reuters-news-wire-category.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# reuters news categorize"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "from keras.datasets import reuters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading data from https://s3.amazonaws.com/text-datasets/reuters.npz\n",
      "2113536/2110848 [==============================] - 1s 1us/step\n"
     ]
    }
   ],
   "source": [
    "(train_data, train_labels), (test_data, test_labels) = reuters.load_data(num_words=10_000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((8982,), (8982,), (2246,), (2246,))"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_data.shape, train_labels.shape, test_data.shape, test_labels.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading data from https://s3.amazonaws.com/text-datasets/reuters_word_index.json\n",
      "557056/550378 [==============================] - 0s 1us/step\n"
     ]
    }
   ],
   "source": [
    "# decode to words\n",
    "word_index = reuters.get_word_index()\n",
    "rev_word_index = dict([(val, key) for (key, val) in word_index.items()])\n",
    "def decode(news_id):\n",
    "    res = ' '.join([rev_word_index.get(i - 3, '?') for i in train_data[news_id]])\n",
    "    return res\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n",
      "? ? ? said as a result of its december acquisition of space co it expects earnings per share in 1987 of 1 15 to 1 30 dlrs per share up from 70 cts in 1986 the company said pretax net should rise to nine to 10 mln dlrs from six mln dlrs in 1986 and rental operation revenues to 19 to 22 mln dlrs from 12 5 mln dlrs it said cash flow per share this year should be 2 50 to three dlrs reuter 3\n",
      "3\n",
      "\n",
      "1\n",
      "? generale de banque sa lt ? br and lt heller overseas corp of chicago have each taken 50 pct stakes in ? company sa ? factors generale de banque said in a statement it gave no financial details of the transaction sa ? ? turnover in 1986 was 17 5 billion belgian francs reuter 3\n",
      "4\n",
      "\n",
      "2\n",
      "? shr 3 28 dlrs vs 22 cts shr diluted 2 99 dlrs vs 22 cts net 46 0 mln vs 3 328 000 avg shrs 14 0 mln vs 15 2 mln year shr 5 41 dlrs vs 1 56 dlrs shr diluted 4 94 dlrs vs 1 50 dlrs net 78 2 mln vs 25 9 mln avg shrs 14 5 mln vs 15 1 mln note earnings per share reflect the two for one split effective january 6 1987 per share amounts are calculated after preferred stock dividends loss continuing operations for the qtr 1986 includes gains of sale of investments in ? corp of 14 mln dlrs and associated companies of 4 189 000 less writedowns of investments in national ? inc of 11 8 mln and ? corp of 15 6 mln reuter 3\n",
      "3\n",
      "\n",
      "3\n",
