misc.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import keras\n",
    "from keras.datasets import mnist"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((60000, 28, 28), (60000,))"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X, y = mnist.load_data()[0]\n",
    "X.shape, y.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "%config InlineBackend.figure_format = 'retina'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x720 with 5 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 234,
       "width": 1038
      },
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "_, axes = plt.subplots(nrows=1, ncols=5, figsize=(18, 10))\n",
    "for ax, image, label in zip(axes[:5], X, y):\n",
    "    ax.imshow(image, cmap='gray')\n",
    "    ax.set_title('Label: {}'.format(label), fontsize=20, y=1.05)\n",
    "    ax.axis('off')\n",
    "\n",
    "plt.savefig('../erasmus/images/mnist.png', dpi=150, bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}