simple GAN(1D).ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# demonstrate simple x^2 function\n",
    "from matplotlib import pyplot\n",
    "\n",
    "# simple function\n",
    "def calculate(x):\n",
    "\treturn x * x*x\n",
    "\n",
    "# define inputs\n",
    "inputs = [-0.5, -0.4, -0.3, -0.2, -0.1, 0, 0.1, 0.2, 0.3, 0.4, 0.5]\n",
    "# calculate outputs\n",
    "outputs = [calculate(x) for x in inputs]\n",
    "# plot the result\n",
    "pyplot.plot(inputs, outputs)\n",
    "pyplot.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# example of generating random samples from X^2\n",
    "from numpy.random import rand\n",
    "from numpy import hstack\n",
    "from matplotlib import pyplot\n",
    "\n",
    "# generate randoms sample from x^2\n",
    "def generate_samples(n=100):\n",
    "\t# generate random inputs in [-0.5, 0.5]\n",
    "\tX1 = rand(n) - 0.5\n",
    "\t# generate outputs X^2 (quadratic)\n",
    "\tX2 = X1 * X1 *X1\n",
    "\t# stack arrays\n",
    "\tX1 = X1.reshape(n, 1)\n",
    "\tX2 = X2.reshape(n, 1)\n",
    "\treturn hstack((X1, X2))\n",
    "\n",
    "# generate samples\n",
    "data = generate_samples()\n",
    "# plot samples\n",
    "pyplot.scatter(data[:, 0], data[:, 1])\n",
    "pyplot.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 1.0 0.203125\n",
      "1 1.0 0.171875\n",
      "2 1.0 0.171875\n",
      "3 1.0 0.234375\n",
      "4 0.984375 0.265625\n",
      "5 1.0 0.28125\n",
      "6 0.875 0.515625\n",
      "7 0.578125 0.5\n",
      "8 0.65625 0.390625\n",
      "9 0.65625 0.4375\n",
      "10 0.5 0.484375\n",
      "11 0.421875 0.59375\n",
      "12 0.5 0.53125\n",
      "13 0.515625 0.578125\n",
      "14 0.390625 0.671875\n",
      "15 0.5625 0.5625\n",
      "16 0.53125 0.78125\n",
      "17 0.328125 0.703125\n",
      "18 0.421875 0.625\n",
      "19 0.296875 0.671875\n",
      "20 0.4375 0.71875\n",
      "21 0.359375 0.609375\n",
      "22 0.421875 0.625\n",
      "23 0.40625 0.75\n",
      "24 0.21875 0.734375\n",
      "25 0.34375 0.84375\n",
      "26 0.296875 0.734375\n",
      "27 0.40625 0.703125\n",
      "28 0.34375 0.65625\n",
      "29 0.359375 0.75\n",
      "30 0.3125 0.859375\n",
      "31 0.375 0.828125\n",
      "32 0.390625 0.84375\n",
      "33 0.328125 0.796875\n",
      "34 0.40625 0.890625\n",
      "35 0.375 0.890625\n",
      "36 0.3125 0.953125\n",
      "37 0.234375 0.984375\n",
      "38 0.234375 1.0\n",
      "39 0.171875 1.0\n",
      "40 0.265625 0.984375\n",
      "41 0.28125 1.0\n",
      "42 0.375 1.0\n",
      "43 0.3125 1.0\n",
      "44 0.21875 1.0\n",
      "45 0.28125 1.0\n",
      "46 0.234375 1.0\n",
      "47 0.28125 1.0\n",
      "48 0.234375 1.0\n",
      "49 0.28125 1.0\n",
      "50 0.328125 1.0\n",
      "51 0.171875 1.0\n",
      "52 0.296875 1.0\n",
      "53 0.296875 1.0\n",
      "54 0.375 1.0\n",
      "55 0.28125 1.0\n",
      "56 0.296875 0.984375\n",
      "57 0.421875 1.0\n",
      "58 0.40625 0.953125\n",
      "59 0.28125 1.0\n",
      "60 0.53125 0.984375\n",
