CNN_Model.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "e74aa718",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "f495b137",
   "metadata": {},
   "outputs": [],
   "source": [
    "sym = pd.read_pickle('DATA/data.pickle', compression='infer')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "8ee2bea0",
   "metadata": {},
   "outputs": [],
   "source": [
    "sym = pd.DataFrame(sym)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "4850a4e2",
   "metadata": {},
   "outputs": [],
   "source": [
    "sym.img = sym.img.apply(lambda x: x.ravel())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "f33f60f7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0       +\n",
       "1       -\n",
       "2       x\n",
       "3       /\n",
       "4       (\n",
       "       ..\n",
       "4895    x\n",
       "4896    /\n",
       "4897    (\n",
       "4898    )\n",
       "4899    ,\n",
       "Name: label, Length: 4900, dtype: object"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sym['label']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "385cfb50",
   "metadata": {},
   "outputs": [],
   "source": [
    "def sym_to_num(x):\n",
    "    if x == '+':\n",
    "        return 10\n",
    "    if x == '-':\n",
    "        return 11\n",
    "    if x == 'x':\n",
    "        return 12\n",
    "    if x == '/':\n",
    "        return 13\n",
    "    if x == '(':\n",
    "        return 14\n",
    "    if x == ')':\n",
    "        return 15\n",
    "    if x == ',':\n",
    "        return 16"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "de4dc78e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0       10\n",
       "1       11\n",
       "2       12\n",
       "3       13\n",
       "4       14\n",
       "        ..\n",
       "4895    12\n",
       "4896    13\n",
       "4897    14\n",
       "4898    15\n",
       "4899    16\n",
       "Name: label, Length: 4900, dtype: int64"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sym['label'] = sym['label'].apply(sym_to_num)\n",
    "sym['label']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "a370639d",
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "eda86a69",
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_digit(X, y, idx):\n",
    "    img = X[idx].reshape(28,28)\n",
    "    plt.imshow(img, cmap='Greys',  interpolation='nearest')\n",
    "    plt.title(y[idx])\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "f68a6d52",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAARYAAAEnCAYAAABsa2xHAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8qNh9FAAAACXBIWXMAAA9hAAAPYQGoP6dpAAAT7UlEQVR4nO3df2hV9/3H8ddV61nqbi5kMbk3NKZhi3SdQ9A5s6z1x8BLM5Cl22CtY+hghW6JEEIpS0dJ2g0jwoSNtHVlQyhdp7Cp8w/3I8P86JY6nKQ0aGcti/YOk4XIem7M6s00n+8fxfvdNTH1Ju+be258PuDzxz3n3Hvffsx55XM/53NPQs45JwAwtCTfBQBYfAgWAOYIFgDmCBYA5ggWAOYIFgDmCBYA5ggWAOYIFgDmCBbMy/j4uJ5++mnF43GtXLlSoVBI7e3t04776U9/qtraWpWWlsrzPK1atUqPPfaYzp49u/BFI+cIFszLlStX9PLLLyuVSqmhoWHW4+rr6/Xzn/9cf/zjH/Xcc89pYGBAGzdu1Pnz5xeuYCyIEN8Vwnzc/PEJhUIaGxvTypUr1dbWNuOo5VZvv/22HnzwQT377LN6/vnnc1wpFtKyfBeAwhYKheb83JUrV0qSli3jx3Cx4aMQFtSNGzeUSqX097//Xd/5zndUVlamb3/72/kuC8b4VYEFtWLFCqVSKUnS6tWr1dPTo8rKyjxXBWuMWLCg+vv79cYbb+jVV19VOBzW1q1buTK0CBEsWFDr1q1TbW2tvvnNb6q7u1vOOT3zzDP5LgvGCBbkTTgc1gMPPKB33nkn36XAGMGCvBkbG9Pg4KA+9alP5bsUGGPyFvP2u9/9ThMTExofH5cknTt3Tr/+9a8lSV/+8pf13//+V9u2bdOOHTtUU1OjoqIivfPOO/rJT36iVCqltra2fJaPHGCBHObt/vvv16VLl2bcNzQ0pFgspqamJv3lL39RIpHQtWvXFI1GtWXLFrW2turBBx9c4IqRawQLAHPMsQAwR7AAMEewADBHsAAwR7AAMEewADAXuAVyU1NTunz5ssLh8Lzu9QHAlnNO4+Pjqqio0JIlHzEmcTnywgsvuPvvv995nufWrVvn+vr67uh5iUTCSaLRaAFtiUTiI8/jnIxYDh8+rObmZr344ov64he/qJ/97Geqr6/XuXPntGrVqlmfGw6HJUmJRELFxcW5KA/AHCSTSVVWVqbP0dnkZOXtxo0btW7dOr300kvpbZ/+9KfV0NCgjo6OWZ+bTCYViUTk+z7BAgRINuem+eTt5OSkzpw5o3g8nrE9Ho+rv79/2vGpVErJZDKjAShs5sEyNjamGzduqLy8PGN7eXm5RkZGph3f0dGhSCSSbtymECh8ObvcfOsVHefcjFd5Wltb5ft+uiUSiVyVBGCBmE/elpaWaunSpdNGJ6Ojo9NGMZLkeZ48z7MuA0AemY9Yli9frvXr16urqytje1dXl+rq6qzfDkAA5eRyc0tLi771rW/pc5/7nL7whS/o5Zdf1nvvvacnn3wyF28HIGByEizf+MY3dOXKFT3//PMaHh7WmjVrdOLECVVVVeXi7QAETODuIMc6FiCY8rqOBQAIFgDmCBYA5ggWAOYIFgDmCBYA5ggWAOYIFgDmCBYA5ggWAOYIFgDmCBYA5ggWAOYIFgDmCBYA5ggWAOYIFgDmCBYA5ggWAOYIFgDmCBYA5ggWAOYIFgDmCBYA5ggWAOYIFgDmCBYA5ggWAOYIFgDmCBYA5syDpb29XaFQKKNFo1HrtwEC5daf+bm0xWRZLl70M5/5jP70pz+lHy9dujQXbwMgoHISLMuWLWOUAtzFcjLHcuHCBVVUVKi6ulqPPfaY/vGPf+TibQAElPmIZePGjXrllVe0evVq/etf/9KPfvQj1dXV6ezZs/rEJz4x7fhUKqVUKpV+nEwmrUsCsMBCzjmXyzeYmJjQJz/5ST399NNqaWmZtr+9vV3PPffctO2+76u4uDiXpQFmLCZfc3wqzlsymVQkErmjczPnl5tXrFihz372s7pw4cKM+1tbW+X7frolEolclwQgx3Iyefu/UqmU3n77bT388MMz7vc8T57n5boMAAvIfMTy1FNPqbe3V0NDQ/rrX/+qr3/960omk9q5c6f1WwELZiHWoCymdS7mI5Z//vOfevzxxzU2NqaVK1eqtrZWp06dUlVVlfVbAQgo82A5dOiQ9UsCKDB8VwiAOYIFgDmCBYA5ggWAOYIFgDmCBYC5nK+8BQpBoS1ACzpGLADMESwAzBEsAMwRLADMESwAzBEsAMwRLADMsY4Fi16hrFEJ+j1vs8GIBYA5ggWAOYIFgDmCBYA5ggWAOYIFgDmCBYA51rGg4BXKOpW7CSMWAOYIFgDmCBYA5ggWAOYIFgDmCBYA5ggWAOayDpa+vj5t375dFRUVCoVCOnbsWMZ+55za29tVUVGhoqIibdmyRWfPnrWqF3ehUCg0aysEzrmPbItJ1sEyMTGhtWvXqrOzc8b9+/bt0/79+9XZ2anTp08rGo1q27ZtGh8fn3exAAqEmwdJ7ujRo+nHU1NTLhqNur1796a3Xbt2zUUiEXfgwIE7ek3f950k5/v+fErDIiKp4NtikM25aTrHMjQ0pJGREcXj8fQ2z/O0efNm9ff3W74VgAAz/a7QyMiIJKm8vDxje3l5uS5dujTjc1KplFKpVPpxMpm0LAlAHuTkqtCtE2rOudtOsnV0dCgSiaRbZWVlLkoCsIBMgyUajUr6/5HLTaOjo9NGMTe1trbK9/10SyQSliUByAPTYKmurlY0GlVXV1d62+TkpHp7e1VXVzfjczzPU3FxcUYDUNiynmO5evWq3n333fTjoaEhvfnmmyopKdGqVavU3NysPXv2qKamRjU1NdqzZ4/uvfde7dixw7RwAAGW7SWn7u7uGS+n7dy50zn34SXntrY2F41Gned5btOmTW5wcDAnl7SwOMz081Ro7W6QzbkZci5YS/6SyaQikYh83+dj0V2iUFbPziZgp1FOZHNu8l0hAOYIFgDmCBYA5ggWAOYIFgDmCBYA5viDZcAduBsuJ1tixALAHMECwBzBAsAcwQLAHMECwBzBAsAcwQLAHOtYkFOFcksE1qnYYsQCwBzBAsAcwQLAHMECwBzBAsAcwQLAHMECwBzrWDAvhbBOhTUqC48RCwBzBAsAcwQLAHMECwBzBAsAcwQLAHMECwBzBAsAc1kHS19fn7Zv366KigqFQiEdO3YsY/+uXbsUCoUyWm1trVW9MHbr/1W2LQicc7M2LLysg2ViYkJr165VZ2fnbY955JFHNDw8nG4nTpyYV5EACkvWS/rr6+tVX18/6zGe5ykajc65KACFLSdzLD09PSorK9Pq1av1xBNPaHR09LbHplIpJZPJjAagsJkHS319vX75y1/q5MmT+vGPf6zTp0/rS1/6klKp1IzHd3R0KBKJpFtlZaV1SQAWWMjNY3YrFArp6NGjamhouO0xw8PDqqqq0qFDh/TVr3512v5UKpUROslkUpWVlfJ9X8XFxXMtDXcoKBOw88EE7cJIJpOKRCJ3dG7m/LYJsVhMVVVVunDhwoz7Pc+T53m5LgPAAsr5OpYrV64okUgoFovl+q0ABETWI5arV6/q3XffTT8eGhrSm2++qZKSEpWUlKi9vV1f+9rXFIvFdPHiRT3zzDMqLS3Vo48+alo47h581Ck8WQfL3/72N23dujX9uKWlRZK0c+dOvfTSSxocHNQrr7yi999/X7FYTFu3btXhw4cVDoftqgYQaPOavM2FbCaIMH+FMHkbsB/Ru1Y25ybfFQJgjmABYI5gAWCOYAFgjmABYI4/WLaIFcIVH4mrPosRIxYA5ggWAOYIFgDmCBYA5ggWAOYIFgDmCBYA5ljHUsAKYZ0Ka1TuToxYAJgjWACYI1gAmCNYAJgjWACYI1gAmCNYAJgjWACYY4FcgLEADoWKEQsAcwQLAHMECwBzBAsAcwQLAHMECwBzBAsAc1kFS0dHhzZs2KBwOKyysjI1NDTo/PnzGcc459Te3q6KigoVFRVpy5YtOnv2rGnRi0UoFJq1BYFzbtYGzCSrYOnt7VVjY6NOnTqlrq4uXb9+XfF4XBMTE+lj9u3bp/3796uzs1OnT59WNBrVtm3bND4+bl48gIBy8zA6Ouokud7eXuecc1NTUy4ajbq9e/emj7l27ZqLRCLuwIEDd/Savu87Sc73/fmUVhAkBb4BN2Vzbs5rjsX3fUlSSUmJJGloaEgjIyOKx+PpYzzP0+bNm9Xf3z/ja6RSKSWTyYwGoLDNOVicc2ppadFDDz2kNWvWSJJGRkYkSeXl5RnHlpeXp/fdqqOjQ5FIJN0qKyvnWhKAgJhzsDQ1Nemtt97Sr371q2n7bp14dM7ddjKytbVVvu+nWyKRmGtJAAJiTt9u3r17t44fP66+vj7dd9996e3RaFTShyOXWCyW3j46OjptFHOT53nyPG8uZQAIqKxGLM45NTU16ciRIzp58qSqq6sz9ldXVysajaqrqyu9bXJyUr29vaqrq7OpGEDgZTViaWxs1Guvvabf/va3CofD6XmTSCSioqIihUIhNTc3a8+ePaqpqVFNTY327Nmje++9Vzt27MjJPwBAAGVzuUm3uSR58ODB9DFTU1Oura3NRaNR53me27RpkxscHMzJJa1Cd7v+DFIDbsrm3Aw5F6zlk8lkUpFIRL7vq7i4ON/l5FRQVtfOJmA/HsijbM5NvisEwBzBAsAcwQLAHMECwBzBAsAcf1coRwrhio/EVR/kBiMWAOYIFgDmCBYA5ggWAOYIFgDmCBYA5ggWAOYIFgDmWCC3iLH4DfnCiAWAOYIFgDmCBYA5ggWAOYIFgDmCBYA5ggWAOdaxFDDWqSCoGLEAMEewADBHsAAwR7AAMEewADBHsAAwR7AAMJdVsHR0dGjDhg0Kh8MqKytTQ0ODzp8/n3HMrl27FAqFMlptba1p0YXAOZfzBgRVVsHS29urxsZGnTp1Sl1dXbp+/bri8bgmJiYyjnvkkUc0PDycbidOnDAtGkCwZbXy9ve//33G44MHD6qsrExnzpzRpk2b0ts9z1M0GrWpEEDBmdcci+/7kqSSkpKM7T09PSorK9Pq1av1xBNPaHR0dD5vA6DAhNwcP6w75/SVr3xF//73v/X666+ntx8+fFgf//jHVVVVpaGhIT377LO6fv26zpw5I8/zpr1OKpVSKpVKP04mk6qsrJTv+youLp5LaQByIJlMKhKJ3Nm56eboe9/7nquqqnKJRGLW4y5fvuzuuece95vf/GbG/W1tbU7StOb7/lxLA5ADvu/f8bk5p49Cu3fv1vHjx9Xd3a377rtv1mNjsZiqqqp04cKFGfe3trbK9/10SyQScykJQIBkNXnrnNPu3bt19OhR9fT0qLq6+iOfc+XKFSUSCcVisRn3e54340ckAIUrqxFLY2OjXn31Vb322msKh8MaGRnRyMiIPvjgA0nS1atX9dRTT+mNN97QxYsX1dPTo+3bt6u0tFSPPvpoTv4BAIInq8nbUCg04/aDBw9q165d+uCDD9TQ0KCBgQG9//77isVi2rp1q374wx+qsrLyjt4jqwkiAAsmm3Mz649CsykqKtIf/vCHbF4SwCLEd4UAmCNYAJgjWACYI1gAmCNYAJgjWACYI1gAmCNYAJgjWACYI1gAmCNYAJgjWACYI1gAmMvq280L4eY3qJPJZJ4rAfC/bp6Td3KnlcAFy/j4uCTd8f1bACys8fFxRSKRWY+Z8136c2VqakqXL19WOBxO31jq5p37E4kEN3+aB/rRzt3Yl845jY+Pq6KiQkuWzD6LErgRy5IlS257g+7i4uK75j8xl+hHO3dbX37USOUmJm8BmCNYAJgriGDxPE9tbW38mZB5oh/t0JezC9zkLYDCVxAjFgCFhWABYI5gAWCOYAFgLvDB8uKLL6q6ulof+9jHtH79er3++uv5Linw+vr6tH37dlVUVCgUCunYsWMZ+51zam9vV0VFhYqKirRlyxadPXs2P8UGWEdHhzZs2KBwOKyysjI1NDTo/PnzGcfQlzMLdLAcPnxYzc3N+sEPfqCBgQE9/PDDqq+v13vvvZfv0gJtYmJCa9euVWdn54z79+3bp/3796uzs1OnT59WNBrVtm3b0t/Twod6e3vV2NioU6dOqaurS9evX1c8HtfExET6GPryNlyAff7zn3dPPvlkxrYHHnjAff/7389TRYVHkjt69Gj68dTUlItGo27v3r3pbdeuXXORSMQdOHAgDxUWjtHRUSfJ9fb2Oufoy9kEdsQyOTmpM2fOKB6PZ2yPx+Pq7+/PU1WFb2hoSCMjIxn96nmeNm/eTL9+BN/3JUklJSWS6MvZBDZYxsbGdOPGDZWXl2dsLy8v18jISJ6qKnw3+45+zY5zTi0tLXrooYe0Zs0aSfTlbAL37eZb3bx1wk3OuWnbkD36NTtNTU1666239Oc//3naPvpyusCOWEpLS7V06dJpyT86OjrtNwTuXDQalST6NQu7d+/W8ePH1d3dnXFLD/ry9gIbLMuXL9f69evV1dWVsb2rq0t1dXV5qqrwVVdXKxqNZvTr5OSkent76ddbOOfU1NSkI0eO6OTJk6qurs7YT1/OIq9Txx/h0KFD7p577nG/+MUv3Llz51xzc7NbsWKFu3jxYr5LC7Tx8XE3MDDgBgYGnCS3f/9+NzAw4C5duuScc27v3r0uEom4I0eOuMHBQff444+7WCzmkslknisPlu9+97suEom4np4eNzw8nG7/+c9/0sfQlzMLdLA459wLL7zgqqqq3PLly926devSl/pwe93d3U7StLZz507n3IeXSdva2lw0GnWe57lNmza5wcHB/BYdQDP1oSR38ODB9DH05cy4bQIAc4GdYwFQuAgWAOYIFgDmCBYA5ggWAOYIFgDmCBYA5ggWAOYIFgDmCBYA5ggWAOYIFgDm/g9Fj/b9AOdJvwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 2000x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2000x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2000x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2000x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2000x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2000x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2000x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2000x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2000x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2000x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.rcParams[\"figure.figsize\"] = (20,3)\n",
