04chollet_getting_started_reuters_topic_detection.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import tensorflow as tf\n",
    "import seaborn as sns\n",
    "# from tensorflow.keras.datasets import mnist\n",
    "from tensorflow.keras.datasets import imdb\n",
    "from tensorflow.keras.datasets import reuters\n",
    "\n",
    "from tensorflow.keras import models\n",
    "from tensorflow.keras import layers\n",
    "from tensorflow.keras import optimizers\n",
    "from tensorflow.keras import losses\n",
    "from tensorflow.keras import metrics\n",
    "\n",
    "from tensorflow.keras.utils import to_categorical\n",
    "\n",
    "plt.style.use('ggplot')\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Single Class Multilabel classification"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/reuters.npz\n",
      "2113536/2110848 [==============================] - 152s 72us/step\n"
     ]
    }
   ],
   "source": [
    "# loading the data\n",
    "(train_data, train_labels),(test_data,test_labels) = reuters.load_data(num_words=10000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(8982, 2246)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(train_data),len(test_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/reuters_word_index.json\n",
      "557056/550378 [==============================] - 19s 33us/step\n"
     ]
    }
   ],
   "source": [
    "# index to word\n",
    "word_index = reuters.get_word_index()\n",
    "reverse_word_index = dict([(value,key) \n",
    "                           for (key,value) \n",
    "                           in word_index.items() ])\n",
    "def decoded_newswire(array_list):\n",
    "    newswire = ' '.join([reverse_word_index.get(i - 3,'?')\n",
    "                        for i in array_list])\n",
    "    return newswire\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'? the great atlantic and pacific tea co said its three year 345 mln dlr capital program will be be substantially increased to ? growth and expansion plans for ? inc and ? inc over the next two years a and p said the acquisition of ? in august 1986 and ? in december helped us achieve better than expected results in the fourth quarter ended february 28 its net income from continuing operations jumped 52 6 pct to 20 7 mln dlrs or 55 cts a share in the latest quarter as sales increased 48 3 pct to 1 58 billion dlrs a and p gave no details on the expanded capital program but it did say it completed the first year of the program during 1986 a and p is 52 4 pct owned by lt ? ? of west germany reuter 3'"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "decoded_newswire(test_data[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# encoding the features\n",
    "def vectorize_sequences(sequences,dimensions=10000):\n",
    "    results = np.zeros([len(sequences),dimensions])\n",
    "    for i, sequence in enumerate(sequences):\n",
    "        results[i, sequence] = 1.\n",
    "    return results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train = vectorize_sequences(train_data)\n",
    "x_test = vectorize_sequences(test_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# encoding the category (one hot)\n",
    "def to_one_hot(labels,dimensions=46):\n",
    "    results = np.zeros([len(labels),dimensions])\n",
    "    for i, label in enumerate(labels):\n",
    "        results[i, label] = 1.\n",
    "    return results\n",
    "\n",
    "# # alternate way using inbuilt keras tools\n",
    "# one_hot_train_label = to_categorical(train_labels)\n",
    "# one_hot_test_labels = to_categorical(test_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Model definition\n",
    "\n",
    "model = models.Sequential()\n",
    "\n",
    "model.add(layers.Dense(64, \n",
    "                       activation='relu', \n",
    "                       input_shape=(10000,)))\n",
    "\n",
    "model.add(layers.Dense(64, \n",
    "                       activation='relu'))\n",
    "\n",
    "model.add(layers.Dense(46,\n",
    "                       activation='softmax'))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# compiling the model\n",
    "model.compile(optimizer='rmsprop',\n",
    "              loss='categorical_crossentropy',\n",
    "              metrics=['accuracy'])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Validation set\n",
    "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": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 7982 samples, validate on 1000 samples\n",
      "Epoch 1/20\n",
      "7982/7982 [==============================] - 3s 399us/sample - loss: 2.5486 - accuracy: 0.5219 - val_loss: 1.6995 - val_accuracy: 0.6400\n",
      "Epoch 2/20\n",
      "7982/7982 [==============================] - 2s 240us/sample - loss: 1.4016 - accuracy: 0.7124 - val_loss: 1.3102 - val_accuracy: 0.7090\n",
      "Epoch 3/20\n",
      "7982/7982 [==============================] - 2s 229us/sample - loss: 1.0505 - accuracy: 0.7766 - val_loss: 1.1324 - val_accuracy: 0.7510\n",
      "Epoch 4/20\n",
      "7982/7982 [==============================] - 2s 228us/sample - loss: 0.8252 - accuracy: 0.8269 - val_loss: 1.1052 - val_accuracy: 0.7530\n",
      "Epoch 5/20\n",
      "7982/7982 [==============================] - 2s 227us/sample - loss: 0.6559 - accuracy: 0.8619 - val_loss: 0.9573 - val_accuracy: 0.7970\n",
      "Epoch 6/20\n",