      "61 0.359375 0.984375\n",
      "62 0.421875 1.0\n",
      "63 0.484375 1.0\n",
      "64 0.515625 1.0\n",
      "65 0.46875 0.984375\n",
      "66 0.515625 1.0\n",
      "67 0.515625 1.0\n",
      "68 0.46875 0.984375\n",
      "69 0.4375 1.0\n",
      "70 0.453125 1.0\n",
      "71 0.546875 1.0\n",
      "72 0.515625 0.984375\n",
      "73 0.515625 1.0\n",
      "74 0.5 1.0\n",
      "75 0.5 1.0\n",
      "76 0.546875 0.984375\n",
      "77 0.546875 0.984375\n",
      "78 0.609375 0.984375\n",
      "79 0.59375 1.0\n",
      "80 0.625 0.96875\n",
      "81 0.53125 0.984375\n",
      "82 0.625 0.96875\n",
      "83 0.6875 0.984375\n",
      "84 0.609375 1.0\n",
      "85 0.640625 0.984375\n",
      "86 0.640625 0.984375\n",
      "87 0.578125 1.0\n",
      "88 0.703125 1.0\n",
      "89 0.671875 0.96875\n",
      "90 0.671875 0.96875\n",
      "91 0.59375 0.984375\n",
      "92 0.671875 0.984375\n",
      "93 0.65625 1.0\n",
      "94 0.796875 0.984375\n",
      "95 0.671875 1.0\n",
      "96 0.8125 0.96875\n",
      "97 0.828125 0.984375\n",
      "98 0.8125 1.0\n",
      "99 0.703125 0.984375\n",
      "100 0.6875 0.9375\n",
      "101 0.875 0.984375\n",
      "102 0.890625 0.984375\n",
      "103 0.828125 0.953125\n",
      "104 0.890625 0.984375\n",
      "105 0.859375 1.0\n",
      "106 0.84375 0.96875\n",
      "107 0.890625 0.9375\n",
      "108 0.859375 1.0\n",
      "109 0.875 1.0\n",
      "110 0.90625 0.984375\n",
      "111 0.890625 0.96875\n",
      "112 0.859375 0.96875\n",
      "113 0.90625 0.984375\n",
      "114 0.90625 0.9375\n",
      "115 0.921875 0.984375\n",
      "116 0.90625 1.0\n",
      "117 0.9375 0.953125\n",
      "118 0.890625 0.984375\n",
      "119 0.875 0.984375\n",
      "120 0.96875 0.984375\n",
      "121 0.9375 1.0\n",
      "122 0.875 0.984375\n",
      "123 0.90625 0.921875\n",
      "124 1.0 0.984375\n",
      "125 0.953125 0.953125\n",
      "126 0.9375 1.0\n",
      "127 0.96875 0.96875\n",
      "128 0.984375 0.9375\n",
      "129 1.0 0.96875\n",
      "130 0.984375 0.96875\n",
      "131 1.0 0.953125\n",
      "132 1.0 0.921875\n",
      "133 0.953125 0.890625\n",
      "134 0.984375 0.9375\n",
      "135 1.0 0.953125\n",
      "136 1.0 1.0\n",
      "137 1.0 0.984375\n",
      "138 1.0 0.96875\n",
      "139 1.0 0.96875\n",
      "140 1.0 0.9375\n",
      "141 1.0 0.9375\n",
      "142 1.0 0.9375\n",
      "143 1.0 0.921875\n",
      "144 1.0 0.953125\n",
      "145 1.0 0.9375\n",
      "146 1.0 0.96875\n",
      "147 1.0 0.953125\n",
      "148 1.0 0.921875\n",
      "149 1.0 0.921875\n",
      "150 1.0 0.96875\n",
      "151 1.0 0.96875\n",
      "152 1.0 0.953125\n",
      "153 1.0 0.84375\n",
      "154 1.0 0.90625\n",
      "155 1.0 0.90625\n",
      "156 1.0 0.90625\n",
      "157 1.0 0.96875\n",
      "158 1.0 0.90625\n",
      "159 1.0 0.9375\n",
      "160 1.0 0.96875\n",
      "161 1.0 0.90625\n",
      "162 1.0 0.875\n",
      "163 1.0 0.953125\n",
      "164 1.0 0.9375\n",
      "165 1.0 0.875\n",
      "166 1.0 0.953125\n",