    "\n",
    "for i in range(10,20):\n",
    "    plot_digit(sym.img, sym.label, i)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "e6987c80",
   "metadata": {},
   "outputs": [],
   "source": [
    "import struct"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "677c27eb",
   "metadata": {},
   "outputs": [],
   "source": [
    "num = pd.DataFrame()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "9afae033",
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('DATA/train-labels.idx1-ubyte', 'rb') as lbpath:\n",
    "    magic, n = struct.unpack('>II', lbpath.read(8))\n",
    "    label = np.fromfile(lbpath, dtype=np.uint8)\n",
    "    num['label'] = label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "58dab766",
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('DATA/train-images.idx3-ubyte', 'rb') as imgpath:\n",
    "    magic, n, rows, cols = struct.unpack('>IIII', imgpath.read(16))\n",
    "    img = np.fromfile(imgpath, dtype=np.uint8).reshape(len(label), 784)\n",
    "    img = img/255\n",
    "    num['img'] = img.tolist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "c1784088",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2000x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2000x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2000x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2000x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2000x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2000x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2000x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2000x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2000x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2000x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for i in range(10,20):\n",
    "    plot_digit(img, num.label, i)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "93b44f4f",
   "metadata": {},
   "outputs": [],
   "source": [
    "DF2 = pd.DataFrame(num['img'].tolist())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "ea62d9c7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>774</th>\n",
       "      <th>775</th>\n",
       "      <th>776</th>\n",
       "      <th>777</th>\n",
       "      <th>778</th>\n",
       "      <th>779</th>\n",
       "      <th>780</th>\n",
       "      <th>781</th>\n",
       "      <th>782</th>\n",
       "      <th>783</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59995</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59996</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59997</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59998</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59999</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>60000 rows × 784 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       0    1    2    3    4    5    6    7    8    9    ...  774  775  776  \\\n",
       "0      0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0   \n",
       "1      0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0   \n",
       "2      0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0   \n",
       "3      0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0   \n",
       "4      0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0   \n",
       "...    ...  ...  ...  ...  ...  ...  ...  ...  ...  ...  ...  ...  ...  ...   \n",
       "59995  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0   \n",
       "59996  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0   \n",
       "59997  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0   \n",
       "59998  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0   \n",
       "59999  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0   \n",
       "\n",
       "       777  778  779  780  781  782  783  \n",
       "0      0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "1      0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "2      0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "3      0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "4      0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "...    ...  ...  ...  ...  ...  ...  ...  \n",
       "59995  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "59996  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "59997  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "59998  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "59999  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "\n",
       "[60000 rows x 784 columns]"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "DF2"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fc2d2ead",
   "metadata": {},
   "source": [
    "### Transforming and adding data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "128df7f2",
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.preprocessing.image import ImageDataGenerator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "503aaac7",
