/boeeIDklJiZme0mpQyO2UBXSV0kNQeGA5NrbDMZuDx9Pwx4OiJCUktJ+wNIOg2oiohF6amrdunyZsBZwIIS9sHMzGoo2VVVEVEl6WpgKtAEuDciFkoaC1RExGTgV8BvJS0BVpOEC8DBwFRJm4Hl/ON01L7p8mZpndOBu0vVBzMz25Yiag47NDzl5eVRUeGrd83M8pA0JyK2uaLVd46bmVkuDg4zM8vFwWFmZrk4OMzMLBcHh5mZ5eLgMDOzXBwcZmaWi4PDzMxycXCYmVkuDg4zM8vFwWFmZrk4OMzMLBcHh5mZ5eLgMDOzXBwcZmaWi4PDzMxycXCYmVkuDg4zM8vFwWFmZrk4OMzMLBcHh5mZ5eLgMDOzXBwcZmaWi4PDzMxycXCYmVkuDg4zM8vFwWFmZrk0re8G7NGeGg3vvFzfrTAz2zWH9IAzbqnzan3EYWZmufiIY0dKkNRmZns7H3GYmVkuDg4zM8vFwWFmZrk4OMzMLJeSBoekIZJek7RE0ugi6/eV9HC6/kVJndPlzSXdJ+llSX+TNLCgTN90+RJJP5OkUvbBzMy2VrLgkNQEuB04AzgGuEjSMTU2+zqwJiI+D/wU+FG6/EqAiOgBnAb8RFJ1W+8ERgJd09eQUvXBzMy2Vcojjv7Akoh4IyI+BR4CzqmxzTnA/en7ScDg9AjiGGAGQES8B6wFyiV1AA6MiBciIoDfAOeWsA9mZlZDKYOjI7Cs4HNluqzoNhFRBawD2gJ/A86R1FRSF6AvcFi6feVO6jQzsxIq5Q2AxcYeIuM29wJHAxXAm8DzQFXGOpOKpZEkp7QAPpT0WoY2F9MOWLWLZfdm7nfj0lj7DY2371n6fUSxhaUMjkqSo4RqnYAV29mmUlJT4CBgdXoa6rrqjSQ9DywG1qT17KhOACJiPDC+ln1AUkVElNe2nr2N+924NNZ+Q+Pte236XcpTVbOBrpK6SGoODAcm19hmMnB5+n4Y8HREhKSWkvYHkHQaUBURiyLibWC9pAHpWMjXgMdL2AczM6uhZEccEVEl6WpgKtAEuDciFkoaC1RExGTgV8BvJS0BVpOEC8DBwFRJm4HlwGUFVY8Cfg3sBzyVvszMbDcp6SSHEfEk8GSNZT8seL8ROL9IuaXAF7dTZwXQvU4bumO1Pt21l3K/G5fG2m9ovH3f5X4rGU4wMzPLxlOOmJlZLg4OMzPLxcGxAzuba6uhkHSvpPckLShY1kbSNEmL069l9dnGUpB0mKSZkl6RtFDSt9PlDbrvklpI+ms6D9xCSTely7ukc8YtTueQa17fbS0FSU0kvSTpifRzg++3pKXpHH/zJFWky3b559zBsR0Z59pqKH7NtnN+jQZmRERXkulfGmJwVgHfiYijgQHAt9J/44be90+AUyLiWKAXMETSAJK54n6a9nsNyVxyDdG3gVcKPjeWfg+KiF4F927s8s+5g2P7ssy11SBExDMkl0MXKpxH7H4a4JxgEfF2RMxN368n+WXSkQbe90h8mH5slr4COIVkzjhogP0GkNQJOBO4J/0sGkG/t2OXf84dHNuXZa6thuxz6Q2XpF8Pruf2lFQ6pX9v4EUaQd/T0zXzgPeAacDrwNp0zjhouD/v44DvA5vTz21pHP0O4E+S5qTTMUEtfs5Leh/HXi7zvFi2d5N0APAIcG1EfNAYHvESEZuAXpJaA4+SzA23zWa7t1WlJeks4L2ImFPwjJ/G8v/8+IhYIelgYJqkV2tTmY84ti/LXFsN2bvpNPakX9+r5/aUhKRmJKExISL+kC5uFH0HiIi1wCySMZ7W6Zxx0DB/3o8HhkpaSnLq+RSSI5CG3m8iYkX69T2SPxT6U4ufcwfH9mWZa6shK5xH7HIa4Jxg6fntXwGvRMRtBasadN8ltU+PNJC0H3AqyfjOTJI546AB9jsiboiIThHRmeT/89MRcQkNvN+S9pfUqvo9cDqwgFr8nPvO8R2Q9GWSv0iq59q6uZ6bVBKSHgQGkkyz/C4wBngMmAgcDrwFnB8RNQfQ92qSTgCeBV7mH+e8f0AyztFg+y6pJ8lgaBOSPx4nRsRYSUeS/CXeBngJuDQiPqm/lpZOeqrquxFxVkPvd9q/R9OPTYEHIuJmSW3ZxZ9zB4eZmeXiU1VmZpaLg8PMzHJxcJiZWS4ODjMzy8XBYWZmuTg4zPZwkgZWz+RqtidwcJiZWS4ODrM6IunS9DkX8yTdlU4k+KGkn0iaK2mGpPbptr0k/UXSfEmPVj8LQdLnJU1Pn5UxV9I/pdUfIGmSpFclTVBjmFDL9lgODrM6IOlo4EKSyeR6AZuAS4D9gbkR0Qf4M8ld+QC/Aa6PiJ4kd65XL58A3J4+K+NfgLfT5b2Ba0meDXMkybxLZvXCs+Oa1Y3BQF9gdnowsB/JpHGbgYfTbX4H/EHSQUDriPhzuvx+4PfpfEIdI+JRgIjYCJDW99eIqEw/zwM6A8+Vvltm23JwmNUNAfdHxA1bLZT+s8Z2O5rjZ0ennwrnTtqE/+9aPfKpKrO6MQMYlj7voPp5zkeQ/B+rnnn1YuC5iFgHrJF0Yrr8MuDPEfEBUCnp3LSOfSW13K29MMvAf7WY1YGIWCTpP0iesrYP8BnwLeAjoJukOcA6knEQSKax/mUaDG8AV6TLLwPukjQ2reP83dgNs0w8O65ZCUn6MCIOqO92mNUln6oyM7NcfMRhZma5+IjDzMxycXCYmVkuDg4zM8vFwWFmZrk4OMzMLJf/D41e4031xWCTAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test loss: 1.78\n",
      "Test accuracy: 0.568\n",
      "\n",
      "dropout_rate: 0.2\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEWCAYAAABrDZDcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8li6FKAAAaW0lEQVR4nO3dfZRV1Z3m8e8TXqR5UwScQVCLdOOKgCWUJSFNNBDUhaTFRAmW0XRwjNqkbSfpjBNNoqg9mXES29h2qxGjiUkbkWBr6DSIY4KJdquh8IUGlCUxKCVRCgJoRBT0N3/cY3kpbhW3oM69FPv5rMVa95yzz7m/XVXUU3ufe/dVRGBmZun6ULULMDOz6nIQmJklzkFgZpY4B4GZWeIcBGZmiXMQmJklzkFgViZJP5T0v8psu1bSyft6HbNKcBCYmSXOQWBmljgHgR1QsimZyyQtl/SmpDsk/RdJiyS9IelhSQOK2k+TtFLSFkmPSDqm6NhYSU9l590L9Gr1XH8h6Zns3P+QVLuXNV8oaY2kP0haIOnwbL8kfVfSBklbsz6Nzo5NlbQqq+0VSf9jr75gZjgI7MB0FnAKcDRwOrAI+DowiMLP/KUAko4G7gG+DAwGFgL/KqmnpJ7AA8CPgUOBn2bXJTu3DrgTuBgYCNwGLJB0UEcKlfRJ4P8AM4AhwEvA3OzwqcBJWT8OAc4GNmXH7gAujoh+wGjglx15XrNiDgI7EP1jRLwWEa8AjwJPRsTTEfE2cD8wNmt3NvBvEfH/ImIHcD3wJ8CfA+OBHsCNEbEjIuYDS4ue40Lgtoh4MiLejYi7gLez8zriXODOiHgqq+8K4GOSaoAdQD/gI4Ai4rmI+H123g5gpKT+EbE5Ip7q4POatXAQ2IHotaLHb5XY7ps9PpzCX+AARMR7wDpgaHbsldh1VcaXih4fBXw1mxbaImkLcER2Xke0ruGPFP7qHxoRvwT+CbgZeE3SHEn9s6ZnAVOBlyT