02_01MNIST.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# ! pip install tensorflow_datasets\n",
    "import tensorflow_datasets as tfds\n",
    "import tensorflow as tf\n",
    "import math\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1mDownloading and preparing dataset Unknown size (download: Unknown size, generated: Unknown size, total: Unknown size) to ~\\tensorflow_datasets\\mnist\\3.0.1...\u001b[0m\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4534a1474eb14ec9a5f7c18535d4adc8",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Dl Completed...: 0 url [00:00, ? url/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "10e0fb04c99f4b7d80a90ca6f3e98d68",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Dl Size...: 0 MiB [00:00, ? MiB/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a93e3e8d778148fa932ea6c4ef8e8eb6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Extraction completed...: 0 file [00:00, ? file/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Generating splits...:   0%|          | 0/2 [00:00<?, ? splits/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Generating train examples...: 0 examples [00:00, ? examples/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Shuffling ~\\tensorflow_datasets\\mnist\\3.0.1.incompleteYI96L8\\mnist-train.tfrecord*...:   0%|          | 0/6000…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Generating test examples...: 0 examples [00:00, ? examples/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Shuffling ~\\tensorflow_datasets\\mnist\\3.0.1.incompleteYI96L8\\mnist-test.tfrecord*...:   0%|          | 0/10000…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1mDataset mnist downloaded and prepared to ~\\tensorflow_datasets\\mnist\\3.0.1. Subsequent calls will reuse this data.\u001b[0m\n"
     ]
    }
   ],
   "source": [
    "dataset, metadata = tfds.load('mnist', as_supervised=True, with_info=True)\n",
    "train_dataset, test_dataset = dataset['train'], dataset['test']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [113]\n",
      "  [254]\n",
      "  [ 95]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [183]\n",
      "  [250]\n",
      "  [190]\n",
      "  [ 11]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [ 19]\n",
      "  [198]\n",
      "  [252]\n",
      "  [235]\n",
      "  [ 39]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [ 60]\n",
      "  [207]\n",
      "  [254]\n",
      "  [159]\n",
      "  [ 13]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  2]\n",
      "  [138]\n",
      "  [246]\n",
      "  [250]\n",
      "  [109]\n",
      "  [  1]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [ 24]\n",
      "  [153]\n",
      "  [254]\n",
      "  [225]\n",
      "  [ 96]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  8]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [ 23]\n",
      "  [208]\n",
      "  [254]\n",
      "  [228]\n",
      "  [ 40]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [ 84]\n",
      "  [207]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [157]\n",
      "  [254]\n",
      "  [209]\n",
      "  [  6]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  5]\n",
      "  [ 57]\n",
      "  [204]\n",
      "  [246]\n",
      "  [ 93]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [ 83]\n",
      "  [230]\n",
      "  [252]\n",
      "  [189]\n",
      "  [ 65]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [ 15]\n",
      "  [ 32]\n",
      "  [123]\n",
      "  [210]\n",
      "  [219]\n",
      "  [234]\n",
      "  [134]\n",
      "  [ 47]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [ 48]\n",
      "  [198]\n",
      "  [254]\n",
      "  [241]\n",
      "  [146]\n",
      "  [ 28]\n",
      "  [  0]\n",
      "  [ 11]\n",
      "  [ 95]\n",
      "  [189]\n",
      "  [222]\n",
      "  [254]\n",
      "  [225]\n",
      "  [159]\n",
      "  [118]\n",
      "  [ 20]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  4]\n",
      "  [123]\n",
      "  [225]\n",