   "metadata": {},
   "outputs": [],
   "source": [
    "DF1 = pd.DataFrame(sym['img'].tolist())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "87faacce",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>774</th>\n",
       "      <th>775</th>\n",
       "      <th>776</th>\n",
       "      <th>777</th>\n",
       "      <th>778</th>\n",
       "      <th>779</th>\n",
       "      <th>780</th>\n",
       "      <th>781</th>\n",
       "      <th>782</th>\n",
       "      <th>783</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4895</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4896</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4897</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4898</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4899</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>4900 rows × 784 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      0    1    2    3    4    5    6    7    8    9    ...  774  775  776  \\\n",
       "0       0    0    0    0    0    0    0    0    0    0  ...    0    0    0   \n",
       "1       0    0    0    0    0    0    0    0    0    0  ...    0    0    0   \n",
       "2       0    0    0    0    0    0    0    0    0    0  ...    0    0    0   \n",
       "3       0    0    0    0    0    0    0    0    0    0  ...    0    0    0   \n",
       "4       0    0    0    0    0    0    0    0    0    0  ...    0    0    0   \n",
       "...   ...  ...  ...  ...  ...  ...  ...  ...  ...  ...  ...  ...  ...  ...   \n",
       "4895    0    0    0    0    0    0    0    0    0    0  ...    0    0    0   \n",
       "4896    0    0    0    0    0    0    0    0    0    0  ...    0    0    0   \n",
       "4897    0    0    0    0    0    0    0    0    0    0  ...    0    0    0   \n",
       "4898    0    0    0    0    0    0    0    0    0    0  ...    0    0    0   \n",
       "4899    0    0    0    0    0    0    0    0    0    0  ...    0    0    0   \n",
       "\n",
       "      777  778  779  780  781  782  783  \n",
       "0       0    0    0    0    0    0    0  \n",
       "1       0    0    0    0    0    0    0  \n",
       "2       0    0    0    0    0    0    0  \n",
       "3       0    0    0    0    0    0    0  \n",
       "4       0    0    0    0    0    0    0  \n",
       "...   ...  ...  ...  ...  ...  ...  ...  \n",
       "4895    0    0    0    0    0    0    0  \n",
       "4896    0    0    0    0    0    0    0  \n",
       "4897    0    0    0    0    0    0    0  \n",
       "4898    0    0    0    0    0    0    0  \n",
       "4899    0    0    0    0    0    0    0  \n",
       "\n",
       "[4900 rows x 784 columns]"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "DF1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "47c0bfa2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4900, 784)"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "DF1.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "04ffa0ac",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4900, 28, 28, 1)"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "DF1 = DF1.values.reshape(-1,28,28,1)\n",
    "DF1.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "45ed5f92",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[[[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]]],\n",
       "\n",
       "\n",
       "       [[[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]]],\n",
       "\n",
       "\n",
       "       [[[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]]],\n",
       "\n",
       "\n",
       "       ...,\n",
       "\n",
       "\n",
       "       [[[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]]],\n",
       "\n",
       "\n",
       "       [[[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]]],\n",
       "\n",
       "\n",
       "       [[[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]],\n",
       "\n",
       "        [[0],\n",
       "         [0],\n",
       "         [0],\n",
       "         ...,\n",
       "         [0],\n",
       "         [0],\n",
       "         [0]]]], dtype=int8)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "DF1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "4ace695e",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Applying transformations for adding data\n",
    "\n",
    "datagen = ImageDataGenerator(\n",
    "        featurewise_center=False,  # set input mean to 0 over the dataset\n",
    "        samplewise_center=False,  # set each sample mean to 0\n",
    "        featurewise_std_normalization=False,  # divide inputs by std of the dataset\n",
    "        samplewise_std_normalization=False,  # divide each input by its std\n",
    "        zca_whitening=False,  # apply ZCA whitening\n",
    "        rotation_range=10,  # randomly rotate images in the range (degrees, 0 to 180)\n",
    "        zoom_range = 0.1, # Randomly zoom image \n",
    "        width_shift_range=0.1,  # randomly shift images horizontally (fraction of total width)\n",