      "7982/7982 [==============================] - 2s 224us/sample - loss: 0.5241 - accuracy: 0.8870 - val_loss: 0.9313 - val_accuracy: 0.8060\n",
      "Epoch 7/20\n",
      "7982/7982 [==============================] - 2s 233us/sample - loss: 0.4184 - accuracy: 0.9139 - val_loss: 0.9273 - val_accuracy: 0.8080\n",
      "Epoch 8/20\n",
      "7982/7982 [==============================] - 2s 244us/sample - loss: 0.3449 - accuracy: 0.9288 - val_loss: 0.9014 - val_accuracy: 0.8070\n",
      "Epoch 9/20\n",
      "7982/7982 [==============================] - 2s 243us/sample - loss: 0.2839 - accuracy: 0.9376 - val_loss: 0.8915 - val_accuracy: 0.8210\n",
      "Epoch 10/20\n",
      "7982/7982 [==============================] - 2s 228us/sample - loss: 0.2388 - accuracy: 0.9455 - val_loss: 0.9052 - val_accuracy: 0.8130\n",
      "Epoch 11/20\n",
      "7982/7982 [==============================] - 2s 272us/sample - loss: 0.2038 - accuracy: 0.9494 - val_loss: 0.9172 - val_accuracy: 0.8110\n",
      "Epoch 12/20\n",
      "7982/7982 [==============================] - 3s 349us/sample - loss: 0.1851 - accuracy: 0.9503 - val_loss: 0.9541 - val_accuracy: 0.8130\n",
      "Epoch 13/20\n",
      "7982/7982 [==============================] - 2s 217us/sample - loss: 0.1669 - accuracy: 0.9511 - val_loss: 0.9675 - val_accuracy: 0.8170\n",
      "Epoch 14/20\n",
      "7982/7982 [==============================] - 2s 197us/sample - loss: 0.1507 - accuracy: 0.9531 - val_loss: 1.0054 - val_accuracy: 0.8050\n",
      "Epoch 15/20\n",
      "7982/7982 [==============================] - 2s 200us/sample - loss: 0.1371 - accuracy: 0.9563 - val_loss: 0.9806 - val_accuracy: 0.8120\n",
      "Epoch 16/20\n",
      "7982/7982 [==============================] - 2s 193us/sample - loss: 0.1314 - accuracy: 0.9545 - val_loss: 0.9616 - val_accuracy: 0.8230\n",
      "Epoch 17/20\n",
      "7982/7982 [==============================] - 2s 200us/sample - loss: 0.1257 - accuracy: 0.9559 - val_loss: 1.0720 - val_accuracy: 0.7940\n",
      "Epoch 18/20\n",
      "7982/7982 [==============================] - 2s 188us/sample - loss: 0.1221 - accuracy: 0.9572 - val_loss: 0.9958 - val_accuracy: 0.8100\n",
      "Epoch 19/20\n",
      "7982/7982 [==============================] - 2s 192us/sample - loss: 0.1146 - accuracy: 0.9557 - val_loss: 1.0513 - val_accuracy: 0.8010\n",
      "Epoch 20/20\n",
      "7982/7982 [==============================] - 1s 181us/sample - loss: 0.1161 - accuracy: 0.9558 - val_loss: 1.0367 - val_accuracy: 0.8130\n"
     ]
    }
   ],
   "source": [
    "# Training the model\n",
    "\n",
    "history = model.fit(partial_x_train,\n",
    "                   partial_y_train,\n",
    "                   epochs=20,\n",
    "                   batch_size=512,\n",
    "                   validation_data=(x_val, y_val))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot training and validation loss\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "epochs = range(1,len(loss)+1)\n",
    "\n",
    "plt.figure(figsize=(10,10))\n",
    "plt.plot(epochs,loss,'r--',label='Training Loss')\n",
    "plt.plot(epochs,val_loss,'b-',label='Validation Loss')\n",
    "plt.title('Training and Validation Loss')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot training and validation accuracy\n",
    "acc = history.history['accuracy']\n",
    "val_acc = history.history['val_accuracy']\n",
    "epochs = range(1,len(acc)+1)\n",
    "\n",
    "plt.figure(figsize=(10,10))\n",
    "plt.plot(epochs,acc,'r--',label='Training Accuracy')\n",
    "plt.plot(epochs,val_acc,'b-',label='Validation Accuracy')\n",
    "plt.title('Training and Validation Accuracy')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.legend();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 7982 samples, validate on 1000 samples\n",
      "Epoch 1/9\n",
      "7982/7982 [==============================] - 2s 311us/sample - loss: 2.5864 - accuracy: 0.5326 - val_loss: 1.7238 - val_accuracy: 0.6520\n",
      "Epoch 2/9\n",
      "7982/7982 [==============================] - 1s 168us/sample - loss: 1.4159 - accuracy: 0.7131 - val_loss: 1.3049 - val_accuracy: 0.7130\n",
      "Epoch 3/9\n",
      "7982/7982 [==============================] - 1s 173us/sample - loss: 1.0538 - accuracy: 0.7767 - val_loss: 1.1460 - val_accuracy: 0.7520\n",
      "Epoch 4/9\n",
      "7982/7982 [==============================] - 2s 191us/sample - loss: 0.8343 - accuracy: 0.8239 - val_loss: 1.0367 - val_accuracy: 0.7960\n",
      "Epoch 5/9\n",
      "7982/7982 [==============================] - 1s 177us/sample - loss: 0.6688 - accuracy: 0.8619 - val_loss: 0.9669 - val_accuracy: 0.8100\n",
      "Epoch 6/9\n",
      "7982/7982 [==============================] - 2s 197us/sample - loss: 0.5325 - accuracy: 0.8896 - val_loss: 0.9404 - val_accuracy: 0.8100\n",
      "Epoch 7/9\n",
      "7982/7982 [==============================] - 2s 220us/sample - loss: 0.4294 - accuracy: 0.9110 - val_loss: 0.9424 - val_accuracy: 0.7980\n",
      "Epoch 8/9\n",
      "7982/7982 [==============================] - 2s 195us/sample - loss: 0.3532 - accuracy: 0.9235 - val_loss: 0.8955 - val_accuracy: 0.8210\n",
      "Epoch 9/9\n",
      "7982/7982 [==============================] - 1s 187us/sample - loss: 0.2885 - accuracy: 0.9361 - val_loss: 0.9068 - val_accuracy: 0.8220\n",