9StLHOvi8Zi0cBJay9RR+oQOFOXkKv8xfAX4PDM32ve/IosfrgG9FxCFF/3pHxD37WEMfClNNrwBExE0RcTwwisIU0WXZ/qURcQZwGIUprHkdfF6zFg4CS9k84FOSJkvqAXyVwvTOfwCPAzuBSyV1l3QmMK7o3NuBv5L00eymbh9Jn5LUr4M1/AQ4X9KY7P7C/6YwlbVW0gnZ9XsAbwLbgXezexjnSjo4m9J6HXh3H74OljgHgSUrIlYD5wH/CGykcGP59Ih4JyLeAc4EZgKbKdxP+Jeicxsp3Cf4p+z4mqxtR2v4BXAlcB+FUcifAg3Z4f4UAmczhemjTRTuYwB8Hlgr6XXgr7J+mO0V+YNpzMzS5hGBmVniHARmZolzEJiZJc5BYGaWuO7VLqCjBg0aFDU1NdUuw8ysS1m2bNnGiBhc6liXC4KamhoaGxurXYaZWZci6aW2jnlqyMwscQ4CM7PEOQjMzBLX5e4RmNmBZceOHTQ1NbF9+/Zql3JA6NWrF8OGDaNHjx5ln+MgMLOqampqol+/ftTU1LDrYq/WURHBpk2baGpqYvjw4WWf56khM6uq7du3M3DgQIdAJ5DEwIEDOzy6chCYWdU5BDrP3nwtHQRmZolzEJhZ0rZs2cItt9zS4fOmTp3Kli1b2m1z1VVX8fDDD+9taRXjIDCzpLUVBO++2/6Hvi1cuJBDDjmk3TbXXnstJ5988j7VVwkOAjNL2uWXX85vf/tbxowZwwknnMCkSZP43Oc+x7HHHgvApz/9aY4//nhGjRrFnDlzWs6rqalh48aNrF27lmOOOYYLL7yQUaNGceqpp/LWW28BMHPmTObPn9/Sfvbs2dTV1XHsscfy/PPPA9Dc3Mwpp5xCXV0dF198MUcddRQbN26s6NfALx81s/3GNf+6klXrX+/Ua448vD+zTx/V5vHrrruOFStW8Mwzz/DII4/wqU99ihUrVrS8/PLOO+/k0EMP5a233uKEE07grLPOYuDAgbtc44UXXuCee+7h9ttvZ8aMGdx3332cd97unx46aNAgnnrqKW655Rauv/56vv/973PNNdfwyU9+kiuuuIIHH3xwl7CpFI8IzMyKjBs3bpfX4N90000cd9xxjB8/nnXr1vHCCy/sds7w4cMZM2YMAMcffzxr164tee0zzzxztzaPPfYYDQ2Fj6meMmUKAwYM6MTelMcjAjPbb7T3l3ul9OnTp+XxI488wsMPP8zjjz9O7969mThxYsnX6B900EEtj7t169YyNdRWu27durFz506g8CawasttRCDpTkkbJK1o47gk3SRpjaTlkuryqsXMrC39+vXjjTfeKHls69atDBgwgN69e/P888/zxBNPdPrzf/zjH2fevHkAPPTQQ2zevLnTn2NP8pwa+iEwpZ3jpwEjsn8XAbfmWIuZWUkDBw5kwoQJjB49mssuu2yXY1OmTGHnzp3U1tZy5ZVXMn78+E5//tmzZ/PQQw9RV1fHokWLGDJkCP369ev052mP8hyWSKoBfh4Ro0scuw14JCLuybZXAxMj4vftXbO+vj78wTRmB47nnnuOY445ptplVM3bb79Nt27d6N69O48//jizZs3imWee2adrlvqaSloWEfWl2lfzHsFQYF3RdlO2b7cgkHQRhVEDRx55ZEWKMzOrhJdffpkZM2bw3nvv0bNnT26//faK11DNICi1IEbJ4UlEzAHmQGFEkGdRZmaVNGLECJ5++umq1lDNl482AUcUbQ8D1lepFjOzZFUzCBYAf5m9emg8sHVP9wfMzKzz5TY1JOkeYCIwSFITMBvoARAR3wMWAlOBNcA24Py8ajEzs7blFgQRcc4ejgfw13k9v5mZlcdLTJiZdUDfvn0BWL9+PdOnTy/ZZuLEiezpZe433ngj27Zta9kuZ1nrvDgIzMz2wuGHH