    "        height_shift_range=0.1,  # randomly shift images vertically (fraction of total height)\n",
    "        horizontal_flip=False,  # randomly flip images\n",
    "        vertical_flip=False)  # randomly flip images"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "b232fb51",
   "metadata": {},
   "outputs": [],
   "source": [
    "augmentation_factor = 9\n",
    "\n",
    "# Generate augmented data\n",
    "aug = []\n",
    "labl = []\n",
    "\n",
    "for i in range(augmentation_factor):\n",
    "    aug.extend(datagen.flow(DF1, batch_size=len(DF1), shuffle=False).next())\n",
    "    labl.append(sym.label)\n",
    "\n",
    "labl = pd.concat(labl, ignore_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "2a956f4b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0        10\n",
       "1        11\n",
       "2        12\n",
       "3        13\n",
       "4        14\n",
       "         ..\n",
       "44095    12\n",
       "44096    13\n",
       "44097    14\n",
       "44098    15\n",
       "44099    16\n",
       "Name: label, Length: 44100, dtype: int64"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "labl"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "36ba5a91",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(44100, 28, 28, 1)"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "aug = np.array(aug).squeeze()\n",
    "aug = aug.reshape(-1,28,28,1)\n",
    "aug.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "eeecd039",
   "metadata": {},
   "outputs": [],
   "source": [
    "DF1 = aug"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "cdaf812b",
   "metadata": {},
   "outputs": [],
   "source": [
    "DF1 = DF1.reshape(44100, 784)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "61cda226",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(44100, 784)"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "DF1.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "6e07dad7",
   "metadata": {},
   "outputs": [],
   "source": [
    "DF1 = pd.DataFrame(DF1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "d2219464",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>774</th>\n",
       "      <th>775</th>\n",
       "      <th>776</th>\n",
       "      <th>777</th>\n",
       "      <th>778</th>\n",
       "      <th>779</th>\n",
       "      <th>780</th>\n",
       "      <th>781</th>\n",
       "      <th>782</th>\n",
       "      <th>783</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44095</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44096</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44097</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44098</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44099</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>44100 rows × 784 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       0    1    2    3    4    5    6    7    8    9    ...  774  775  776  \\\n",
       "0      0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0   \n",
       "1      0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0   \n",
       "2      0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0   \n",
       "3      0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0   \n",
       "4      0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0   \n",
       "...    ...  ...  ...  ...  ...  ...  ...  ...  ...  ...  ...  ...  ...  ...   \n",
       "44095  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0   \n",
       "44096  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0   \n",
       "44097  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0   \n",
       "44098  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0   \n",
       "44099  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  0.0  ...  0.0  0.0  0.0   \n",
       "\n",
       "       777  778  779  780  781  782  783  \n",
       "0      0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "1      0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "2      0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "3      0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "4      0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "...    ...  ...  ...  ...  ...  ...  ...  \n",
       "44095  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "44096  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "44097  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "44098  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "44099  0.0  0.0  0.0  0.0  0.0  0.0  0.0  \n",
       "\n",
       "[44100 rows x 784 columns]"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "DF1"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e98b5be2",
   "metadata": {},
   "source": [
    "### Appending"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "3aa6d4af",
   "metadata": {},
   "outputs": [],
   "source": [