      "? the farmers home administration the u s agriculture department's farm lending arm could lose about seven billion dlrs in outstanding principal on its severely ? borrowers or about one fourth of its farm loan portfolio the general accounting office gao said in remarks prepared for delivery to the senate agriculture committee brian crowley senior associate director of gao also said that a preliminary analysis of proposed changes in ? financial eligibility standards indicated as many as one half of ? borrowers who received new loans from the agency in 1986 would be ? under the proposed system the agency has proposed evaluating ? credit using a variety of financial ratios instead of relying solely on ? ability senate agriculture committee chairman patrick leahy d vt ? the proposed eligibility changes telling ? administrator ? clark at a hearing that they would mark a dramatic shift in the agency's purpose away from being farmers' lender of last resort toward becoming a big city bank but clark defended the new regulations saying the agency had a responsibility to ? its 70 billion dlr loan portfolio in a ? yet ? manner crowley of gao ? ? arm said the proposed credit ? system attempted to ensure that ? would make loans only to borrowers who had a reasonable change of repaying their debt reuter 3\n",
      "4\n",
      "\n",
      "4\n",
      "? seton co said its board has received a proposal from chairman and chief executive officer philip d ? to acquire seton for 15 75 dlrs per share in cash seton said the acquisition bid is subject to ? arranging the necessary financing it said he intends to ask other members of senior management to participate the company said ? owns 30 pct of seton stock and other management members another 7 5 pct seton said it has formed an independent board committee to consider the offer and has deferred the annual meeting it had scheduled for march 31 reuter 3\n",
      "4\n",
      "\n"
     ]
    }
   ],
   "source": [
    "for i in range(5):\n",
    "    print(i)\n",
    "    print(decode(i))\n",
    "    print(train_labels[i])\n",
    "    print()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# one hot encode input"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "def vectorize_seqs(seqs, dims=10_000):\n",
    "    results = np.zeros((len(seqs), dims))\n",
    "    for i, seq in enumerate(seqs):\n",
    "        results[i, seq] = 1.\n",
    "    return results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train = vectorize_seqs(train_data)\n",
    "x_test = vectorize_seqs(test_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((8982, 10000), (2246, 10000))"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_train.shape, x_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "# one hot encode category"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "def to_one_hot(labels, dims=46):\n",
    "    results = np.zeros((len(labels), dims))\n",
    "    for i, label in enumerate(labels):\n",
    "        results[i, label] = 1.\n",
    "    return results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "one_hot_train_labels = to_one_hot(train_labels)\n",