      "2246/1 [====================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================================] - 0s 187us/sample - loss: 1.7273 - accuracy: 0.7867\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "test loss:  0.9731497410037957  test accuracy:  0.78673196\n"
     ]
    }
   ],
   "source": [
    "# Retraining the model from scratch\n",
    "model1 = models.Sequential()\n",
    "model1.add(layers.Dense(64,\n",
    "                        activation='relu',\n",
    "                        input_shape=(10000,)))\n",
    "model1.add(layers.Dense(64,\n",
    "                        activation='relu'))\n",
    "model1.add(layers.Dense(46,\n",
    "                        activation='softmax'))\n",
    "\n",
    "model1.compile(optimizer='rmsprop',\n",
    "               loss='categorical_crossentropy',\n",
    "               metrics=['accuracy'])\n",
    "model1.fit(partial_x_train,\n",
    "           partial_y_train,\n",
    "           epochs=9,\n",
    "           batch_size=512,\n",
    "           validation_data=(x_val, y_val))\n",
    "\n",
    "results = model1.evaluate(x_test, one_hot_test_labels)\n",
    "print('\\ntest loss: ',results[0],' test accuracy: ',results[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train = np.array(train_labels)\n",
    "y_test = np.array(test_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 3,  4,  3, ..., 25,  3, 25])"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 7982 samples, validate on 1000 samples\n",
      "Epoch 1/20\n",
      "7982/7982 [==============================] - 2s 194us/step - loss: 2.6070 - acc: 0.4648 - val_loss: 1.7200 - val_acc: 0.6350\n",
      "Epoch 2/20\n",
      "7982/7982 [==============================] - 1s 133us/step - loss: 1.3996 - acc: 0.7105 - val_loss: 1.2722 - val_acc: 0.7280\n",
      "Epoch 3/20\n",
      "7982/7982 [==============================] - 1s 134us/step - loss: 1.0386 - acc: 0.7804 - val_loss: 1.1025 - val_acc: 0.7620\n",
      "Epoch 4/20\n",
      "7982/7982 [==============================] - 1s 153us/step - loss: 0.8144 - acc: 0.8310 - val_loss: 1.0283 - val_acc: 0.7870\n",
      "Epoch 5/20\n",
      "7982/7982 [==============================] - 1s 133us/step - loss: 0.6477 - acc: 0.8641 - val_loss: 0.9536 - val_acc: 0.8040\n",
      "Epoch 6/20\n",
      "7982/7982 [==============================] - 1s 138us/step - loss: 0.5278 - acc: 0.8904 - val_loss: 0.9219 - val_acc: 0.8060\n",
      "Epoch 7/20\n",
      "7982/7982 [==============================] - 1s 133us/step - loss: 0.4197 - acc: 0.9109 - val_loss: 0.9027 - val_acc: 0.8120\n",
      "Epoch 8/20\n",
      "7982/7982 [==============================] - 1s 133us/step - loss: 0.3481 - acc: 0.9261 - val_loss: 0.8939 - val_acc: 0.8170\n",
      "Epoch 9/20\n",
      "7982/7982 [==============================] - 1s 137us/step - loss: 0.2862 - acc: 0.9346 - val_loss: 0.8857 - val_acc: 0.8200\n",
      "Epoch 10/20\n",
      "7982/7982 [==============================] - 1s 135us/step - loss: 0.2439 - acc: 0.9424 - val_loss: 0.8856 - val_acc: 0.8140\n",
      "Epoch 11/20\n",
      "7982/7982 [==============================] - 1s 133us/step - loss: 0.2092 - acc: 0.9496 - val_loss: 0.9072 - val_acc: 0.8150\n",
      "Epoch 12/20\n",
      "7982/7982 [==============================] - 1s 149us/step - loss: 0.1827 - acc: 0.9518 - val_loss: 0.9368 - val_acc: 0.8050\n",
      "Epoch 13/20\n",
      "7982/7982 [==============================] - 1s 150us/step - loss: 0.1706 - acc: 0.9528 - val_loss: 0.9458 - val_acc: 0.8160\n",
      "Epoch 14/20\n",
      "7982/7982 [==============================] - 1s 136us/step - loss: 0.1553 - acc: 0.9546 - val_loss: 0.9855 - val_acc: 0.8100\n",
      "Epoch 15/20\n",
      "7982/7982 [==============================] - 1s 142us/step - loss: 0.1416 - acc: 0.9577 - val_loss: 0.9716 - val_acc: 0.8060\n",
      "Epoch 16/20\n",
      "7982/7982 [==============================] - 1s 155us/step - loss: 0.1313 - acc: 0.9560 - val_loss: 1.0117 - val_acc: 0.7950\n",
      "Epoch 17/20\n",
      "7982/7982 [==============================] - 1s 160us/step - loss: 0.1295 - acc: 0.9563 - val_loss: 1.0141 - val_acc: 0.8040\n",
      "Epoch 18/20\n",
      "7982/7982 [==============================] - 1s 154us/step - loss: 0.1201 - acc: 0.9575 - val_loss: 1.1047 - val_acc: 0.7900\n",
      "Epoch 19/20\n",
      "7982/7982 [==============================] - 1s 140us/step - loss: 0.1177 - acc: 0.9572 - val_loss: 1.0345 - val_acc: 0.8070\n",
      "Epoch 20/20\n",
      "7982/7982 [==============================] - 1s 146us/step - loss: 0.1154 - acc: 0.9569 - val_loss: 1.0640 - val_acc: 0.8030\n"
     ]
    }
   ],
   "source": [
    "# alternate way to encode labels\n",
    "y_train = np.array(train_labels)\n",
    "y_test = np.array(test_labels)\n",
    "\n",
    "# Model definition\n",
    "\n",
    "model2 = models.Sequential()\n",
    "\n",
    "model2.add(layers.Dense(64, \n",
    "                       activation='relu', \n",
    "                       input_shape=(10000,)))\n",
    "\n",
    "model2.add(layers.Dense(64, \n",
    "                       activation='relu'))\n",
    "\n",
    "model2.add(layers.Dense(46,\n",
    "                       activation='softmax'))\n",
    "\n",
    "# compiling the model\n",
    "# loss function sparse_categorical_crossentropy\n",
    "model2.compile(optimizer='rmsprop',\n",
    "              loss='sparse_categorical_crossentropy',\n",
    "              metrics=['accuracy'])\n",
