uchidama/keras_mnist_simple_viewer.ipynb

## keras_mnist_simple_viewer.ipynb
{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "# import modules\n",
    "from __future__ import print_function\n",
    "import keras\n",
    "from keras.datasets import mnist\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib import cm\n",
    "%matplotlib inline\n",
    "\n",
    "import random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# the data, shuffled and split between train and test sets\n",
    "(x_train, y_train), (x_test, y_test) = mnist.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "index: 28704\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP4AAAD8CAYAAABXXhlaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ncOI7HCWEE9/hKCGc+A5HCeHEdzhKiP8PITeBBTfkrs4AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x12c3ad128>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# show one mnist data\n",
    "index = random.randint(0, len(x_train)-1)\n",
    "print(\"index:\", index)\n",
    "plt.imshow(x_train[index].reshape(28, 28), cmap=cm.gray_r)\n",
    "plt.show()"
   ]
  }
 ],
 "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": "3.5.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}
	{
	"cells": [
	{
	"cell_type": "code",
	"execution_count": 1,
	"metadata": {
	"collapsed": false
	},
	"outputs": [
	{
	"name": "stderr",
	"output_type": "stream",
	"text": [
	"Using TensorFlow backend.\n"
	]
	}
	],
	"source": [
	"# import modules\n",
	"from __future__ import print_function\n",
	"import keras\n",
	"from keras.datasets import mnist\n",
	"import matplotlib.pyplot as plt\n",
	"from matplotlib import cm\n",
	"%matplotlib inline\n",
	"\n",
	"import random"
	]
	},
	{
	"cell_type": "code",
	"execution_count": 2,
	"metadata": {
	"collapsed": true
	},
	"outputs": [],
	"source": [
	"# the data, shuffled and split between train and test sets\n",
	"(x_train, y_train), (x_test, y_test) = mnist.load_data()"
	]
	},
	{
	"cell_type": "code",
	"execution_count": 3,
	"metadata": {
	"collapsed": false
	},
	"outputs": [
	{
	"name": "stdout",
	"output_type": "stream",
	"text": [
	"index: 28704\n"
	]
	},
	{
	"data": {
	"image/png": "iVBORw0KGgoAAAANSUhEUgAAAP4AAAD8CAYAAABXXhlaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ncOI7HCWEE9/hKCGc+A5HCeHEdzhKiP8PITeBBTfkrs4AAAAASUVORK5CYII=\n",
	"text/plain": [
	"<matplotlib.figure.Figure at 0x12c3ad128>"
	]
	},
	"metadata": {},
	"output_type": "display_data"
	}
	],
	"source": [
	"# show one mnist data\n",
	"index = random.randint(0, len(x_train)-1)\n",
	"print(\"index:\", index)\n",
	"plt.imshow(x_train[index].reshape(28, 28), cmap=cm.gray_r)\n",
	"plt.show()"
	]
	}
	],
	"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": "3.5.1"
	}
	},
	"nbformat": 4,
	"nbformat_minor": 0
	}