96ysujeaB0E5SxrnRcHgZkl7Wtf+9oun0dw9dVXc8011zB58uSWJaN/9rOf7Xbe2rVrGT268F7Zt956i4aGBmprazn77LN3WWto1qxZ1NfXM2rUKGbPng0UFrJbv349kyZNYtKkScAHy1oD3HDDDYwePZrRo0dz4403tjxfW8td7ysvOmdm+49Fl8Or/9m51/yvx8Jp17V5uKGhgS9/+ct86UtfAmDevHk8+OCDfOUrX6F///5s3LiR8ePHM23atDY/D/jWW2+ld+/eLF++nOXLl1NX98HSad/61rc49NBDeffdd5k8eTLLly/n0ksv5YYbbmDJkiUMGjRol2stW7aMH/zgBzz55JNEBB/96Ef5xCc+wYABA8pe7rqjPCIws6SNHTuWDRs2sH79ep599lkGDBjAkCFD+PrXv05tbS0nn3wyr7zyCq+99lqb1/j1r3/d8gu5traW2tralmPz5s2jrq6OsWPHsnLlSlatWtVuPY899hif+cxn6NOnD3379uXMM8/k0UcfBcpf7rqjPCIws/1HO3+552n69OnMnz+fV199lYaGBu6++26am5tZtmwZPXr0oKampuTy08VKjRZ+97vfcf3117N06VIGDBjAzJkz93id9tZ/K3e5647yiMDMktfQ0MDcuXOZP38+06dPZ+vWrRx22GH06NGDJUuW8NJLL7V7/kknncTdd98NwIoVK1i+fDkAr7/+On369OHggw/mtddeY9GiRS3ntLX89UknncQDDzzAtm3bePPNN7n//vs58cQTO7G3u/OIwMySN2rUKN544w2GDh3KkCFDOPfcczn99NOpr69nzJgxfOQjH2n3/FmzZnH++edTW1vLmDFjGDduHADHHXccY8eOZdSoUXz4wx9mwoQJLedcdNFFnHbaaQwZMoQlS5a07K+rq2PmzJkt1/jiF7/I2LFjO20aqJRcl6HOg5ehNjuwpL4MdR46ugy1p4bMzBLnIDAzS5yDwMyqrqtNUe/P9uZr6SAws6rq1asXmzZtchh0gohg06ZN9OrVq0Pn+VVDZlZVw4YNo6mpiebm5mqXckDo1asXw4YN69A5DgIzq6oePXowfPjwapeRNE8NmZklzkFgZpY4B4GZWeIcBGZmiXMQmJklzkFgZpY4B4GZWeIcBGZmiXMQmJklzkFgZpY4B4GZWeIcBGZmics1CCRNkbRa0hpJl5c4fqSkJZKelrRc0tQ86zEzs93lFgSSugE3A6cBI4FzJI1s1eybwLyIGAs0ALfkVY+ZmZWW54hgHLAmIl6MiHeAucAZrdoE0D97fDCwPsd6zMyshDyDYCiwrmi7KdtX7GrgPElNwELgb0pdSNJFkholNfrDK8zMOleeQaAS+1p/Ft05wA8jYhgwFfixpN1qiog5EVEfEfWDBw/OoVQzs3TlGQRNwBFF28PYfernAmAeQEQ8DvQCBuVYk5mZtZJnECwFRkgaLqknhZvBC1q1eRmYDCDpGApB4LkfM7MKyi0IImIncAmwGHiOwquDVkq6VtK0rNlXgQslPQvcA8yMiNbTR2ZmlqNcP7w+IhZSuAlcvO+qosergAl51mBmZu3zO4vNzBLnIDAzS5yDwMwscQ4CM7PEOQjMzBLnIDAzS5yDwMwscQ4CM7PEOQjMzBLnIDAzS5yDwMwscQ4CM7PEOQjMzBLnIDAzS5yDwMwscQ4CM7PEOQjMzBLnIDAzS5yDwMwscQ4CM7PEOQjMzBLnIDAzS5yDwMwscQ4CM7PEOQjMzBLnIDAzS5yDwMwscbkGgaQpklZLWiPp8jbazJC0StJKST/Jsx4zM9td97wuLKkbcDNwCtAELJW0ICJWFbUZAVwBTIiIzZIOy6seMzMrLc8RwThgTUS8GBHvAHOBM1q1uRC4OSI2A0TEhhzrMTOzEvIMgqHAuqLtpmxfsaOBoyX9u6QnJE0pdSFJF0lqlNTY3NycU7lmZmnKMwhUYl+02u4OjAAmAucA35d0yG4nRcyJiPqIqB88eHCnF2pmlrI8g6AJOKJoexiwvkSbn0XEjoj4HbCaQjCYmVmF5BkES4ERkoZL6gk0AAtatXkAmAQgaRCFqaIXc6zJzMxayS0IImIncAmwGHgOmBcRKyVdK2la1mwxsEnSKmAJcFlEbMqrJjMz250iWk/b79/q6+ujsbGx2mWYmXUpkpZFRH2pY35nsZlZ4hwEZmaJcxCYmSXOQWBmljgHgZlZ4hwEZmaJKysIJP13Sf1VcIekpySdmndxZmaWv3JHBP8tIl4HTgUGA+cD1+VWlZmZVUy5QfD+AnJTgR9ExLOUXlTOzMy6mHKDYJmkhygEwWJJ/YD38ivLzMwqpdxPKLsAGAO8GBHbJB1KYXrIzMy6uHJHBB8DVkfEFknnAd8EtuZXlpmZVUq5QXArsE3SccD/BF4CfpRbVWZmVjHlBsHOKCxTegbwDxHxD0C//MoyM7NKKfcewRuSrgA+D5woqRvQI7+yzMysUsodEZwNvE3h/QSvUvgQ+u/kVpWZmVVMWUGQ/fK/GzhY0l8A2yPC9wjMzA4A5S4xMQP4DfBZYAbwpKTpeRZmZmaVUe49gm8AJ0TEBgBJg4GHgfl5FWZmZpVR7j2CD70fAplNHTjXzMz2Y+WOCB6UtBi4J9s+G1iYT0lmZlZJZQVBRFwm6SxgAoXF5uZExP25VmZmZhVR7oiAiLgPuC/HWszMrAraDQJJbwBR6hAQEdE/l6rMzKxi2g2CiPAyEmZmBzi/8sfMLHEOAjOzxOUaBJKmSFotaY2ky9tpN11SSKrPsx4zM9tdbkGQrVB6M3AaMBI4R9LIEu36AZcCT+ZVi5mZtS3PEcE4YE1EvBgR7wBzKXyeQWt/B3wb2J5jLWZm1oY8g2AosK5ouynb10LSWOCIiPh5exeSdJGkRkmNzc3NnV+pmVnC8gwCldjX8p4ESR8Cvgt8dU8Xiog5EVEfEfWDBw/uxBLNzCzPIGgCjijaHgasL9ruB4wGHpG0FhgPLPANYzOzysozCJYCIyQNl9QTaAAWvH8wIrZGxKCIqImIGuAJYFpENOZYk5mZtZJbEETETuASYDHwHDAvIlZKulbStLye18zMOqbsRef2RkQspNVy1RFxVRttJ+ZZi5mZleZ3FpuZJc5BYGaWOAeBmVniHARmZolzEJiZJc5BYGaWOAeBmVniHARmZolzEJiZJc5BYGaWOAeBmVniHARmZolzEJiZJc5BYGaWOAeBmVniHARmZolzEJiZJc5BYGaWOAeBmVniHARmZolzEJiZJc5BYGaWOAeBmVniHARmZolzEJiZJc5BYGaWOAeBmVnicg0CSVMkrZa0RtLlJY7/raRVkpZL+oWko/Ksx8zMdpdbEEjqBtwMnAaMBM6RNLJVs6eB+oioBeYD386rHjMzKy3PEcE4YE1EvBgR7wBzgTOKG0TEkojYlm0+AQzLsR4zMyshzyAYCqwr2m7K9rXlAmBRqQOSLpLUKKmxubm5E0s0M7M8g0Al9kXJhtJ5QD3wnVLHI2JORNRHRP3gwYM7sUQzM+ue47WbgCOKtocB61s3knQy8A3gExHxdo71mJlZCXmOCJYCIyQNl9QTaAAWFDeQNBa4DZgWERtyrMXMzNqQWxBExE7gEmAx8BwwLyJWSrpW0rSs2XeAvsBPJT0jaUEblzMzs5zkOTVERCwEFrbad1XR45PzfH4zM9szv7PYzCxxDgIzs8Q5CMzMEucgMDNLnIPAzCxxDgIzs8Q5CMzMEucgMDNLnIPAzCxxDgIzs8Q5CMzMEucgMDNLnIPAzCxxDgIzs8Q5CMzMEucgMDNLnIPAzCxxDgIzs8Q5CMzMEucgMDNLnIPAzCxxDgIzs8