    "\n",
    "\n",
    "y_val = y_train[:1000]\n",
    "partial_y_train = y_train[1000:]\n",
    "\n",
    "# Training the model\n",
    "history = model2.fit(partial_x_train,\n",
    "                   partial_y_train,\n",
    "                   epochs=20,\n",
    "                   batch_size=512,\n",
    "                   validation_data=(x_val, y_val))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot training and validation loss\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "epochs = range(1,len(loss)+1)\n",
    "\n",
    "plt.figure(figsize=(10,10))\n",
    "plt.plot(epochs,loss,'r--',label='Training Loss')\n",
    "plt.plot(epochs,val_loss,'b-',label='Validation Loss')\n",
    "plt.title('Training and Validation Loss')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot training and validation accuracy\n",
    "acc = history.history['acc']\n",
    "val_acc = history.history['val_acc']\n",
    "epochs = range(1,len(acc)+1)\n",
    "\n",
    "plt.figure(figsize=(10,10))\n",
    "plt.plot(epochs,acc,'r--',label='Training Accuracy')\n",
    "plt.plot(epochs,val_acc,'b-',label='Validation Accuracy')\n",
    "plt.title('Training and Validation Accuracy')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.legend();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 7982 samples, validate on 1000 samples\n",
      "Epoch 1/9\n",
      "7982/7982 [==============================] - 2s 245us/step - loss: 2.6496 - acc: 0.5431 - val_loss: 1.7413 - val_acc: 0.6330\n",
      "Epoch 2/9\n",
      "7982/7982 [==============================] - 1s 171us/step - loss: 1.4260 - acc: 0.7068 - val_loss: 1.3115 - val_acc: 0.7130\n",
      "Epoch 3/9\n",
      "7982/7982 [==============================] - 1s 162us/step - loss: 1.0547 - acc: 0.7786 - val_loss: 1.1289 - val_acc: 0.7500\n",
      "Epoch 4/9\n",
      "7982/7982 [==============================] - 1s 139us/step - loss: 0.8386 - acc: 0.8188 - val_loss: 1.0553 - val_acc: 0.7770\n",
      "Epoch 5/9\n",
      "7982/7982 [==============================] - 1s 125us/step - loss: 0.6726 - acc: 0.8558 - val_loss: 0.9873 - val_acc: 0.7830\n",
      "Epoch 6/9\n",
      "7982/7982 [==============================] - 1s 157us/step - loss: 0.5386 - acc: 0.8891 - val_loss: 0.9244 - val_acc: 0.8080\n",
      "Epoch 7/9\n",
      "7982/7982 [==============================] - 1s 166us/step - loss: 0.4325 - acc: 0.9090 - val_loss: 0.9277 - val_acc: 0.7980\n",
      "Epoch 8/9\n",
      "7982/7982 [==============================] - 1s 145us/step - loss: 0.3570 - acc: 0.9241 - val_loss: 0.8869 - val_acc: 0.8110\n",
      "Epoch 9/9\n",
      "7982/7982 [==============================] - 1s 154us/step - loss: 0.2941 - acc: 0.9334 - val_loss: 0.8940 - val_acc: 0.8080\n",
      "2246/2246 [==============================] - 0s 157us/step\n",
      "\n",
      "test loss:  0.9979781796649855  test accuracy:  0.7827248441674087\n"
     ]
    }
   ],
   "source": [
    "# Retraining the model from scratch\n",
    "model2 = models.Sequential()\n",
    "model2.add(layers.Dense(64,\n",
    "                        activation='relu',\n",
    "                        input_shape=(10000,)))\n",
    "model2.add(layers.Dense(64,\n",
    "                        activation='relu'))\n",
    "model2.add(layers.Dense(46,\n",
    "                        activation='softmax'))\n",
    "\n",
    "model2.compile(optimizer='rmsprop',\n",
    "               loss='sparse_categorical_crossentropy',\n",
    "               metrics=['accuracy'])\n",
    "model2.fit(partial_x_train,\n",
    "           partial_y_train,\n",
    "           epochs=9,\n",
    "           batch_size=512,\n",
    "           validation_data=(x_val, y_val))\n",
    "\n",
    "results = model2.evaluate(x_test, y_test)\n",
    "print('\\ntest loss: ',results[0],' test accuracy: ',results[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 7982 samples, validate on 1000 samples\n",
      "Epoch 1/50\n",
      "7982/7982 [==============================] - 2s 206us/step - loss: 3.3569 - acc: 0.4243 - val_loss: 2.9186 - val_acc: 0.5730\n",
      "Epoch 2/50\n",
      "7982/7982 [==============================] - 1s 127us/step - loss: 2.6419 - acc: 0.5737 - val_loss: 2.3750 - val_acc: 0.5830\n",
      "Epoch 3/50\n",
      "7982/7982 [==============================] - 1s 150us/step - loss: 2.1531 - acc: 0.5898 - val_loss: 2.0021 - val_acc: 0.5860\n",
      "Epoch 4/50\n",
      "7982/7982 [==============================] - 1s 140us/step - loss: 1.8228 - acc: 0.6041 - val_loss: 1.7652 - val_acc: 0.5950\n",
      "Epoch 5/50\n",
      "7982/7982 [==============================] - 1s 140us/step - loss: 1.6021 - acc: 0.6201 - val_loss: 1.6208 - val_acc: 0.6120\n",
      "Epoch 6/50\n",
      "7982/7982 [==============================] - 1s 139us/step - loss: 1.4470 - acc: 0.6426 - val_loss: 1.5199 - val_acc: 0.6290\n",
      "Epoch 7/50\n",
      "7982/7982 [==============================] - 1s 132us/step - loss: 1.3290 - acc: 0.6589 - val_loss: 1.4498 - val_acc: 0.6300\n",
      "Epoch 8/50\n",
      "7982/7982 [==============================] - 1s 132us/step - loss: 1.2330 - acc: 0.6689 - val_loss: 1.4056 - val_acc: 0.6380\n",
      "Epoch 9/50\n",
      "7982/7982 [==============================] - 1s 133us/step - loss: 1.1501 - acc: 0.6931 - val_loss: 1.3729 - val_acc: 0.6570\n",
      "Epoch 10/50\n",