    "imgs = pd.concat([DF2, DF1], ignore_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "80abed2a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "104100"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(imgs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "77632b86",
   "metadata": {},
   "outputs": [],
   "source": [
    "labels = pd.concat([num['label'], labl], ignore_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "ab184f81",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "104100"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(labels)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3a0d0ee9",
   "metadata": {},
   "source": [
    "### Reshaping and encoding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "5d029493",
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.utils.np_utils import to_categorical\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "d71209bd",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = imgs\n",
    "y = labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "56a6cd2f",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = X.values.reshape(-1,28,28,1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "36678df5",
   "metadata": {},
   "outputs": [],
   "source": [
    "y = to_categorical(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "09558f4a",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train,X_test,y_train,y_test=train_test_split(X,y,random_state=42,test_size=0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "510fc649",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x19065ccf880>"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 200x200 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(2,2))\n",
    "plt.imshow(X_train[78195][:,:,0], cmap = 'Greys_r')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "00ed15d1",
   "metadata": {},
   "source": [
    "### CNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "21a49ce4",
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Dense,Conv2D,Flatten,MaxPool2D,Dropout,BatchNormalization\n",
    "from keras.optimizers import RMSprop,Adam\n",
    "from keras.callbacks import ReduceLROnPlateau"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "fdfff23b",
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "\n",
    "model.add(Conv2D(filters = 32, kernel_size = (5,5),padding = 'Same', \n",
    "                 activation ='relu', input_shape = (28,28,1)))\n",
    "model.add(Conv2D(filters = 32, kernel_size = (5,5),padding = 'Same', \n",
    "                 activation ='relu'))\n",
    "model.add(BatchNormalization(momentum=.15))\n",
    "model.add(MaxPool2D(pool_size=(2,2)))\n",
    "model.add(Dropout(0.25))\n",
    "\n",
    "model.add(Conv2D(filters = 64, kernel_size = (3,3),padding = 'Same', \n",
    "                 activation ='relu'))\n",
    "model.add(Conv2D(filters = 64, kernel_size = (3,3),padding = 'Same', \n",
    "                 activation ='relu'))\n",
    "model.add(BatchNormalization(momentum=0.15))\n",
    "model.add(MaxPool2D(pool_size=(2,2), strides=(2,2)))\n",
    "model.add(Dropout(0.25))\n",
    "\n",
    "model.add(Conv2D(filters = 32, kernel_size = (5,5),padding = 'Same', \n",
    "                 activation ='relu', input_shape = (28,28,1)))\n",
    "model.add(Conv2D(filters = 32, kernel_size = (5,5),padding = 'Same', \n",
    "                 activation ='relu'))\n",
    "model.add(BatchNormalization(momentum=.15))\n",
    "model.add(MaxPool2D(pool_size=(2,2)))\n",
    "model.add(Dropout(0.25))\n",
    "\n",
    "model.add(Flatten())\n",
    "model.add(Dense(256, activation = \"relu\"))\n",
    "model.add(Dropout(0.4))\n",
    "model.add(Dense(17, activation = \"softmax\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "4d7833a1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " conv2d (Conv2D)             (None, 28, 28, 32)        832       \n",
      "                                                                 \n",
      " conv2d_1 (Conv2D)           (None, 28, 28, 32)        25632     \n",
      "                                                                 \n",
      " batch_normalization (BatchN  (None, 28, 28, 32)       128       \n",
      " ormalization)                                                   \n",
      "                                                                 \n",
      " max_pooling2d (MaxPooling2D  (None, 14, 14, 32)       0         \n",
      " )                                                               \n",
      "                                                                 \n",
      " dropout (Dropout)           (None, 14, 14, 32)        0         \n",
      "                                                                 \n",
      " conv2d_2 (Conv2D)           (None, 14, 14, 64)        18496     \n",
      "                                                                 \n",
      " conv2d_3 (Conv2D)           (None, 14, 14, 64)        36928     \n",
      "                                                                 \n",