    "one_hot_test_labels = to_one_hot(test_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((8982, 46), (2246, 46))"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "one_hot_train_labels.shape, one_hot_test_labels.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras import models\n",
    "from keras import layers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING: Logging before flag parsing goes to stderr.\n",
      "W0916 13:23:31.052654 4663387584 deprecation_wrapper.py:119] From /anaconda3/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:74: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n",
      "\n",
      "W0916 13:23:31.084854 4663387584 deprecation_wrapper.py:119] From /anaconda3/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n",
      "\n",
      "W0916 13:23:31.087738 4663387584 deprecation_wrapper.py:119] From /anaconda3/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "model = models.Sequential()\n",
    "model.add(layers.Dense(64, activation='relu', input_shape=(10_000,)))\n",
    "model.add(layers.Dense(64, activation='relu'))\n",
    "model.add(layers.Dense(46, activation='softmax'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "W0916 13:24:08.987415 4663387584 deprecation_wrapper.py:119] From /anaconda3/lib/python3.7/site-packages/keras/optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
      "\n",
      "W0916 13:24:09.014467 4663387584 deprecation_wrapper.py:119] From /anaconda3/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:3295: The name tf.log is deprecated. Please use tf.math.log instead.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "model.compile(optimizer='adam',\n",
    "             loss='categorical_crossentropy',\n",
    "             metrics=['acc'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_1 (Dense)              (None, 64)                640064    \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 64)                4160      \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 46)                2990      \n",
      "=================================================================\n",
      "Total params: 647,214\n",
      "Trainable params: 647,214\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "# validation - set apart 1000 samples"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_val = x_train[:1000]\n",
    "partial_x_train = x_train[1000:]\n",
    "\n",
    "y_val = one_hot_train_labels[:1000]\n",
    "partial_y_train = one_hot_train_labels[1000:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((1000, 10000), (7982, 10000), (1000, 46), (7982, 46))"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_val.shape, partial_x_train.shape, y_val.shape, partial_y_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "# train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 7982 samples, validate on 1000 samples\n",
      "Epoch 1/9\n",
      "7982/7982 [==============================] - 1s 110us/step - loss: 0.0972 - acc: 0.9638 - val_loss: 1.0223 - val_acc: 0.8060\n",