      "7982/7982 [==============================] - 1s 133us/step - loss: 1.0779 - acc: 0.7233 - val_loss: 1.3347 - val_acc: 0.6730\n",
      "Epoch 11/50\n",
      "7982/7982 [==============================] - 1s 138us/step - loss: 1.0140 - acc: 0.7319 - val_loss: 1.3263 - val_acc: 0.6810\n",
      "Epoch 12/50\n",
      "7982/7982 [==============================] - 1s 145us/step - loss: 0.9575 - acc: 0.7483 - val_loss: 1.3135 - val_acc: 0.6920\n",
      "Epoch 13/50\n",
      "7982/7982 [==============================] - 1s 148us/step - loss: 0.9053 - acc: 0.7716 - val_loss: 1.3048 - val_acc: 0.6960\n",
      "Epoch 14/50\n",
      "7982/7982 [==============================] - 1s 146us/step - loss: 0.8586 - acc: 0.7786 - val_loss: 1.3027 - val_acc: 0.7030\n",
      "Epoch 15/50\n",
      "7982/7982 [==============================] - 1s 146us/step - loss: 0.8193 - acc: 0.7795 - val_loss: 1.3030 - val_acc: 0.7000\n",
      "Epoch 16/50\n",
      "7982/7982 [==============================] - 1s 136us/step - loss: 0.7826 - acc: 0.7831 - val_loss: 1.3140 - val_acc: 0.6910\n",
      "Epoch 17/50\n",
      "7982/7982 [==============================] - 1s 132us/step - loss: 0.7503 - acc: 0.7865 - val_loss: 1.3340 - val_acc: 0.6960\n",
      "Epoch 18/50\n",
      "7982/7982 [==============================] - 1s 156us/step - loss: 0.7211 - acc: 0.7884 - val_loss: 1.3480 - val_acc: 0.6970\n",
      "Epoch 19/50\n",
      "7982/7982 [==============================] - 1s 186us/step - loss: 0.6946 - acc: 0.7957 - val_loss: 1.3286 - val_acc: 0.7060\n",
      "Epoch 20/50\n",
      "7982/7982 [==============================] - 1s 143us/step - loss: 0.6678 - acc: 0.8047 - val_loss: 1.3698 - val_acc: 0.7020\n",
      "Epoch 21/50\n",
      "7982/7982 [==============================] - 1s 134us/step - loss: 0.6461 - acc: 0.8146 - val_loss: 1.3687 - val_acc: 0.7030\n",
      "Epoch 22/50\n",
      "7982/7982 [==============================] - 1s 148us/step - loss: 0.6229 - acc: 0.8211 - val_loss: 1.3854 - val_acc: 0.7070\n",
      "Epoch 23/50\n",
      "7982/7982 [==============================] - 1s 153us/step - loss: 0.6014 - acc: 0.8292 - val_loss: 1.3883 - val_acc: 0.7070\n",
      "Epoch 24/50\n",
      "7982/7982 [==============================] - 1s 153us/step - loss: 0.5815 - acc: 0.8370 - val_loss: 1.4179 - val_acc: 0.7080\n",
      "Epoch 25/50\n",
      "7982/7982 [==============================] - 1s 156us/step - loss: 0.5616 - acc: 0.8435 - val_loss: 1.4160 - val_acc: 0.7080\n",
      "Epoch 26/50\n",
      "7982/7982 [==============================] - 1s 148us/step - loss: 0.5469 - acc: 0.8453 - val_loss: 1.4432 - val_acc: 0.7120\n",
      "Epoch 27/50\n",
      "7982/7982 [==============================] - 1s 139us/step - loss: 0.5271 - acc: 0.8520 - val_loss: 1.4470 - val_acc: 0.7110\n",
      "Epoch 28/50\n",
      "7982/7982 [==============================] - 1s 141us/step - loss: 0.5133 - acc: 0.8568 - val_loss: 1.4683 - val_acc: 0.7080\n",
      "Epoch 29/50\n",
      "7982/7982 [==============================] - 1s 150us/step - loss: 0.4971 - acc: 0.8609 - val_loss: 1.4786 - val_acc: 0.7090\n",
      "Epoch 30/50\n",
      "7982/7982 [==============================] - 1s 153us/step - loss: 0.4844 - acc: 0.8663 - val_loss: 1.4981 - val_acc: 0.7100\n",
      "Epoch 31/50\n",
      "7982/7982 [==============================] - 1s 157us/step - loss: 0.4693 - acc: 0.8664 - val_loss: 1.5181 - val_acc: 0.7060\n",
      "Epoch 32/50\n",
      "7982/7982 [==============================] - 1s 161us/step - loss: 0.4574 - acc: 0.8730 - val_loss: 1.5192 - val_acc: 0.7030\n",
      "Epoch 33/50\n",
      "7982/7982 [==============================] - 1s 157us/step - loss: 0.4456 - acc: 0.8790 - val_loss: 1.5528 - val_acc: 0.7060\n",
      "Epoch 34/50\n",
      "7982/7982 [==============================] - 1s 140us/step - loss: 0.4345 - acc: 0.8829 - val_loss: 1.5535 - val_acc: 0.7110\n",
      "Epoch 35/50\n",
      "7982/7982 [==============================] - 1s 140us/step - loss: 0.4223 - acc: 0.8849 - val_loss: 1.5629 - val_acc: 0.7140\n",
      "Epoch 36/50\n",
      "7982/7982 [==============================] - 1s 151us/step - loss: 0.4128 - acc: 0.8893 - val_loss: 1.6066 - val_acc: 0.7060\n",
      "Epoch 37/50\n",
      "7982/7982 [==============================] - 1s 141us/step - loss: 0.4017 - acc: 0.8908 - val_loss: 1.6213 - val_acc: 0.7100\n",
      "Epoch 38/50\n",
      "7982/7982 [==============================] - 1s 148us/step - loss: 0.3909 - acc: 0.8938 - val_loss: 1.6354 - val_acc: 0.7150\n",
      "Epoch 39/50\n",
      "7982/7982 [==============================] - 1s 145us/step - loss: 0.3800 - acc: 0.8945 - val_loss: 1.6584 - val_acc: 0.7080\n",
      "Epoch 40/50\n",
      "7982/7982 [==============================] - 1s 147us/step - loss: 0.3748 - acc: 0.8971 - val_loss: 1.6472 - val_acc: 0.7100\n",
      "Epoch 41/50\n",
      "7982/7982 [==============================] - 1s 139us/step - loss: 0.3623 - acc: 0.8998 - val_loss: 1.6847 - val_acc: 0.7080\n",
      "Epoch 42/50\n",
      "7982/7982 [==============================] - 1s 144us/step - loss: 0.3549 - acc: 0.9027 - val_loss: 1.7206 - val_acc: 0.7080\n",
      "Epoch 43/50\n",
      "7982/7982 [==============================] - 1s 146us/step - loss: 0.3476 - acc: 0.9035 - val_loss: 1.7280 - val_acc: 0.7080\n",
      "Epoch 44/50\n",