      " batch_normalization_1 (Batc  (None, 14, 14, 64)       256       \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " max_pooling2d_1 (MaxPooling  (None, 7, 7, 64)         0         \n",
      " 2D)                                                             \n",
      "                                                                 \n",
      " dropout_1 (Dropout)         (None, 7, 7, 64)          0         \n",
      "                                                                 \n",
      " conv2d_4 (Conv2D)           (None, 7, 7, 32)          51232     \n",
      "                                                                 \n",
      " conv2d_5 (Conv2D)           (None, 7, 7, 32)          25632     \n",
      "                                                                 \n",
      " batch_normalization_2 (Batc  (None, 7, 7, 32)         128       \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " max_pooling2d_2 (MaxPooling  (None, 3, 3, 32)         0         \n",
      " 2D)                                                             \n",
      "                                                                 \n",
      " dropout_2 (Dropout)         (None, 3, 3, 32)          0         \n",
      "                                                                 \n",
      " flatten (Flatten)           (None, 288)               0         \n",
      "                                                                 \n",
      " dense (Dense)               (None, 256)               73984     \n",
      "                                                                 \n",
      " dropout_3 (Dropout)         (None, 256)               0         \n",
      "                                                                 \n",
      " dense_1 (Dense)             (None, 17)                4369      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 237,617\n",
      "Trainable params: 237,361\n",
      "Non-trainable params: 256\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "e8fa30ca",
   "metadata": {},
   "outputs": [],
   "source": [
    "optimizer=Adam(learning_rate=0.001,beta_1=0.9,beta_2=0.999)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "505f1aa0",
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(optimizer=optimizer,loss=['categorical_crossentropy'],metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "95e6427f",
   "metadata": {},
   "outputs": [],
   "source": [
    "learning_rate_reduction = ReduceLROnPlateau(monitor='val_accuracy', \n",
    "                                            patience=3, \n",
    "                                            verbose=1, \n",
    "                                            factor=0.5, \n",
    "                                            min_lr=0.00001)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "05d2a259",
   "metadata": {},
   "outputs": [],
   "source": [
    "epochs=3\n",
    "batch_size=64\n",
    "\n",
    "# Actual model being used has been trained for 30 epochs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "16f8f93b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/3\n",
      "1301/1301 - 455s - loss: 0.3631 - accuracy: 0.8853 - val_loss: 0.0676 - val_accuracy: 0.9777 - lr: 0.0010 - 455s/epoch - 349ms/step\n",
      "Epoch 2/3\n",
      "1301/1301 - 355s - loss: 0.1198 - accuracy: 0.9610 - val_loss: 0.0564 - val_accuracy: 0.9818 - lr: 0.0010 - 355s/epoch - 273ms/step\n",
      "Epoch 3/3\n",
      "1301/1301 - 344s - loss: 0.0984 - accuracy: 0.9675 - val_loss: 0.0448 - val_accuracy: 0.9856 - lr: 0.0010 - 344s/epoch - 264ms/step\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(datagen.flow(X_train,y_train, batch_size=batch_size),\n",
    "                    epochs = epochs, validation_data = (X_test,y_test),\n",
    "                    verbose = 2, steps_per_epoch=X_train.shape[0] // batch_size,\n",
    "                    callbacks=[learning_rate_reduction])\n",
    "\n",
    "model.save(filepath = \"example.h5\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "2decd2b7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'accuracy')"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2000x300 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig,ax=plt.subplots(2,1)\n",
    "fig.set\n",
    "x=range(1,1+epochs)\n",
    "ax[0].plot(x,history.history['loss'],color='red')\n",
    "ax[0].plot(x,history.history['val_loss'],color='blue')\n",
    "\n",
    "ax[1].plot(x,history.history['accuracy'],color='red')\n",
    "ax[1].plot(x,history.history['val_accuracy'],color='blue')\n",
    "ax[0].legend(['trainng loss','validation loss'])\n",
    "ax[1].legend(['trainng acc','validation acc'])\n",
    "plt.xlabel('Number of epochs')\n",
    "plt.ylabel('accuracy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "94ddbd99",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "651/651 [==============================] - 24s 36ms/step\n"
     ]
    }
   ],
   "source": [
    "y_pre_test=model.predict(X_test)\n",
    "y_pre_test=np.argmax(y_pre_test,axis=1)\n",