      "Epoch 2/9\n",
      "7982/7982 [==============================] - 1s 107us/step - loss: 0.0946 - acc: 0.9614 - val_loss: 1.0177 - val_acc: 0.8100\n",
      "Epoch 3/9\n",
      "7982/7982 [==============================] - 1s 108us/step - loss: 0.0935 - acc: 0.9607 - val_loss: 1.0411 - val_acc: 0.8010\n",
      "Epoch 4/9\n",
      "7982/7982 [==============================] - 1s 109us/step - loss: 0.0893 - acc: 0.9634 - val_loss: 1.0584 - val_acc: 0.8020\n",
      "Epoch 5/9\n",
      "7982/7982 [==============================] - 1s 104us/step - loss: 0.0862 - acc: 0.9624 - val_loss: 1.0672 - val_acc: 0.8080\n",
      "Epoch 6/9\n",
      "7982/7982 [==============================] - 1s 103us/step - loss: 0.0853 - acc: 0.9609 - val_loss: 1.0696 - val_acc: 0.8070\n",
      "Epoch 7/9\n",
      "7982/7982 [==============================] - 1s 104us/step - loss: 0.0843 - acc: 0.9633 - val_loss: 1.0586 - val_acc: 0.8050\n",
      "Epoch 8/9\n",
      "7982/7982 [==============================] - 1s 106us/step - loss: 0.0870 - acc: 0.9622 - val_loss: 1.0894 - val_acc: 0.8010\n",
      "Epoch 9/9\n",
      "7982/7982 [==============================] - 1s 104us/step - loss: 0.0839 - acc: 0.9623 - val_loss: 1.0723 - val_acc: 0.8060\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(partial_x_train,\n",
    "                   partial_y_train,\n",
    "                   epochs= 9,  # 20,\n",
    "                   batch_size=512,\n",
    "                   validation_data=(x_val, y_val))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "# plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzt3XuYFPWd7/H3h4vAICgCGuQ2mBjlNg4yh5C4q3g0RomiSTRi0FWPG04Sk9VkN8GNMRLdrK5JNtEnxoRjLkYJLkti4kZWk2xQzJ5gBBVWLq7KRQZULhEEAYXhu39UTdEMPTM9wzTdw3xez9PPdFf/qurb1fD7dP2quloRgZmZGUCnUhdgZmblw6FgZmYZh4KZmWUcCmZmlnEomJlZxqFgZmYZh4K1GUnfl3RTW7ctJUmPS/rrIix3taSz0/tflnRvIW1bsZ6/lPRCa+tsYrmVkkJSl7ZetpWW31ADko4H+OuI+F1rlxERnypG28NdRPxjWy1LUgAnRsRL6bKfBE5qq+Xb4c97ClYQfyI06xgcCoak+4EhwL9J2i7pSznDA9dIegX4fdr2XyW9JmmrpPmSRuYs5yeS/iG9P0FSraS/lbRB0quSrm5l276S/k3Sm5KelvQPkv7QxOtprsa7JT0iaZukpyS9O+f5D0pakc77XUCNrON4STslHZMzbYykTZK6Snq3pN9L2pxOmynp6EaWNV3SAzmPr5C0Jp33xgZtx0n6o6Qt6Xb6rqQj0ufmp80Wp+/jpfXbNmf+4emQ2BZJSyVNKnTbNCXdHg9L+rOklyR9skHNC9P373VJ/5xO7y7pgfR1bknf2+MKWZ8Vj0PBiIgrgFeACyLiyIi4I+fpM4DhwIfSx/8OnAgcCzwDzGxi0e8CjgIGAtcAd0vq04q2dwNvpW2uTG9Naa7Gy4CvAX2Al4CvA0jqB/wc+ArQD3gZOC3fCiJiPfBH4GM5kz8BzImI3SRhchtwPMn2GwxMb6ZuJI0A7gGuSOftCwzKaVIHfD6t7/3AWcBn0ppOT9uckr6P/9Jg2V2BfwN+Q7JtPgfMlJQ7vJR32xRgFlCb1nwx8I+SzkqfuxO4MyJ6A+8GZqfTryR5zwenr/NTwM4C12dF4lCw5kyPiLciYidARPwoIrZFxNskndwpko5qZN7dwC0RsTsi5gLbaXx8O29bSZ1JOt6bI2JHRCwD7muq4AJq/EVE/Cki9pAERnU6fSKwLCLqO/bvAK81saqfkXSiSBIwOZ1GRLwUEb+NiLcjYiPwzyQB25yLgV9HxPy0/puAvTmvbVFELIiIPRGxGvhBgcsFGA8cCdweEe9ExO+BX9e/hlRj26ZRkgYDfwFMi4hdEfEccC9JsEHy3r5HUr+I2B4RC3Km9wXeExF16Wt7s8DXYkXiULDmrK2/I6mzpNslvSzpTWB1+lS/RubdnHYu9XaQdEotaduf5ISItTnP5d7fT4E15nb0uTUdn7vsSK4W2ei6gDnA+yUdD5wOBPBkWsexkh6UtC6t4wEa3065GtbwFrA55/W9V9Kv0+GxN4F/LHC52bIjYm/OtDUke2f1Gts2zS33zxGxrZHlXgO8F1iRDhGdn06/H3gMeFDSekl3pHszVkIOBavX2OVyc6d/ArgQOJtkt78ynZ533L2NbAT2sP8QyuAm2h9Mja/mLjv99N/ouiJiC8lQzMfT9c6KfZcdvo1k21WlwyaXt7KGCpJP0/XuAVaQnGHUG/hygcsFWA8MlpT7/34IsK7A+Zta7jGSeuVbbkS8GBGXkQxZ/RMwR1LPdK/waxExAvgAcD7wVwdZix0kh4LVex04oZk2vYC3ST65VpB8Si2qiKgDfgFMl1Qh6WSa7jgOpsZHgJGSPqrkbKu/ITmO0ZSfpfV8LL2fW8d2YIukgcAXC6xhDnC+pL9IDyDfwv7/T3sBbwLb023x6QbzN/U+PkVybOZL6cHwCcAFwIMF1pZXRKwF/j9wW3rwuIpk72AmgKTLJfVP91C2pLPVSTpT0uh0iPBNkuGkuoOpxQ6eQ8Hq3QZ8JT0L5O8aafNTkmGBdcAyYEEj7draZ0k+9b9GMuQwi6Tjz6fVNUbEJuAS4HaSUDkR+M9mZns4bfd6RCzOmf414FRgK0nY/KLAGpYC15IEzKvAGyQHcOv9HcleyTbg/wH/0mAR04H70vfx4w2W/Q4wCTgP2AR8D/iriFhRSG3NuIxkr2w98BDJMaDfps+dCyyVtJ3koPPkiNhFErhzSAJhOfAEyTCblZD8IzvW3kj6J+BdEdHcWUhm1kLeU7CyJ+lkSVVKjCMZmnio1HWZHY78LVVrD3qRDBkdD2wAvgX8qqQVmR2mPHxkZmYZDx+ZmVmm3Q0f9evXLyorK0tdhplZu7Jo0aJNEdG/uXbtLhQqKytZuHBhqcswM2tXJK0ppJ2Hj8zMLONQMDOzjEPBzMwyDgUzM8s4FMzMLONQMDOzjEPBzMwy7e57CmZmbW3nTvjv/4YVK+Dll6FnTzjuODj22OTvccfBMcdApw7wMdqhYGYdxhtvwPLlyW3Fin33V62C5i4D17kz9O+/f1jkhkbu/f794YgjDs1ramsOBTM7rETAunX5O//XX9/Xrls3eO97oaYGrrgChg9Pbu95T7Ln8PrryW3DhgPvb9gAL76Y3N+5M38dffo0HRy593v2PDTbphAOBTNrl/bsSYZ6Gnb+K1bAtm372h11VNLZT5y4r+MfPhwqK5NP//lUVEDfvjBiRNM1RMBbbzUfIEuWJPe3bMm/nIqK5oPj2GNhyJDiB4hDwewg7NmTdArbt8OOHcnjurp9t4aPG7sVs10EHHlk0jn27p38zb2fO61HD5BKvVX399Zb8MILB3b+L74Iu3fvazdwIJx8Mlx55f6d/3HHFe81Scm2PfJIePe7m2//zjv7h0W+AFm9Gp56CjZuhL1795//u9+Fa68tykvJOBSsQ6ir29d5b9/edvffbuyXog8hKfnE27kzdOmy7379TdpXb3O6dElCIl9gFDKt/m+XVvQsmzbt6/hzA2BNzmXcOnVKOt/hw+GCC/Z1/CefnKy73B1xBAwalNyas3cvbN68f2jU1BS/RoeClbWIZJe7/hPV668nnce2bS3rwHftKnyd9Z/+evbc9ymwZ89kjHjw4AOn19+vqDiwU87XSTe8FdKmqXaFfgquq4M330xuW7fu+5t7P9+0detg2bJ90/fsaX5dFRWNB0b9/R49YOXKfZ3/pk375u/RA046CT7wAbjmmv3H+7t1K/y9bM86dUoOWPdv9mLXbcuhYIdcXd2+T0D5xmIbTs8dImgoX+d91FHJUEJjnXdj9+sfl+MQSlvo3DkJtj59Wr+MiCRgCw2U3L/r1++7Xz/mf8wxSWd/0UX7D/kMGdIxTv8sRw4FaxO7dycdeGOde27Hn2+sFKBr1/0PrFVV7f8493S/Xr2Sztsdx6ElJdu9R4/kvWiturokXMrprBtLOBSsUbt2Nd/B19/+/Of8y6jvPI47LjnbY9y4Azv5+jMr+vQ5PD+h24E6d3YglCuHQgcVkXyRZ82axm+5Y7y5evfe15kPHw4TJuTv5I87LhmScUdv1n44FA5Te/fCq6/CK6803uk3PBulRw8YOjS5jR2bjOu+610HfoOzR4/SvCYzKz6HQjv1zjuwdm3jHf7atQceoD3mmKTDP/FEOPvsfQEwZEjyt18/f6o36+g6TCi8+GJyWt0RRzR969Zt/8ddu5amo9y+vemhnVdf3f9aLRIMGJB07uPGwSWX7Ov06zv+Xr0O/esws/alw4TCL38JX/pS6+bt2rX5MGnJrWHw1NUdOMzT8MBt167JOfJDh8I55+zf4Q8dmnwZpqOcv21mxdNhQuGqq+CDH0yGXd5+O/l7MLfGlrFjR2HzNtSz574O/n3vO7DTf9e7Gr9Oi5lZWylaKEj6EXA+sCEiRuV5XsCdwERgB3BVRDxTrHpK8c3AxkQkewf1wSIlX7jyeL6ZlVoxv/rzE+DcJp4/DzgxvU0F7iliLWVFSi5ZUH/phKOPdiCYWXkoWihExHygka80AXAh8NNILACOljSgWPWYmVnzSnmRgIHA2pzHtem0A0iaKmmhpIUbN248JMWZmXVEpQyFfAMmeX8QLyJmRERNRNT0L5cDA2Zmh6FShkItMDjn8SBgfYlqMTMzShsKDwN/pcR4YGtEvFrCeszMOrxinpI6C5gA9JNUC9wMdAWIiO8Dc0lOR32J5JTUq4tVi5mZFaZooRARlzXzfABF/rVRMzNrCf9EiZmZZRwKZmaWcSiYmVnGoWBmZhmHgpmZZRwKZmaWcSiYmVnGoWBmZhmHgpmZZRwKZmaWcSiYmVnGoWBmZhmHgpmZZRwKZmaWcSiYmVnGoWBmZhmHgpmZZRwKZmaWcSiYmVnGoWBmZhmHgpmZZRwKZmaWcSiYmVnGoWBmZhmHgpmZZRwKZmaWcSiYmVnGoWBmZhmHgpmZZRwKZmaWKWooSDpX0guSXpJ0Q57nh0iaJ+lZSUskTSxmPWZm1rSihYKkzsDdwHnACOAySSMaNPsKMDsixgCTge8Vqx4zM2teMfcUxgEvRcTKiHgHeBC4sEGbAHqn948C1hexHjMza0YxQ2EgsDbncW06Ldd04HJJtcBc4HP5FiRpqqSFkhZu3LixGLWamRnFDQXlmRYNHl8G/CQiBgETgfslHVBTRMyIiJqIqOnfv38RSjUzMyhuKNQCg3MeD+LA4aFrgNkAEfFHoDvQr4g1mZlZE4oZCk8DJ0oaJukIkgPJDzdo8wpwFoCk4SSh4PEhM7MSKVooRMQe4LPAY8BykrOMlkq6RdKktNnfAp+UtBiYBVwVEQ2HmMzM7BDpUsyFR8RckgPIudO+mnN/GXBaMWswM7PC+RvNZmaWcSiYmVnGoWBmZhmHgpmZZRwKZmaWcSiYmVnGoWBmZhmHgpmZZRwKZmaWcSiYmVnGoWBmZhmHgpmZZRwKZmaWcSiYmVnGoWBmZhmHgpmZZRwKZmaWcSiYmVnGoWBmZhmHgpmZZbqUugAzM4Ddu3dTW1vLrl27Sl1Ku9a9e3cGDRpE165dWzW/Q8HMykJtbS29evWisrISSaUup12KCDZv3kxtbS3Dhg1r1TI8fGRmZWHXrl307dvXgXAQJNG3b9+D2ttyKJhZ2XAgHLyD3YYOBTMzYMuWLXzve99r1bwTJ05ky5YtBbefPn063/zmN1u1rmJzKJhZuzRzJlRWQqdOyd+ZMw9ueU2FQl1dXZPzzp07l6OPPvrgCigTDgUza3dmzoSpU2HNGohI/k6denDBcMMNN/Dyyy9TXV3NF7/4RR5//HHOPPNMPvGJTzB69GgALrroIsaOHcvIkSOZMWNGNm9lZSWbNm1i9erVDB8+nE9+8pOMHDmSc845h507dza53ueee47x48dTVVXFRz7yEd544w0A7rrrLkaMGEFVVRWTJ08G4IknnqC6uprq6mrGjBnDtm3bWv+CGxMR7eo2duzYMLPDz7JlywpuO3RoRBIH+9+GDm39+letWhUjR47MHs+bNy8qKipi5cqV2bTNmzdHRMSOHTti5MiRsWnTprSeobFx48ZYtWpVdO7cOZ599tmIiLjkkkvi/vvvP2BdN998c3zjG9+IiIjRo0fH448/HhERN910U1x33XURETFgwIDYtWtXRES88cYbERFx/vnnxx/+8IeIiNi2bVvs3r0772vJty2BhVFAH+s9BTNrd155pWXTW2vcuHH7ndp51113ccoppzB+/HjWrl3Liy++eMA8w4YNo7q6GoCxY8eyevXqRpe/detWtmzZwhlnnAHAlVdeyfz58wGoqqpiypQpPPDAA3Tpknx74LTTTuMLX/gCd911F1u2bMmmtyWHgpm1O0OGtGx6a/Xs2TO7//jjj/O73/2OP/7xjyxevJgxY8bkPfWzW7du2f3OnTuzZ8+eVq37kUce4dprr2XRokWMHTuWPXv2cMMNN3Dvvfeyc+dOxo8fz4oVK1q17KYUFAqSrpPUW4kfSnpG0jkFzHeupBckvSTphkbafFzSMklLJf2spS/AzDqer38dKir2n1ZRkUxvrV69ejU5Rr9161b69OlDRUUFK1asYMGCBa1fWeqoo46iT58+PPnkkwDcf//9nHHGGezdu5e1a9dy5plncscdd7Blyxa2b9/Oyy+/zOjRo5k2bRo1NTVFCYVC9z3+T0TcKelDQH/gauDHwG8am0FSZ+Bu4INALfC0pIcjYllOmxOBvwdOi4g3JB3bytdhZh3IlCnJ3xtvTIaMhgxJAqF+emv07duX0047jVGjRnHeeefx4Q9/eL/nzz33XL7//e9TVVXFSSedxPjx4w/iFexz33338alPfYodO3Zwwgkn8OMf/5i6ujouv/xytm7dSkTw+c9/nqOPPpqbbrqJefPm0blzZ0aMGMF5553XJjXkUnL8oZlG0pKIqJJ0J/B4RDwk6dmIGNPEPO8HpkfEh9LHfw8QEbfltLkD+O+IuLfQgmtqamLhwoWFNjezdmL58uUMHz681GUcFvJtS0mLIqKmuXkLPaawSNJvgInAY5J6AXubmWcgsDbncW06Ldd7gfdK+k9JCySdm29BkqZKWihp4caNGwss2czMWqrQ4aNrgGpgZUTskHQMyRBSU/J917rhbkkX4ERgAjAIeFLSqIjY76uBETEDmAHJnkKBNZuZWQsVuqfwfuCFiNgi6XLgK8DWZuapBQbnPB4ErM/T5lcRsTsiVgEvkISEmZmVQKGhcA+wQ9IpwJeANcBPm5nnaeBEScMkHQFMBh5u0OaXwJkAkvqRDCetLLAmMzNrY4WGwp70G3EXAndGxJ1Ar6ZmiIg9wGeBx4DlwOyIWCrpFkmT0maPAZslLQPmAV+MiM2teSFmZnbwCj2msC09e+gK4C/T002b/VmfiJgLzG0w7as59wP4QnozM7MSK3RP4VLgbZLvK7xGchbRN4pWlZlZO3DkkUcCsH79ei6++OK8bSZMmEC+0+gbm15qBYVCGgQzgaMknQ/siojmjimYmXUIxx9/PHPmzCl1GW2i0MtcfBz4E3AJ8HHgKUn5Y9HMrB2aNm3afr+nMH36dL71rW+xfft2zjrrLE499VRGjx7Nr371qwPmXb16NaNGjQJg586dTJ48maqqKi699NJmL50NMGvWLEaPHs2oUaOYNm0akPyGw1VXXcWoUaMYPXo03/72t4H8l9RuS4UeU7gR+F8RsQFAUn/gd8DhEY1mVlauvx6ee65tl1ldDd/5TuPPT548meuvv57PfOYzAMyePZtHH32U7t2789BDD9G7d282bdrE+PHjmTRpUqM/e3nPPfdQUVHBkiVLWLJkCaeeemqTda1fv55p06axaNEi+vTpwznnnMMvf/lLBg8ezLp163j++ecBsl92u/3221m1ahXdunVr0a+9FarQYwqd6gMhtbkF85qZlb0xY8awYcMG1q9fz+LFi+nTpw9DhgwhIvjyl79MVVUVZ599NuvWreP1119vdDnz58/n8ssvB5LLX1dVVTW53qeffpoJEybQv39/unTpwpQpU5g/fz4nnHACK1eu5HOf+xyPPvoovXv3zpbZ8JLabanQJT4q6TFgVvr4UhqcVWRm1laa+kRfTBdffDFz5szhtddey4ZmZs6cycaNG1m0aBFdu3alsrIy7yWzczW2F5FPY9ef69OnD4sXL+axxx7j7rvvZvbs2fzoRz/ikUceYf78+Tz88MPceuutLF26tE3DodADzV8kucxEFXAKMCMiprVZFWZmZWDy5Mk8+OCDzJkzJzubaOvWrRx77LF07dqVefPmsWbNmiaXcfrppzMz/V3Q559/niVLljTZ/n3vex9PPPEEmzZtoq6ujlmzZnHGGWewadMm9u7dy8c+9jFuvfVWnnnmmUYvqd2WCo6XiPg58PM2XbuZWRkZOXIk27ZtY+DAgQwYMACAKVOmcMEFF1BTU0N1dTUnn3xyk8v49Kc/zdVXX01VVRXV1dWMGzeuyfYDBgzgtttu48wzzyQimDhxIhdeeCGLFy/m6quvZu/e5Nqjt912W6OX1G5LTV46W9I2DryIHSQXu4uI6N2m1RTAl842Ozz50tlt52Aund3knkJENHkpCzMzO7z4DCIzM8s4FMzMLONQMLOyUcjPA1vTDnYbOhTMrCx0796dzZs3OxgOQkSwefNmunfv3upltP3X4czMWmHQoEHU1tbi32E/ON27d2fQoEGtnt+hYGZloWvXrgwbNqzUZXR4Hj4yM7OMQ8HMzDIOBTMzyzgUzMws41AwM7OMQ8HMzDIOBTMzyzgUzMws41AwM7OMQ8HMzDIOBTMzyzgUzMws41AwM7OMQ8HMzDIOBTMzyxQ1FCSdK+kFSS9JuqGJdhdLCkk1xazHzMyaVrRQkNQZuBs4DxgBXCZpRJ52vYC/AZ4qVi1mZlaYYu4pjANeioiVEfEO8CBwYZ52twJ3ALuKWIuZmRWgmKEwEFib87g2nZaRNAYYHBG/bmpBkqZKWihpoX+/1cyseIoZCsozLbInpU7At4G/bW5BETEjImoioqZ///5tWKKZmeUqZijUAoNzHg8C1uc87gWMAh6XtBoYDzzsg81mZqVTzFB4GjhR0jBJRwCTgYfrn4yIrRHRLyIqI6ISWABMioiFRazJzMyaULRQiIg9wGeBx4DlwOyIWCrpFkmTirVeMzNrvS7FXHhEzAXmNpj21UbaTihmLWZm1jx/o9nMzDIOBTMzyzgUzMws41AwM7OMQ8HMzDIOBTMzyzgUzMws41AwM7OMQ8HMzDIOBTMzyzgUzMws41AwM7OMQ8HMzDIOBTMzyzgUzMws41AwM7OMQ8HMzDIOBTMzyzgUzMws41AwM7OMQ8HMzDIOBTMzyzgUzMws41AwM7OMQ8HMzDIOBTMzyzgUzMws41AwM7OMQ8HMzDIOBTMzyxQ1FCSdK+kFSS9JuiHP81+QtEzSEkn/IWloMesxM7OmFS0UJHUG7gbOA0YAl0ka0aDZs0BNRFQBc4A7ilWPmZk1r5h7CuOAlyJiZUS8AzwIXJjbICLmRcSO9OECYFAR6zEzs2YUMxQGAmtzHtem0xpzDfDv+Z6QNFXSQkkLN27c2IYlmplZrmKGgvJMi7wNpcuBGuAb+Z6PiBkRURMRNf3792/DEs3MLFeXIi67Fhic83gQsL5hI0lnAzcCZ0TE20Wsx8zMmlHMPYWngRMlDZN0BDAZeDi3gaQxwA+ASRGxoYi1mJlZAYoWChGxB/gs8BiwHJgdEUsl3SJpUtrsG8CRwL9Kek7Sw40szszMDoFiDh8REXOBuQ2mfTXn/tnFXL+ZmbWMv9FsZmYZh4KZmWUcCmZmlnEomJlZxqFgZmYZh4KZmWUcCmZmlnEomJlZxqFgZmYZh4KZmWUcCmZmlukQoTBzJlRWQqdOyd+ZM0tdkZlZeTrsQ2HmTJg6FdasgYjk79Sp5REM5RpW5VqXmRXfYR8KN94IO3bsP23HjmR6KZVrWJVrXfW1lWNYlWtd1nJ+L4GIaFe3sWPHRktIEUn3tv9NatFi2tzQofnrGjrUdeXzwAMRFRX711RRkUx3XY174IHkvZOSv66rceX8XrbF9gIWRgF9bMk7+ZbeWhoK5drJlWtYlWtd5fo+lmtdEeXbyZVrXeX6XrbV9io0FJS0bT9qampi4cKFBbevHw7JHUKqqIAZM2DKlCIUWKDKymRopqGhQ2H16kNdzT7lWlenTsl/h4Yk2Lv30NdTr1zrgvJ9L8u1rnJ9L9tqe0laFBE1zbU77I8pTJmSBMDQocmbO3Ro6QMB4OtfT8IpV0VFMr2UyrWuIUNaNv1QKde6AF55pWXTD5Vyratc38tDvr0K2Z0op1tLh4/KWTmOq0aUZ13lOuRQrnVFlO9wSLnWVa7vZVttL3xMwQ435RhWEeVdVzl2cuVaV31t5fZeHupjCiXv5Ft6cyiYFa4cO7mI8q2rXB3Ks48O+wPNZmbmA81mZtYKDgUzM8s4FMzMLONQMDOzjEPBzMwy7e7sI0kbgTxf+i5IP2BTG5bTVlxXy7iulivX2lxXyxxMXUMjon9zjdpdKBwMSQsLOSXrUHNdLeO6Wq5ca3NdLXMo6vLwkZmZZRwKZmaW6WihMKPUBTTCdbWM62q5cq3NdbVM0evqUMcUzMysaR1tT8HMzJrgUDAzs0yHCAVJP5K0QdLzpa4ll6TBkuZJWi5pqaTrSl0TgKTukv4kaXFa19dKXVMuSZ0lPSvp16WupZ6k1ZL+S9JzksrmMr6SjpY0R9KK9N/Z+8ugppPS7VR/e1PS9aWuC0DS59N/889LmiWpe6lrApB0XVrT0mJvqw5xTEHS6cB24KcRMarU9dSTNAAYEBHPSOoFLAIuiohlJa5LQM+I2C6pK/AH4LqIWFDKuupJ+gJQA/SOiPNLXQ8koQDURERZfeFJ0n3AkxFxr6QjgIqI2FLquupJ6gysA94XEa39Umpb1TKQ5N/6iIjYKWk2MDciflLiukYBDwLjgHeAR4FPR8SLxVhfh9hTiIj5wJ9LXUdDEfFqRDyT3t8GLAcGlrYqSH+TY3v6sGt6K4tPD5IGAR8G7i11LeVOUm/gdOCHABHxTjkFQuos4OVSB0KOLkAPSV2ACmB9iesBGA4siIgdEbEHeAL4SLFW1iFCoT2QVAmMAZ4qbSWJdIjmOWAD8NuIKIu6gO8AXwL2lrqQBgL4jaRFkqaWupjUCcBG4MfpcNu9knqWuqgGJgOzSl0EQESsA74JvAK8CmyNiN+UtioAngdOl9RXUgUwERhcrJU5FMqApCOBnwPXR8Sbpa4HICLqIqIaGASMS3dhS0rS+cCGiFhU6lryOC0iTgXOA65NhyxLrQtwKnBPRIwB3gJuKG1J+6TDWZOAfy11LQCS+gAXAsOA44Geki4vbVUQEcuBfwJ+SzJ0tBjYU6z1ORRKLB2z/zkwMyJ+Uep6GkqHGx4Hzi1xKQCnAZPS8fsHgf8t6YHSlpSIiPXp3w3AQyTjv6VWC9Tm7OXNIQmJcnEe8ExEvF7qQlJnA6siYmNE7AZ+AXygxDUBEBE/jIhTI+J0kqHwohxPAIdCSaUHdH8ILI+Ify51PfUk9Zd0dHq/B8l/lhWlrQoYhqNgAAAC0klEQVQi4u8jYlBEVJIMO/w+Ikr+SU5Sz/REAdLhmXNIdvlLKiJeA9ZKOimddBZQ0pMYGriMMhk6Sr0CjJdUkf7fPIvkOF/JSTo2/TsE+ChF3G5dirXgciJpFjAB6CepFrg5In5Y2qqA5JPvFcB/peP3AF+OiLklrAlgAHBfemZIJ2B2RJTN6Z9l6DjgoaQfoQvws4h4tLQlZT4HzEyHalYCV5e4HgDSsfEPAv+31LXUi4inJM0BniEZnnmW8rncxc8l9QV2A9dGxBvFWlGHOCXVzMwK4+EjMzPLOBTMzCzjUDAzs4xDwczMMg4FMzPLOBTMikzShHK6oqtZUxwKZmaWcSiYpSRdnv6OxHOSfpBeFHC7pG9JekbSf0jqn7atlrRA0hJJD6XXzUHSeyT9Lv0timckvTtd/JE5v2swM/3GLJJul7QsXc43S/TSzTIOBTNA0nDgUpIL21UDdcAUoCfJ9XlOJblk8c3pLD8FpkVEFfBfOdNnAndHxCkk1815NZ0+BrgeGEFy9dLTJB1Dcgnkkely/qG4r9KseQ4Fs8RZwFjg6fSSI2eRdN57gX9J2zwA/IWko4CjI+KJdPp9JJc27gUMjIiHACJiV0TsSNv8KSJqI2Iv8BxQCbwJ7ALulfRRoL6tWck4FMwSAu6LiOr0dlJETM/TrqnrwqiJ597OuV8HdEl/MGUcyVVyLyK5LLJZSTkUzBL/AVycczXKYyQNJfk/cnHa5hPAHyJiK/CGpL9Mp18BPJH+FkatpIvSZXRLL/yWV/o7GkelF0C8Hqguxgsza4kOcZVUs+ZExDJJXyH59bROpFejJPlhmpGSFgFbSY47AFwJfD/t9HOvPnoF8ANJt6TLuKSJ1fYCfpX+OLyAz7fxyzJrMV8l1awJkrZHxJGlrsPsUPHwkZmZZbynYGZmGe8pmJlZxqFgZmYZh4KZmWUcCmZmlnEomJlZ5n8AGomIAZ30uukAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "\n",
    "epochs = range(1, len(loss)+1)\n",
    "\n",
    "plt.plot(epochs, loss, 'bo', label='train loss')\n",
    "plt.plot(epochs, val_loss, 'b', label='valid loss')\n",
    "plt.title('training and validation loss')\n",
    "plt.xlabel('epochs')\n",
    "plt.ylabel('loss')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.clf()\n",
    "\n",
    "acc = history.history['acc']\n",
    "val_acc = history.history['val_acc']\n",
    "\n",
    "plt.plot(epochs, acc, 'bo', label='training acc')\n",
    "plt.plot(epochs, val_acc, 'b', label='validation acc')\n",
    "plt.title('training and validation acc')\n",
    "plt.xlabel('epochs')\n",
    "plt.ylabel('acc')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "# peaks at 9, do run for 9 epochs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2246/2246 [==============================] - 0s 126us/step\n"
     ]
    }
   ],
   "source": [
    "results = model.evaluate(x_test, one_hot_test_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1.1937172574224157, 0.7911843276936776]"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results\n",
    "# [1.1937172574224157, 0.7911843276936776]"
   ]
  },
  {
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}