      "7982/7982 [==============================] - 1s 138us/step - loss: 0.3390 - acc: 0.9074 - val_loss: 1.8167 - val_acc: 0.7060\n",
      "Epoch 45/50\n",
      "7982/7982 [==============================] - 1s 140us/step - loss: 0.3323 - acc: 0.9100 - val_loss: 1.7621 - val_acc: 0.7090\n",
      "Epoch 46/50\n",
      "7982/7982 [==============================] - 1s 137us/step - loss: 0.3253 - acc: 0.9117 - val_loss: 1.7866 - val_acc: 0.7060\n",
      "Epoch 47/50\n",
      "7982/7982 [==============================] - 1s 140us/step - loss: 0.3205 - acc: 0.9131 - val_loss: 1.7894 - val_acc: 0.7050\n",
      "Epoch 48/50\n",
      "7982/7982 [==============================] - 1s 142us/step - loss: 0.3122 - acc: 0.9161 - val_loss: 1.8462 - val_acc: 0.7080\n",
      "Epoch 49/50\n",
      "7982/7982 [==============================] - 1s 148us/step - loss: 0.3071 - acc: 0.9178 - val_loss: 1.8712 - val_acc: 0.7020\n",
      "Epoch 50/50\n",
      "7982/7982 [==============================] - 1s 154us/step - loss: 0.3007 - acc: 0.9198 - val_loss: 1.8979 - val_acc: 0.7110\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7fcb936328d0>"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# alternate models with less hidden nodes\n",
    "# information bottleneck\n",
    "\n",
    "model3 = models.Sequential()\n",
    "model3.add(layers.Dense(64,\n",
    "                        activation='relu',\n",
    "                        input_shape=(10000,)))\n",
    "model3.add(layers.Dense(4,\n",
    "                        activation='relu'))\n",
    "model3.add(layers.Dense(46,\n",
    "                        activation='softmax'))\n",
    "\n",
    "model3.compile(optimizer='rmsprop',\n",
    "               loss='sparse_categorical_crossentropy',\n",
    "               metrics=['accuracy'])\n",
    "model3.fit(partial_x_train,\n",
    "           partial_y_train,\n",
    "           epochs=50,\n",
    "           batch_size=512,\n",
    "           validation_data=(x_val, y_val))\n",
    "# validation accuracy stuck around 70%"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# alternate models with less hidden nodes\n",
    "# information bottleneck\n",
    "\n",
    "model3 = models.Sequential()\n",
    "model3.add(layers.Dense(64,\n",
    "                        activation='relu',\n",
    "                        input_shape=(10000,)))\n",
    "model3.add(layers.Dense(4,\n",
    "                        activation='relu'))\n",
    "model3.add(layers.Dense(46,\n",
    "                        activation='softmax'))\n",
    "\n",
    "model3.compile(optimizer='rmsprop',\n",
    "               loss='sparse_categorical_crossentropy',\n",
    "               metrics=['accuracy'])\n",
    "model3.fit(partial_x_train,\n",
    "           partial_y_train,\n",
    "           epochs=20,\n",
    "           batch_size=512,\n",
    "           validation_data=(x_val, y_val))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 7982 samples, validate on 1000 samples\n",
      "Epoch 1/20\n",
      "7982/7982 [==============================] - 2s 201us/step - loss: 2.7238 - acc: 0.5276 - val_loss: 1.9829 - val_acc: 0.6640\n",
      "Epoch 2/20\n",
      "7982/7982 [==============================] - 1s 147us/step - loss: 1.5816 - acc: 0.7194 - val_loss: 1.4400 - val_acc: 0.7180\n",
      "Epoch 3/20\n",
      "7982/7982 [==============================] - 1s 132us/step - loss: 1.1283 - acc: 0.7843 - val_loss: 1.1851 - val_acc: 0.7560\n",
      "Epoch 4/20\n",
      "7982/7982 [==============================] - 1s 129us/step - loss: 0.8651 - acc: 0.8259 - val_loss: 1.0386 - val_acc: 0.7900\n",
      "Epoch 5/20\n",
      "7982/7982 [==============================] - 1s 140us/step - loss: 0.6861 - acc: 0.8647 - val_loss: 0.9480 - val_acc: 0.8040\n",
      "Epoch 6/20\n",
      "7982/7982 [==============================] - 1s 148us/step - loss: 0.5570 - acc: 0.8903 - val_loss: 0.8967 - val_acc: 0.8160\n",
      "Epoch 7/20\n",
      "7982/7982 [==============================] - 1s 133us/step - loss: 0.4596 - acc: 0.9065 - val_loss: 0.8606 - val_acc: 0.8180\n",
      "Epoch 8/20\n",
      "7982/7982 [==============================] - 1s 147us/step - loss: 0.3839 - acc: 0.9237 - val_loss: 0.8418 - val_acc: 0.8190\n",
      "Epoch 9/20\n",
      "7982/7982 [==============================] - 1s 148us/step - loss: 0.3264 - acc: 0.9349 - val_loss: 0.8317 - val_acc: 0.8180\n",
      "Epoch 10/20\n",
      "7982/7982 [==============================] - 1s 155us/step - loss: 0.2762 - acc: 0.9414 - val_loss: 0.8250 - val_acc: 0.8250\n",
      "Epoch 11/20\n",
      "7982/7982 [==============================] - 1s 149us/step - loss: 0.2416 - acc: 0.9446 - val_loss: 0.8356 - val_acc: 0.8260\n",
      "Epoch 12/20\n",
      "7982/7982 [==============================] - 1s 150us/step - loss: 0.2102 - acc: 0.9494 - val_loss: 0.8356 - val_acc: 0.8240\n",
      "Epoch 13/20\n",
      "7982/7982 [==============================] - 1s 146us/step - loss: 0.1863 - acc: 0.9520 - val_loss: 0.8498 - val_acc: 0.8250\n",
      "Epoch 14/20\n",
      "7982/7982 [==============================] - 1s 146us/step - loss: 0.1690 - acc: 0.9526 - val_loss: 0.8554 - val_acc: 0.8220\n",
      "Epoch 15/20\n",
      "7982/7982 [==============================] - 1s 149us/step - loss: 0.1519 - acc: 0.9549 - val_loss: 0.8623 - val_acc: 0.8270\n",
      "Epoch 16/20\n",