      "167 1.0 0.875\n",
      "168 1.0 0.90625\n",
      "169 1.0 0.828125\n",
      "170 1.0 0.890625\n",
      "171 1.0 0.84375\n",
      "172 1.0 0.875\n",
      "173 1.0 0.9375\n",
      "174 1.0 0.875\n",
      "175 1.0 0.890625\n",
      "176 1.0 0.90625\n",
      "177 1.0 0.9375\n",
      "178 1.0 0.921875\n",
      "179 1.0 0.921875\n",
      "180 1.0 0.828125\n",
      "181 1.0 0.921875\n",
      "182 1.0 0.9375\n",
      "183 1.0 0.90625\n",
      "184 1.0 0.9375\n",
      "185 1.0 0.953125\n",
      "186 1.0 0.8125\n",
      "187 1.0 0.890625\n",
      "188 1.0 0.859375\n",
      "189 1.0 0.890625\n",
      "190 1.0 0.890625\n",
      "191 1.0 0.875\n",
      "192 1.0 0.875\n",
      "193 1.0 0.9375\n",
      "194 1.0 0.921875\n",
      "195 1.0 0.921875\n",
      "196 1.0 0.9375\n",
      "197 1.0 0.875\n",
      "198 1.0 0.859375\n",
      "199 1.0 0.875\n",
      "200 1.0 0.9375\n",
      "201 1.0 0.828125\n",
      "202 1.0 0.953125\n",
      "203 1.0 0.796875\n",
      "204 1.0 0.953125\n",
      "205 1.0 0.828125\n",
      "206 1.0 0.953125\n",
      "207 1.0 0.859375\n",
      "208 1.0 0.796875\n",
      "209 1.0 0.84375\n",
      "210 1.0 0.953125\n",
      "211 1.0 0.90625\n",
      "212 1.0 0.953125\n",
      "213 1.0 0.921875\n",
      "214 1.0 0.875\n",
      "215 1.0 0.859375\n",
      "216 1.0 0.859375\n",
      "217 1.0 0.859375\n",
      "218 1.0 0.90625\n",
      "219 1.0 0.921875\n",
      "220 1.0 0.875\n",
      "221 1.0 0.9375\n",
      "222 1.0 0.875\n",
      "223 1.0 0.9375\n",
      "224 1.0 0.90625\n",
      "225 1.0 0.890625\n",
      "226 1.0 0.921875\n",
      "227 1.0 0.890625\n",
      "228 1.0 0.90625\n",
      "229 1.0 0.890625\n",
      "230 1.0 0.90625\n",
      "231 1.0 0.875\n",
      "232 1.0 0.875\n",
      "233 1.0 0.953125\n",
      "234 1.0 0.859375\n",
      "235 1.0 0.9375\n",
      "236 1.0 0.890625\n",
      "237 1.0 0.859375\n",
      "238 1.0 0.890625\n",
      "239 1.0 0.875\n",
      "240 1.0 0.9375\n",
      "241 1.0 0.828125\n",
      "242 1.0 0.828125\n",
      "243 1.0 0.890625\n",
      "244 1.0 0.828125\n",
      "245 1.0 0.84375\n",
      "246 1.0 0.859375\n",
      "247 1.0 0.875\n",
      "248 1.0 0.875\n",
      "249 1.0 0.875\n",
      "250 1.0 0.90625\n",
      "251 1.0 0.859375\n",
      "252 1.0 0.96875\n",
      "253 1.0 0.96875\n",
      "254 1.0 0.921875\n",
      "255 1.0 0.859375\n",
      "256 1.0 0.890625\n",
      "257 1.0 0.921875\n",
      "258 1.0 0.921875\n",
      "259 1.0 0.875\n",
      "260 1.0 0.90625\n",
      "261 1.0 0.90625\n",
      "262 1.0 0.765625\n",
      "263 1.0 0.953125\n",
      "264 1.0 0.875\n",
      "265 1.0 0.984375\n",
      "266 1.0 0.84375\n",
      "267 1.0 0.828125\n",
      "268 1.0 0.890625\n",
      "269 1.0 0.9375\n",
      "270 1.0 0.921875\n",
      "271 1.0 0.859375\n",
      "272 1.0 0.921875\n",
      "273 1.0 0.9375\n",
      "274 1.0 0.890625\n",
      "275 1.0 0.953125\n",
      "276 1.0 0.859375\n",
      "277 1.0 0.921875\n",
      "278 1.0 0.875\n",
      "279 1.0 0.828125\n",