    "y_test=np.argmax(y_test,axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "524ed932",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import confusion_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "fc36a179",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "      <th>11</th>\n",
       "      <th>12</th>\n",
       "      <th>13</th>\n",
       "      <th>14</th>\n",
       "      <th>15</th>\n",
       "      <th>16</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1191</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>1243</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>62</td>\n",
       "      <td>11</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1196</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>1275</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1208</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>1072</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>1224</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1233</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>1130</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>14</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "      <td>1173</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1257</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1229</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1223</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0</td>\n",
       "      <td>69</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1150</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1272</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1220</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1224</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      0     1     2     3     4     5     6     7     8     9     10    11  \\\n",
       "0   1191     0     2     0     0     0     0     0     0     0     0     0   \n",
       "1      0  1243     2     1     2     0     0     5     1     1     0     0   \n",
       "2      0     0  1196     1     0     0     0     4     1     0     0     0   \n",
       "3      0     0     2  1275     0     4     0     0     5     2     0     0   \n",
       "4      0     1     1     0  1208     0     0     1     0     6     0     0   \n",
       "5      2     0     0     2     2  1072     7     0     6     1     0     0   \n",
       "6      2     0     1     0     0     2  1224     0     2     1     0     0   \n",
       "7      0     1     8     8     2     0     0  1233     0     4     2     0   \n",
       "8      3     0     3     1     1     3     2     0  1130     4     0     0   \n",
       "9      3     1     0     0    14     0     0     4     2  1173     0     0   \n",
       "10     0     0     0     0     1     0     0     0     0     0  1257     0   \n",
       "11     0     0     0     0     0     0     0     0     0     0     0  1229   \n",
       "12     0     0     0     0     0     0     0     0     0     0     1     0   \n",
       "13     0    69     0     0     0     0     0     0     0     0     0     0   \n",
       "14     0     0     0     0     0     0     1     0     0     0     0     0   \n",
       "15     0     0     0     0     0     0     0     0     0     0     0     0   \n",
       "16     0     0     0     0     0     0     0     0     0     0     0     0   \n",
       "\n",
       "      12    13    14    15    16  \n",
       "0      0     0     0     0     3  \n",
       "1      1    62    11     4     2  \n",
       "2      2     0     0     0     0  \n",
       "3      0     0     0     0     0  \n",
       "4      0     0     0     0     0  \n",
       "5      0     0     0     0     0  \n",
       "6      0     0     0     0     0  \n",
       "7      0     0     0     6     0  \n",
       "8      1     0     0     0     0  \n",
       "9      0     0     0     0     0  \n",
       "10     0     0     0     0     0  \n",
       "11     0     0     0     0     0  \n",
       "12  1223     0     0     0     0  \n",
       "13     0  1150     0     0     0  \n",
       "14     0     0  1272     0     0  \n",
       "15     0     0     0  1220     0  \n",
       "16     0     0     0     0  1224  "
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "conf=confusion_matrix(y_test,y_pre_test)\n",
    "conf=pd.DataFrame(conf,index=range(0,17),columns=range(0,17))\n",
    "conf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "ffa7dd92",
   "metadata": {},
   "outputs": [],
   "source": [
    "x=(y_pre_test-y_test!=0).tolist()\n",
    "x=[i for i,l in enumerate(x) if l!=False]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "9f65ff17",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 5000x5000 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig,ax=plt.subplots(1,10,sharey=False,figsize=(50,50))\n",
    "\n",
    "for i in range(10):\n",
    "    ax[i].imshow(X_test[x[i]][:,:,0])\n",
    "    ax[i].set_xlabel('Real {}, Predicted {}'.format(y_test[x[i]],y_pre_test[x[i]]))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}