      "7982/7982 [==============================] - 1s 147us/step - loss: 0.1443 - acc: 0.9539 - val_loss: 0.8744 - val_acc: 0.8220\n",
      "Epoch 17/20\n",
      "7982/7982 [==============================] - 1s 137us/step - loss: 0.1318 - acc: 0.9564 - val_loss: 0.8853 - val_acc: 0.8240\n",
      "Epoch 18/20\n",
      "7982/7982 [==============================] - 1s 157us/step - loss: 0.1245 - acc: 0.9575 - val_loss: 0.8917 - val_acc: 0.8190\n",
      "Epoch 19/20\n",
      "7982/7982 [==============================] - 1s 160us/step - loss: 0.1161 - acc: 0.9579 - val_loss: 0.9049 - val_acc: 0.8160\n",
      "Epoch 20/20\n",
      "7982/7982 [==============================] - 1s 134us/step - loss: 0.1115 - acc: 0.9593 - val_loss: 0.9179 - val_acc: 0.8200\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7fcb9337a208>"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# alternate models with one hidden layer\n",
    "# less represenations\n",
    "\n",
    "model4 = models.Sequential()\n",
    "model4.add(layers.Dense(64,\n",
    "                        activation='relu',\n",
    "                        input_shape=(10000,)))\n",
    "model4.add(layers.Dense(46,\n",
    "                        activation='softmax'))\n",
    "\n",
    "model4.compile(optimizer='rmsprop',\n",
    "               loss='sparse_categorical_crossentropy',\n",
    "               metrics=['accuracy'])\n",
    "model4.fit(partial_x_train,\n",
    "           partial_y_train,\n",
    "           epochs=20,\n",
    "           batch_size=512,\n",
    "           validation_data=(x_val, y_val))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 7982 samples, validate on 1000 samples\n",
      "Epoch 1/50\n",
      "7982/7982 [==============================] - 2s 243us/step - loss: 2.5968 - acc: 0.4530 - val_loss: 1.6720 - val_acc: 0.6090\n",
      "Epoch 2/50\n",
      "7982/7982 [==============================] - 1s 150us/step - loss: 1.4583 - acc: 0.6668 - val_loss: 1.3625 - val_acc: 0.7000\n",
      "Epoch 3/50\n",
      "7982/7982 [==============================] - 1s 149us/step - loss: 1.1389 - acc: 0.7333 - val_loss: 1.1882 - val_acc: 0.7360\n",
      "Epoch 4/50\n",
      "7982/7982 [==============================] - 1s 149us/step - loss: 0.9300 - acc: 0.7883 - val_loss: 1.1067 - val_acc: 0.7510\n",
      "Epoch 5/50\n",
      "7982/7982 [==============================] - 1s 137us/step - loss: 0.7321 - acc: 0.8299 - val_loss: 1.1139 - val_acc: 0.7610\n",
      "Epoch 6/50\n",
      "7982/7982 [==============================] - 1s 155us/step - loss: 0.6253 - acc: 0.8497 - val_loss: 1.0449 - val_acc: 0.7780\n",
      "Epoch 7/50\n",
      "7982/7982 [==============================] - 1s 149us/step - loss: 0.4988 - acc: 0.8881 - val_loss: 1.1465 - val_acc: 0.7570\n",
      "Epoch 8/50\n",
      "7982/7982 [==============================] - 1s 142us/step - loss: 0.3971 - acc: 0.9093 - val_loss: 1.0349 - val_acc: 0.7910\n",
      "Epoch 9/50\n",
      "7982/7982 [==============================] - 1s 160us/step - loss: 0.3612 - acc: 0.9178 - val_loss: 1.0262 - val_acc: 0.7950\n",
      "Epoch 10/50\n",
      "7982/7982 [==============================] - 1s 149us/step - loss: 0.2850 - acc: 0.9341 - val_loss: 1.5105 - val_acc: 0.7100\n",
      "Epoch 11/50\n",
      "7982/7982 [==============================] - 1s 145us/step - loss: 0.2698 - acc: 0.9357 - val_loss: 1.1142 - val_acc: 0.7820\n",
      "Epoch 12/50\n",
      "7982/7982 [==============================] - 1s 139us/step - loss: 0.2395 - acc: 0.9412 - val_loss: 1.1339 - val_acc: 0.7690\n",
      "Epoch 13/50\n",
      "7982/7982 [==============================] - 1s 146us/step - loss: 0.1967 - acc: 0.9496 - val_loss: 1.2110 - val_acc: 0.7810\n",
      "Epoch 14/50\n",
      "7982/7982 [==============================] - 1s 135us/step - loss: 0.1869 - acc: 0.9503 - val_loss: 1.2680 - val_acc: 0.7750\n",
      "Epoch 15/50\n",
      "7982/7982 [==============================] - 1s 139us/step - loss: 0.1927 - acc: 0.9483 - val_loss: 1.3278 - val_acc: 0.7670\n",
      "Epoch 16/50\n",
      "7982/7982 [==============================] - 1s 135us/step - loss: 0.1566 - acc: 0.9569 - val_loss: 1.2512 - val_acc: 0.7790\n",
      "Epoch 17/50\n",
      "7982/7982 [==============================] - 1s 134us/step - loss: 0.1591 - acc: 0.9577 - val_loss: 1.5036 - val_acc: 0.7580\n",
      "Epoch 18/50\n",
      "7982/7982 [==============================] - 1s 139us/step - loss: 0.1454 - acc: 0.9579 - val_loss: 1.5529 - val_acc: 0.7400\n",
      "Epoch 19/50\n",
      "7982/7982 [==============================] - 1s 135us/step - loss: 0.1467 - acc: 0.9570 - val_loss: 1.4541 - val_acc: 0.7570\n",
      "Epoch 20/50\n",
      "7982/7982 [==============================] - 1s 138us/step - loss: 0.1486 - acc: 0.9582 - val_loss: 1.4523 - val_acc: 0.7640\n",
      "Epoch 21/50\n",
      "7982/7982 [==============================] - 1s 138us/step - loss: 0.1279 - acc: 0.9577 - val_loss: 1.2968 - val_acc: 0.7700\n",
      "Epoch 22/50\n",
      "7982/7982 [==============================] - 1s 139us/step - loss: 0.1304 - acc: 0.9599 - val_loss: 1.3932 - val_acc: 0.7740\n",
      "Epoch 23/50\n",
      "7982/7982 [==============================] - 1s 144us/step - loss: 0.1348 - acc: 0.9584 - val_loss: 1.5648 - val_acc: 0.7440\n",
      "Epoch 24/50\n",
      "7982/7982 [==============================] - 1s 135us/step - loss: 0.1300 - acc: 0.9589 - val_loss: 1.5060 - val_acc: 0.7460\n",