      "280 1.0 0.90625\n",
      "281 1.0 0.890625\n",
      "282 1.0 0.890625\n",
      "283 1.0 0.90625\n",
      "284 1.0 0.890625\n",
      "285 1.0 0.890625\n",
      "286 1.0 0.875\n",
      "287 1.0 0.859375\n",
      "288 1.0 0.9375\n",
      "289 1.0 0.875\n",
      "290 1.0 0.859375\n",
      "291 1.0 0.890625\n",
      "292 1.0 0.875\n",
      "293 1.0 0.875\n",
      "294 1.0 0.875\n",
      "295 1.0 0.890625\n",
      "296 1.0 0.90625\n",
      "297 1.0 0.8125\n",
      "298 1.0 0.859375\n",
      "299 1.0 0.953125\n",
      "300 1.0 0.84375\n",
      "301 1.0 0.953125\n",
      "302 1.0 0.84375\n",
      "303 1.0 0.890625\n",
      "304 1.0 0.953125\n",
      "305 1.0 0.984375\n",
      "306 1.0 0.953125\n",
      "307 1.0 0.921875\n",
      "308 1.0 0.9375\n",
      "309 1.0 0.90625\n",
      "310 1.0 0.9375\n",
      "311 1.0 0.921875\n",
      "312 1.0 0.84375\n",
      "313 1.0 0.921875\n",
      "314 1.0 0.796875\n",
      "315 1.0 0.84375\n",
      "316 1.0 0.96875\n",
      "317 1.0 0.96875\n",
      "318 1.0 0.875\n",
      "319 1.0 0.859375\n",
      "320 1.0 0.921875\n",
      "321 1.0 0.90625\n",
      "322 1.0 0.9375\n",
      "323 1.0 0.90625\n",
      "324 1.0 0.90625\n",
      "325 1.0 0.828125\n",
      "326 1.0 0.875\n",
      "327 1.0 0.90625\n",
      "328 1.0 0.9375\n",
      "329 1.0 0.84375\n",
      "330 1.0 0.859375\n",
      "331 1.0 0.84375\n",
      "332 1.0 0.8125\n",
      "333 1.0 0.953125\n",
      "334 1.0 0.890625\n",
      "335 1.0 0.90625\n",
      "336 1.0 0.90625\n",
      "337 1.0 0.890625\n",
      "338 1.0 0.953125\n",
      "339 1.0 0.890625\n",
      "340 1.0 0.96875\n",
      "341 1.0 0.9375\n",
      "342 1.0 0.90625\n",
      "343 1.0 0.921875\n",
      "344 1.0 0.875\n",
      "345 1.0 0.9375\n",
      "346 1.0 0.890625\n",
      "347 1.0 0.890625\n",
      "348 1.0 0.9375\n",
      "349 1.0 0.90625\n",
      "350 1.0 0.859375\n",
      "351 1.0 0.953125\n",
      "352 1.0 0.9375\n",
      "353 1.0 0.9375\n",
      "354 1.0 0.90625\n",
      "355 1.0 0.90625\n",
      "356 1.0 0.90625\n",
      "357 1.0 0.859375\n",
      "358 1.0 0.890625\n",
      "359 1.0 0.890625\n",
      "360 1.0 0.875\n",
      "361 1.0 0.890625\n",
      "362 1.0 0.953125\n",
      "363 1.0 0.890625\n",
      "364 1.0 0.890625\n",
      "365 1.0 0.84375\n",
      "366 1.0 0.921875\n",
      "367 1.0 0.9375\n",
      "368 1.0 0.9375\n",
      "369 1.0 0.953125\n",
      "370 1.0 0.953125\n",
      "371 1.0 0.90625\n",
      "372 1.0 0.90625\n",
      "373 1.0 0.953125\n",
      "374 1.0 0.90625\n",
      "375 1.0 0.9375\n",
      "376 1.0 0.921875\n",
      "377 1.0 0.875\n",
      "378 1.0 0.96875\n",
      "379 1.0 0.90625\n",
      "380 1.0 0.875\n",
      "381 1.0 0.921875\n",
      "382 1.0 0.953125\n",
      "383 1.0 0.90625\n",
      "384 1.0 0.90625\n",
      "385 1.0 0.90625\n",
      "386 1.0 0.859375\n",
      "387 1.0 0.859375\n",
      "388 1.0 0.875\n",
      "389 1.0 0.890625\n",
      "390 1.0 0.953125\n",
      "391 1.0 0.96875\n",
