Lesson 04 - Part 2 - Keras Neural Nets.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Successfully downloaded train-images-idx3-ubyte.gz 9912422 bytes.\n",
      "Extracting ./train-images-idx3-ubyte.gz\n",
      "Successfully downloaded train-labels-idx1-ubyte.gz 28881 bytes.\n",
      "Extracting ./train-labels-idx1-ubyte.gz\n",
      "Successfully downloaded t10k-images-idx3-ubyte.gz 1648877 bytes.\n",
      "Extracting ./t10k-images-idx3-ubyte.gz\n",
      "Successfully downloaded t10k-labels-idx1-ubyte.gz 4542 bytes.\n",
      "Extracting ./t10k-labels-idx1-ubyte.gz\n"
     ]
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "\n",
    "import numpy as np\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "%matplotlib inline\n",
    "\n",
    "from tensorflow.examples.tutorials.mnist import input_data\n",
    "\n",
    "mnist = input_data.read_data_sets('./', one_hot=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(20, 784)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x, y = mnist.train.next_batch(20)\n",
    "x.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "model.add(Dense(100, activation='relu', input_dim=784))\n",
    "model.add(Dense(10, activation='softmax'))\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer='adam',\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def get_batch(dataset, batch_size = 256):\n",
    "    while (1):\n",
    "        yield dataset.next_batch(batch_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[2.3045627826299424, 0.11247996794871795]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "batch_size = 256\n",
    "test_gen = get_batch(mnist.test, batch_size)\n",
    "steps_per_epoch = mnist.test.num_examples// batch_size\n",
    "model.evaluate_generator(test_gen, steps_per_epoch)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "214/214 [==============================] - 1s - loss: 0.4851 - acc: 0.8688     \n",
      "Epoch 2/10\n",
      "214/214 [==============================] - 1s - loss: 0.2301 - acc: 0.9363     \n",
      "Epoch 3/10\n",
      "214/214 [==============================] - 1s - loss: 0.1793 - acc: 0.9493     \n",
      "Epoch 4/10\n",
      "214/214 [==============================] - 1s - loss: 0.1464 - acc: 0.9584     \n",
      "Epoch 5/10\n",
      "214/214 [==============================] - 1s - loss: 0.1248 - acc: 0.9651     \n",
      "Epoch 6/10\n",
      "214/214 [==============================] - 1s - loss: 0.1069 - acc: 0.9702     \n",
      "Epoch 7/10\n",
      "214/214 [==============================] - 1s - loss: 0.0927 - acc: 0.9744     \n",
      "Epoch 8/10\n",
      "214/214 [==============================] - 1s - loss: 0.0827 - acc: 0.9769     \n",
      "Epoch 9/10\n",
      "214/214 [==============================] - 1s - loss: 0.0727 - acc: 0.9801     \n",
      "Epoch 10/10\n",
      "214/214 [==============================] - 1s - loss: 0.0647 - acc: 0.9815     - ETA: 1s - loss: 0\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7fd498304a90>"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "batch_size = 256\n",
    "data_gen = get_batch(mnist.train, batch_size)\n",
    "steps_per_epoch = mnist.train.num_examples//batch_size\n",
    "\n",
    "model.fit_generator(data_gen, steps_per_epoch, epochs=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.092215876573714134, 0.97073963200934577]"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.evaluate_generator(test_gen, steps_per_epoch)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It is really important to be able to reload the model after you've been training it for hours on end (usually)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from keras.models import load_model\n",
    "\n",
    "model.save('my_model.h5')\n",
    "\n",
    "model2 = load_model('my_model.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.092751604739889915, 0.97064836448598135]"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model2.evaluate_generator(test_gen, steps_per_epoch)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.1"
  },
  "latex_envs": {
   "bibliofile": "biblio.bib",
   "cite_by": "apalike",
   "current_citInitial": 1,
   "eqLabelWithNumbers": true,
   "eqNumInitial": 0
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}