Q5CMzMEucgMDNLnIPAzCxxDgIzs8Q5CMzMEpdrEEiaImm1pDWSLi9x/CBJ92bHn5RUk2c9Zma2u9yCQFI34GbgNGAkcI6kka2aXQBsjog/A74L/N+86jEzs9LyHBGMA9ZExIsR8Q4wFzijVZszgLuyx/OByZKUY01mZtZKnkEwFFhXtN2U7SvZJiJ2AluBga0vJOkiSY2SGpubm3Mq18wsTXkGQam/7GMv2hARcyKiPiLqBw8e3CnFmZlZQZ5B0AQcUbQ9DFjfVhtJ3YGDgT/kWJOZmbWSZxAsBUZIGi6pJ9AALGjVZgHwhezxdOCXEbHbiMDMzPLTPa8LR8ROSZcAi4FuwJ0RsVLStUBjRCwA7gB+LGkNhZFAQ171mJlZabkFAUBELAQWttp3VdHj7cBn86zBzMza53cWm5klzkFgZpY4B4GZWeIcBGZmiXMQmJklzkFgZpY4B4GZWeIcBGZmiXMQmJklzkFgZpY4B4GZWeIcBGZmiXMQmJklTl1t+X9JzcBLFX7aQcDGCj9npRzIfYMDu3/uW9dVjf4dFRElP+KxywVBNUhqjIj6ateRhwO5b3Bg989967r2t/55asjMLHEOAjOzxDkIyjOn2gXk6EDuGxzY/XPfuq79qn++R2BmljiPCMzMEucgMDNLnIOgiKQpklZLWiPp8hLHD5J0b3b8SUk1la9y75TRt7+VtErSckm/kHRUNercW3vqX1G76ZJC0n7z0r09KadvkmZk37+Vkn5S6Rr3Vhk/l0dKWiLp6exnc2o16twbku6UtEHSijaOS9JNWd+XS6qrdI0tIsL/CvdJugG/BT4M9ASeBUa2avMl4HvZ4wbg3mrX3Yl9mwT0zh7P6ip9K7d/Wbt+wK+BJ4D6atfdid+7EcDTwIBs+7Bq192JfZsDzMoejwTWVrvuDvTvJKAOWNHG8anAIkDAeODJatXqEcEHxgFrIuLFiHgHmAuc0arNGcBd2eP5wGRJqmCNe2uPfYuIJRGxLdt8AhhW4Rr3RTnfO4C/A74NbK9kcfuonL5dCNwcEZsBImJDhWvcW+X0LYD+2eODgfUVrG+fRMSvgT+00+QM4EdR8ARwiKQhlaluVw6CDwwF1hVtN2X7SraJiJ3AVmBgRarbN+X0rdgFFP5S6Sr22D9JY4EjIuLnlSysE5TzvTsaOFrSv0t6QtKUilW3b8rp29XAeZKagIXA31SmtIro6P/L3HSvxpPup0r9Zd/6tbXltNkflV23pPOAeuATuVbUudrtn6QPAd8FZlaqoE5UzveuO4XpoYkURnKPShodEVtyrm1fldO3c4AfRsTfS/oY8OOsb+/lX17u9pvfJx4RfKAJOKJoexi7D0Nb2kjqTmGo2t7Qb39RTt+QdDLwDWBaRLxdodo6w5761w8YDTwiaS2F+dgFXeSGcbk/lz+LiB0R8TtgNYVg2N+V07cLgHkAEfE40IvCgm0HgrL+X1aCg+ADS4ERkoZL6knhZvCCVm0WAF/IHk8HfhnZXZ/93B77lk2d3EYhBLrKHPP72u1fRGyNiEERURMRNRTugUyLiMbqlNsh5fxcPkDhZj+SBlGYKnqxolXunXL69jIwGUDSMRSCoLmiVeZnAfCX2auHxgNbI+L31SjEU0OZiNgp6RJgMYVXM9wZESslXQs0RsQC4A4KQ9M1FEYCDdWruHxl9u07QF/gp9n975cjYlrViu6AMvvXJZXZt8XAqZJWAe8Cl0XEpupVXZ4y+/ZV4HZJX6EwbTKzi/zxhaR7KEzXDcruccwGegBExPco3POYCqwBtgHnV6dSLzFhZpY8Tw2ZmSXOQWBmljgHgZlZ4hwEZmaJcxCYmSXOQWBWQZImSupqy1zYAc5BYGaWOAeBWQmSzpP0G0nPSLpNUjdJf5T095Keyj6zYXDWdky22NtySfdLGpDt