      "392 1.0 0.9375\n",
      "393 1.0 0.953125\n",
      "394 1.0 0.890625\n",
      "395 1.0 0.875\n",
      "396 1.0 0.859375\n",
      "397 1.0 0.890625\n",
      "398 1.0 0.890625\n",
      "399 1.0 0.90625\n",
      "400 1.0 0.84375\n",
      "401 1.0 0.921875\n",
      "402 1.0 0.90625\n",
      "403 1.0 0.921875\n",
      "404 1.0 0.875\n",
      "405 1.0 0.875\n",
      "406 1.0 0.90625\n",
      "407 1.0 0.90625\n",
      "408 1.0 0.890625\n",
      "409 1.0 0.90625\n",
      "410 1.0 0.9375\n",
      "411 1.0 0.9375\n",
      "412 1.0 0.890625\n",
      "413 1.0 0.953125\n",
      "414 1.0 0.953125\n",
      "415 1.0 0.84375\n",
      "416 1.0 0.859375\n",
      "417 1.0 0.921875\n",
      "418 1.0 0.90625\n",
      "419 1.0 0.90625\n",
      "420 1.0 0.96875\n",
      "421 1.0 0.921875\n",
      "422 1.0 0.9375\n",
      "423 1.0 0.9375\n",
      "424 1.0 0.890625\n",
      "425 1.0 0.84375\n",
      "426 1.0 0.90625\n",
      "427 1.0 0.875\n",
      "428 1.0 0.9375\n",
      "429 1.0 0.890625\n",
      "430 1.0 0.90625\n",
      "431 1.0 0.875\n",
      "432 1.0 0.90625\n",
      "433 1.0 0.890625\n",
      "434 1.0 0.9375\n",
      "435 1.0 0.875\n",
      "436 1.0 0.921875\n",
      "437 1.0 0.953125\n",
      "438 1.0 0.9375\n",
      "439 1.0 0.90625\n",
      "440 1.0 0.90625\n",
      "441 1.0 0.96875\n",
      "442 1.0 0.90625\n",
      "443 1.0 0.890625\n",
      "444 1.0 0.953125\n",
      "445 1.0 0.890625\n",
      "446 1.0 0.90625\n",
      "447 1.0 0.890625\n",
      "448 1.0 0.9375\n",
      "449 1.0 0.890625\n",
      "450 1.0 0.90625\n",
      "451 1.0 0.84375\n",
      "452 1.0 0.890625\n",
      "453 1.0 0.875\n",
      "454 1.0 0.859375\n",
      "455 1.0 0.953125\n",
      "456 1.0 0.890625\n",
      "457 1.0 0.890625\n",
      "458 1.0 0.953125\n",
      "459 1.0 0.953125\n",
      "460 1.0 0.859375\n",
      "461 1.0 0.953125\n",
      "462 1.0 0.90625\n",
      "463 1.0 0.96875\n",
      "464 1.0 0.875\n",
      "465 1.0 0.9375\n",
      "466 1.0 0.9375\n",
      "467 1.0 0.890625\n",
      "468 1.0 0.90625\n",
      "469 1.0 0.890625\n",
      "470 1.0 0.890625\n",
      "471 1.0 0.953125\n",
      "472 1.0 0.9375\n",
      "473 1.0 0.890625\n",
      "474 1.0 0.90625\n",
      "475 1.0 0.9375\n",
      "476 1.0 0.96875\n",
      "477 1.0 0.875\n",
      "478 1.0 0.9375\n",
      "479 1.0 0.953125\n",
      "480 1.0 0.859375\n",
      "481 1.0 0.96875\n",
      "482 1.0 0.96875\n",
      "483 1.0 0.9375\n",
      "484 1.0 0.9375\n",
      "485 1.0 0.96875\n",
      "486 1.0 0.859375\n",
      "487 1.0 0.921875\n",
      "488 1.0 0.921875\n",
      "489 1.0 0.953125\n",
      "490 1.0 0.9375\n",
      "491 1.0 0.921875\n",
      "492 1.0 0.9375\n",
      "493 1.0 0.875\n",
      "494 1.0 0.953125\n",
      "495 1.0 0.953125\n",
      "496 1.0 0.921875\n",
      "497 1.0 0.890625\n",
      "498 1.0 0.890625\n",
      "499 1.0 0.90625\n",
      "500 1.0 0.953125\n",
      "501 1.0 0.9375\n",
      "502 1.0 0.921875\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "503 1.0 0.953125\n",