      "  [254]\n",
      "  [254]\n",
      "  [163]\n",
      "  [241]\n",
      "  [254]\n",
      "  [238]\n",
      "  [139]\n",
      "  [ 56]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [ 16]\n",
      "  [ 38]\n",
      "  [158]\n",
      "  [254]\n",
      "  [254]\n",
      "  [254]\n",
      "  [254]\n",
      "  [169]\n",
      "  [ 56]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [ 55]\n",
      "  [200]\n",
      "  [254]\n",
      "  [252]\n",
      "  [237]\n",
      "  [150]\n",
      "  [171]\n",
      "  [252]\n",
      "  [233]\n",
      "  [ 43]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [195]\n",
      "  [254]\n",
      "  [231]\n",
      "  [111]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [126]\n",
      "  [254]\n",
      "  [233]\n",
      "  [ 47]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [115]\n",
      "  [252]\n",
      "  [252]\n",
      "  [ 76]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  3]\n",
      "  [193]\n",
      "  [254]\n",
      "  [139]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [195]\n",
      "  [254]\n",
      "  [167]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [188]\n",
      "  [254]\n",
      "  [230]\n",
      "  [ 21]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [160]\n",
      "  [254]\n",
      "  [150]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [188]\n",
      "  [255]\n",
      "  [184]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [105]\n",
      "  [254]\n",
      "  [171]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [ 55]\n",
      "  [147]\n",
      "  [251]\n",
      "  [240]\n",
      "  [ 60]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [ 10]\n",
      "  [174]\n",
      "  [252]\n",
      "  [189]\n",
      "  [123]\n",
      "  [129]\n",
      "  [217]\n",
      "  [252]\n",
      "  [254]\n",
      "  [232]\n",
      "  [ 63]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [ 10]\n",
      "  [104]\n",
      "  [222]\n",
      "  [254]\n",
      "  [254]\n",
      "  [254]\n",
      "  [162]\n",
      "  [114]\n",
      "  [ 17]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]\n",
      "\n",
      " [[  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]\n",
      "  [  0]]]\n",
      "10000\n"
     ]
    }
   ],
   "source": [
    "count = 0\n",
    "for i in test_dataset.as_numpy_iterator():\n",
    "    if count == 10 :\n",
    "        print (i[0])\n",
    "    count +=1\n",
    "\n",
    "print(count)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'tensorflow.python.data.ops.dataset_ops.PrefetchDataset'> 10000\n",
      "2 \t (28, 28, 1) \t 2\n",
      "2 \t (28, 28, 1) \t 0\n",
      "2 \t (28, 28, 1) \t 4\n",
      "2 \t (28, 28, 1) \t 8\n",
      "2 \t (28, 28, 1) \t 7\n",
      "2 \t (28, 28, 1) \t 6\n",
      "2 \t (28, 28, 1) \t 0\n",
      "2 \t (28, 28, 1) \t 6\n",
      "2 \t (28, 28, 1) \t 3\n",
      "2 \t (28, 28, 1) \t 1\n",
      "10000\n"
     ]
    }
   ],
   "source": [
    "print(type(test_dataset), len(test_dataset))\n",
    "\n",
    "count = 0\n",
    "for i in test_dataset.as_numpy_iterator():\n",
    "    if count < 10 :\n",
    "        print (len(i), '\\t', i[0].shape, '\\t', i[1])\n",
    "    count +=1\n",
    "\n",
    "print(count)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def normalize(images, labels):\n",
    "    images = tf.cast(images, tf.float32)\n",
    "    images /= 255\n",
    "    return images, labels\n",
    "\n",
    "train_dataset =  train_dataset.map(normalize)\n",
    "test_dataset  =  test_dataset.map(normalize)\n",
    "\n",
    "#train_dataset =  train_dataset.cache()\n",
    "#test_dataset  =  test_dataset.cache()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for image, label in test_dataset.take(1):\n",
    "    break\n",
    "image = image.numpy().reshape((28,28))\n",
    "\n",
    "plt.figure()\n",
    "plt.imshow(image, cmap=plt.cm.binary)\n",
    "plt.colorbar()\n",