/zNJD0t6NjvnT7PL95U0X9Lzku7uIivY2gHMQWDWSraUwdnAhIgYQ+HduucCfYCnIqIO+BWFd4oC/Aj4WkTUAv9ZtP9uCstDHwf8OfD+8gFjgS9TWF//w8CE3Dtl1g4vMWG2u8nA8cDS7I/1PwE2AO8B92Zt/hn4F0kHA4dExK+y/XdRWKajHzA0Iu4HiIjtANn1fhMRTdn2M0AN8Fj+3TIrzUFgtjsBd0XEFbvslK5s1a699Vnam+4pXtn1Xfz/0KrMU0Nmu/sFMF3SYQCSDs0+w/lDFFadBfgc8FhEbAU2Szox2/954FcR8TrQJOnT2TUOktS7or0wK5P/EjFrJSJWSfom8FD2oTY7gL8G3gRGSVpG4dPpzs5O+QLwvewX/Yt8sIrk54HbstU0dwCfrWA3zMrm1UfNyiTpjxHRt9p1mHU2Tw2ZmSXOIwIzs8R5RGBmljgHgZlZ4hwEZmaJcxCYmSXOQWBmlrj/DygGuUWwkAv4AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test loss: 1.78\n",
      "Test accuracy: 0.568\n",
      "\n",
      "dropout_rate: 0.1\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test loss: 1.78\n",
      "Test accuracy: 0.568\n"
     ]
    }
   ],
   "source": [
    "epochs = 50\n",
    "batch_size = 32\n",
    "validation_split = 0.2\n",
    "\n",
    "dropout_rates = [.8, .5, .3, .2, .1]\n",
    "\n",
    "for dropout_rate in dropout_rates:\n",
    "    long_model_two = Sequential()\n",
    "    long_model_two.add(Dense(units=512, activation='relu', input_shape=(784,)))\n",
    "    #long_model.add(Dropout(rate=dropout_rate))\n",
    "    long_model_two.add(Dense(units=256, activation='relu'))\n",
    "    long_model.add(Dropout(rate=dropout_rate))\n",
    "    long_model_two.add(Dense(units=128, activation='relu'))\n",
    "    long_model.add(Dropout(rate=dropout_rate))\n",
    "    long_model_two.add(Dense(units=10, activation='softmax'))\n",
    "    long_model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "    history = long_model.fit(X_train,y_train, batch_size=batch_size, epochs=epochs, verbose=False, validation_split=validation_split)\n",
    "    print(f'\\ndropout_rate: {dropout_rate}')\n",
    "    plot_training_history(history, model)"
   ]
  },
  {
   "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": "3.6.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}