      "Epoch 25/50\n",
      "7982/7982 [==============================] - 1s 145us/step - loss: 0.1248 - acc: 0.9585 - val_loss: 1.4215 - val_acc: 0.7610\n",
      "Epoch 26/50\n",
      "7982/7982 [==============================] - 2s 188us/step - loss: 0.1245 - acc: 0.9589 - val_loss: 1.4643 - val_acc: 0.7490\n",
      "Epoch 27/50\n",
      "7982/7982 [==============================] - 1s 183us/step - loss: 0.1223 - acc: 0.9590 - val_loss: 1.4029 - val_acc: 0.7680\n",
      "Epoch 28/50\n",
      "7982/7982 [==============================] - 1s 166us/step - loss: 0.1192 - acc: 0.9593 - val_loss: 1.5204 - val_acc: 0.7610\n",
      "Epoch 29/50\n",
      "7982/7982 [==============================] - 1s 142us/step - loss: 0.1172 - acc: 0.9594 - val_loss: 1.4024 - val_acc: 0.7800\n",
      "Epoch 30/50\n",
      "7982/7982 [==============================] - 1s 148us/step - loss: 0.1193 - acc: 0.9573 - val_loss: 1.4077 - val_acc: 0.7710\n",
      "Epoch 31/50\n",
      "7982/7982 [==============================] - 1s 145us/step - loss: 0.1103 - acc: 0.9604 - val_loss: 1.4005 - val_acc: 0.7800\n",
      "Epoch 32/50\n",
      "7982/7982 [==============================] - 1s 166us/step - loss: 0.1107 - acc: 0.9582 - val_loss: 1.4263 - val_acc: 0.7840\n",
      "Epoch 33/50\n",
      "7982/7982 [==============================] - 1s 161us/step - loss: 0.1120 - acc: 0.9607 - val_loss: 1.3959 - val_acc: 0.7730\n",
      "Epoch 34/50\n",
      "7982/7982 [==============================] - 1s 158us/step - loss: 0.1082 - acc: 0.9592 - val_loss: 1.4190 - val_acc: 0.7720\n",
      "Epoch 35/50\n",
      "7982/7982 [==============================] - 1s 164us/step - loss: 0.1056 - acc: 0.9585 - val_loss: 1.4557 - val_acc: 0.7700\n",
      "Epoch 36/50\n",
      "7982/7982 [==============================] - 1s 146us/step - loss: 0.1028 - acc: 0.9615 - val_loss: 1.6903 - val_acc: 0.7530\n",
      "Epoch 37/50\n",
      "7982/7982 [==============================] - 1s 171us/step - loss: 0.0997 - acc: 0.9602 - val_loss: 1.6386 - val_acc: 0.7630\n",
      "Epoch 38/50\n",
      "7982/7982 [==============================] - 2s 195us/step - loss: 0.1025 - acc: 0.9600 - val_loss: 1.4561 - val_acc: 0.7750\n",
      "Epoch 39/50\n",
      "7982/7982 [==============================] - 1s 171us/step - loss: 0.0993 - acc: 0.9618 - val_loss: 1.5048 - val_acc: 0.7580\n",
      "Epoch 40/50\n",
      "7982/7982 [==============================] - 1s 162us/step - loss: 0.1008 - acc: 0.9618 - val_loss: 1.4843 - val_acc: 0.7540\n",
      "Epoch 41/50\n",
      "7982/7982 [==============================] - 1s 160us/step - loss: 0.1058 - acc: 0.9584 - val_loss: 1.5778 - val_acc: 0.7660\n",
      "Epoch 42/50\n",
      "7982/7982 [==============================] - 1s 140us/step - loss: 0.0978 - acc: 0.9597 - val_loss: 1.4666 - val_acc: 0.7780\n",
      "Epoch 43/50\n",
      "7982/7982 [==============================] - 1s 144us/step - loss: 0.0968 - acc: 0.9594 - val_loss: 1.4569 - val_acc: 0.7670\n",
      "Epoch 44/50\n",
      "7982/7982 [==============================] - 1s 145us/step - loss: 0.0984 - acc: 0.9587 - val_loss: 1.7143 - val_acc: 0.7450\n",
      "Epoch 45/50\n",
      "7982/7982 [==============================] - 1s 174us/step - loss: 0.0943 - acc: 0.9607 - val_loss: 1.4850 - val_acc: 0.7660\n",
      "Epoch 46/50\n",
      "7982/7982 [==============================] - 1s 142us/step - loss: 0.0871 - acc: 0.9635 - val_loss: 1.5295 - val_acc: 0.7550\n",
      "Epoch 47/50\n",
      "7982/7982 [==============================] - 1s 153us/step - loss: 0.0950 - acc: 0.9584 - val_loss: 1.5983 - val_acc: 0.7620\n",
      "Epoch 48/50\n",
      "7982/7982 [==============================] - 1s 144us/step - loss: 0.0924 - acc: 0.9594 - val_loss: 1.5741 - val_acc: 0.7670\n",
      "Epoch 49/50\n",
      "7982/7982 [==============================] - 1s 161us/step - loss: 0.0948 - acc: 0.9592 - val_loss: 1.5114 - val_acc: 0.7740\n",
      "Epoch 50/50\n",
      "7982/7982 [==============================] - 1s 143us/step - loss: 0.0896 - acc: 0.9587 - val_loss: 1.5702 - val_acc: 0.7660\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7fcb926ad0b8>"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# alternate models with four hidden layer\n",
    "\n",
    "\n",
    "model4 = models.Sequential()\n",
    "model4.add(layers.Dense(64,\n",
    "                        activation='relu',\n",
    "                        input_shape=(10000,)))\n",
    "\n",
    "model4.add(layers.Dense(64,\n",
    "                        activation='relu'))\n",
    "\n",
    "model4.add(layers.Dense(64,\n",
    "                        activation='relu'))\n",
    "\n",
    "model4.add(layers.Dense(64,\n",
    "                        activation='relu'))\n",
    "\n",
    "model4.add(layers.Dense(46,\n",
    "                        activation='softmax'))\n",
    "\n",
    "model4.compile(optimizer='rmsprop',\n",
    "               loss='sparse_categorical_crossentropy',\n",
    "               metrics=['accuracy'])\n",
    "\n",
    "model4.fit(partial_x_train,\n",
    "           partial_y_train,\n",
    "           epochs=50,\n",
    "           batch_size=512,\n",
    "           validation_data=(x_val, y_val))\n",
    "\n",
    "# no improvement on adding hidden layers"
   ]
  },
  {
   "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
}