      "504 1.0 0.9375\n",
      "505 1.0 0.9375\n",
      "506 1.0 0.921875\n",
      "507 1.0 0.921875\n",
      "508 1.0 0.9375\n",
      "509 1.0 0.9375\n",
      "510 1.0 0.953125\n",
      "511 1.0 0.875\n",
      "512 1.0 0.921875\n",
      "513 1.0 0.921875\n",
      "514 1.0 0.890625\n",
      "515 1.0 0.890625\n",
      "516 1.0 0.921875\n",
      "517 1.0 0.9375\n",
      "518 1.0 0.921875\n",
      "519 1.0 0.875\n",
      "520 1.0 0.921875\n",
      "521 1.0 0.953125\n",
      "522 1.0 0.9375\n",
      "523 1.0 0.90625\n",
      "524 1.0 0.90625\n",
      "525 1.0 0.90625\n",
      "526 1.0 0.953125\n",
      "527 1.0 0.890625\n",
      "528 1.0 0.953125\n",
      "529 1.0 0.859375\n",
      "530 1.0 0.921875\n",
      "531 1.0 0.921875\n",
      "532 1.0 0.953125\n",
      "533 1.0 0.96875\n",
      "534 1.0 0.9375\n",
      "535 1.0 0.890625\n",
      "536 1.0 0.859375\n",
      "537 1.0 0.921875\n",
      "538 1.0 0.921875\n",
      "539 1.0 0.9375\n",
      "540 1.0 0.96875\n",
      "541 1.0 0.953125\n",
      "542 1.0 0.96875\n",
      "543 1.0 0.9375\n",
      "544 1.0 0.859375\n",
      "545 1.0 0.96875\n",
      "546 1.0 0.875\n",
      "547 1.0 0.9375\n",
      "548 1.0 0.953125\n",
      "549 1.0 0.9375\n",
      "550 1.0 0.90625\n",
      "551 1.0 0.875\n",
      "552 1.0 0.96875\n",
      "553 1.0 0.921875\n",
      "554 1.0 0.875\n",
      "555 1.0 0.90625\n",
      "556 1.0 0.9375\n",
      "557 1.0 0.9375\n",
      "558 1.0 0.9375\n",
      "559 1.0 0.921875\n",
      "560 1.0 0.953125\n",
      "561 1.0 0.875\n",
      "562 1.0 0.859375\n",
      "563 1.0 0.96875\n",
      "564 1.0 0.984375\n",
      "565 1.0 0.9375\n",
      "566 1.0 0.9375\n",
      "567 1.0 0.96875\n",
      "568 1.0 0.921875\n",
      "569 1.0 0.875\n",
      "570 1.0 0.953125\n",
      "571 1.0 1.0\n",
      "572 1.0 0.953125\n",
      "573 1.0 0.96875\n",
      "574 1.0 0.921875\n",
      "575 1.0 0.9375\n",
      "576 1.0 0.921875\n",
      "577 1.0 0.9375\n",
      "578 1.0 0.890625\n",
      "579 1.0 0.953125\n",
      "580 1.0 0.9375\n",
      "581 1.0 0.859375\n",
      "582 1.0 0.90625\n",
      "583 1.0 0.984375\n",
      "584 1.0 0.9375\n",
      "585 1.0 0.9375\n",
      "586 1.0 0.921875\n",
      "587 1.0 0.890625\n",
      "588 1.0 0.90625\n",
      "589 1.0 0.9375\n",
      "590 1.0 0.921875\n",
      "591 1.0 0.96875\n",
      "592 1.0 0.90625\n",
      "593 1.0 0.921875\n",
      "594 1.0 0.90625\n",
      "595 1.0 0.953125\n",
      "596 1.0 0.875\n",
      "597 1.0 0.921875\n",
      "598 1.0 0.921875\n",
      "599 1.0 0.96875\n",
      "600 1.0 0.984375\n",
      "601 1.0 0.875\n",
      "602 1.0 0.90625\n",
      "603 1.0 0.9375\n",
      "604 1.0 0.9375\n",
      "605 1.0 0.875\n",
      "606 1.0 0.96875\n",
      "607 1.0 0.953125\n",
      "608 1.0 0.921875\n",
      "609 1.0 0.953125\n",
      "610 1.0 0.9375\n",
      "611 1.0 0.984375\n",
      "612 1.0 0.984375\n",
      "613 1.0 0.921875\n",
      "614 1.0 0.890625\n",