    "plt.grid(False)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2\t0\t4\t8\t7\t\n",
      "6\t0\t6\t3\t1\t\n",
      "8\t0\t7\t9\t8\t\n",
      "4\t5\t3\t4\t0\t\n",
      "6\t6\t3\t0\t2\t\n"
     ]
    }
   ],
   "source": [
    "count = 0\n",
    "for (image, label) in test_dataset.take(25):\n",
    "    print (label.numpy(), end='\\t')\n",
    "    count += 1\n",
    "    if count % 5 == 0:\n",
    "        print()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x720 with 25 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10,10))\n",
    "i = 0\n",
    "for (image, label) in test_dataset.take(25):\n",
    "    image = image.numpy().reshape((28,28))\n",
    "    plt.subplot(5,5,i+1)\n",
    "    plt.xticks([])\n",
    "    plt.yticks([])\n",
    "    plt.grid(False)\n",
    "    plt.imshow(image, cmap='gray')\n",
    "    i += 1\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "BATCH_SIZE = 32\n",
    "train_dataset = train_dataset.cache().repeat().shuffle(60000).batch(BATCH_SIZE)\n",
    "test_dataset = test_dataset.cache().batch(BATCH_SIZE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<BatchDataset element_spec=(TensorSpec(shape=(None, 28, 28, 1), dtype=tf.float32, name=None), TensorSpec(shape=(None,), dtype=tf.int64, name=None))>\n"
     ]
    }
   ],
   "source": [
    "print (train_dataset)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = tf.keras.Sequential([\n",
    "    tf.keras.layers.Flatten(input_shape=(28, 28, 1)),\n",
    "    tf.keras.layers.Dense(128, activation='relu'),\n",
    "    tf.keras.layers.Dense(10)\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(optimizer='adam',\n",
    "              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "1875/1875 [==============================] - 8s 2ms/step - loss: 0.2606 - accuracy: 0.9257\n",
      "Epoch 2/5\n",
      "1875/1875 [==============================] - 4s 2ms/step - loss: 0.1151 - accuracy: 0.9657\n",
      "Epoch 3/5\n",
      "1875/1875 [==============================] - 4s 2ms/step - loss: 0.0755 - accuracy: 0.9771\n",
      "Epoch 4/5\n",
      "1875/1875 [==============================] - 4s 2ms/step - loss: 0.0623 - accuracy: 0.9810\n",
      "Epoch 5/5\n",
      "1875/1875 [==============================] - 4s 2ms/step - loss: 0.0452 - accuracy: 0.9867\n"
     ]
    }
   ],
   "source": [
    "hist = model.fit(train_dataset, epochs=5, steps_per_epoch=math.ceil(60000/BATCH_SIZE))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 [==============================] - 1s 3ms/step - loss: 0.0815 - accuracy: 0.9749\n",
      "Accuracy on test dataset: 0.9749000072479248\n"
     ]
    }
   ],
   "source": [
    "test_loss, test_accuracy = model.evaluate(test_dataset, steps=math.ceil(10000/BATCH_SIZE))\n",
    "print('Accuracy on test dataset:', test_accuracy)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "plt.xlabel('Epoch Number')\n",
    "plt.plot(hist.history['loss']);\n",
    "plt.plot(hist.history['accuracy']);\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "plt.xlabel('Epoch Number')\n",
    "plt.ylabel(\"Accuracy Magnitude\")\n",
    "plt.plot(hist.history['accuracy']);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.save('mnist_1.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " flatten (Flatten)           (None, 784)               0         \n",
      "                                                                 \n",
      " dense (Dense)               (None, 128)               100480    \n",
      "                                                                 \n",
      " dense_1 (Dense)             (None, 10)                1290      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 101,770\n",
      "Trainable params: 101,770\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.9.13 64-bit (microsoft store)",
   "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": "3.9.13"
  },
  "vscode": {
   "interpreter": {
    "hash": "81aa396d7ed62602242c01b807f9762de13f9c3a404432573a99be8bc4dd23e0"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}