      "615 1.0 0.90625\n",
      "616 1.0 0.90625\n",
      "617 1.0 0.859375\n",
      "618 1.0 0.921875\n",
      "619 1.0 0.9375\n",
      "620 1.0 0.953125\n",
      "621 1.0 0.953125\n",
      "622 1.0 0.953125\n",
      "623 1.0 0.953125\n",
      "624 1.0 0.96875\n",
      "625 1.0 0.859375\n",
      "626 1.0 0.953125\n",
      "627 1.0 0.90625\n",
      "628 1.0 0.921875\n",
      "629 1.0 0.953125\n",
      "630 1.0 0.890625\n",
      "631 1.0 0.953125\n",
      "632 1.0 0.90625\n",
      "633 1.0 0.9375\n",
      "634 1.0 0.96875\n",
      "635 1.0 0.90625\n",
      "636 1.0 0.96875\n",
      "637 1.0 0.984375\n",
      "638 1.0 0.9375\n",
      "639 1.0 0.90625\n",
      "640 1.0 0.921875\n",
      "641 1.0 0.90625\n",
      "642 1.0 0.9375\n",
      "643 1.0 0.859375\n",
      "644 1.0 0.921875\n",
      "645 1.0 0.859375\n",
      "646 1.0 0.921875\n",
      "647 1.0 0.9375\n",
      "648 1.0 0.9375\n",
      "649 1.0 0.9375\n",
      "650 1.0 0.921875\n",
      "651 1.0 0.9375\n",
      "652 1.0 0.96875\n",
      "653 1.0 0.953125\n",
      "654 1.0 0.9375\n",
      "655 1.0 0.953125\n",
      "656 1.0 0.921875\n",
      "657 1.0 0.953125\n",
      "658 1.0 0.96875\n",
      "659 1.0 0.96875\n",
      "660 1.0 0.875\n",
      "661 1.0 0.90625\n",
      "662 1.0 0.828125\n",
      "663 1.0 0.984375\n",
      "664 1.0 0.921875\n",
      "665 1.0 0.921875\n",
      "666 1.0 0.953125\n",
      "667 1.0 0.921875\n",
      "668 1.0 0.890625\n",
      "669 1.0 0.9375\n",
      "670 1.0 0.96875\n",
      "671 1.0 0.90625\n",
      "672 1.0 0.921875\n",
      "673 1.0 0.921875\n",
      "674 1.0 0.90625\n",
      "675 1.0 0.96875\n",
      "676 1.0 0.875\n",
      "677 1.0 0.953125\n",
      "678 1.0 0.90625\n",
      "679 1.0 0.9375\n",
      "680 1.0 0.953125\n",
      "681 1.0 0.96875\n",
      "682 1.0 0.9375\n",
      "683 1.0 0.9375\n",
      "684 1.0 0.921875\n",
      "685 1.0 0.96875\n",
      "686 1.0 0.890625\n",
      "687 1.0 0.96875\n",
      "688 1.0 0.90625\n",
      "689 1.0 0.90625\n",
      "690 1.0 0.9375\n",
      "691 1.0 0.96875\n",
      "692 1.0 0.953125\n",
      "693 1.0 0.953125\n",
      "694 1.0 0.921875\n",
      "695 1.0 0.953125\n",
      "696 1.0 0.96875\n",
      "697 1.0 0.90625\n",
      "698 1.0 0.90625\n",
      "699 1.0 0.90625\n",
      "700 1.0 0.96875\n",
      "701 1.0 0.96875\n",
      "702 1.0 0.921875\n",
      "703 1.0 0.9375\n",
      "704 1.0 0.921875\n",
      "705 1.0 0.96875\n",
      "706 1.0 0.921875\n",
      "707 1.0 0.953125\n",
      "708 1.0 0.953125\n",
      "709 1.0 0.921875\n",
      "710 1.0 0.921875\n",
      "711 1.0 0.953125\n",
      "712 1.0 0.921875\n",
      "713 1.0 0.984375\n",
      "714 1.0 0.9375\n",
      "715 1.0 0.953125\n",
      "716 1.0 0.953125\n",
      "717 1.0 0.921875\n",
      "718 1.0 0.9375\n",
      "719 1.0 0.953125\n",
      "720 1.0 0.90625\n",
      "721 1.0 0.953125\n",
      "722 1.0 0.9375\n",
      "723 1.0 0.921875\n",
      "724 1.0 0.9375\n",
      "725 1.0 0.984375\n",
      "726 1.0 0.96875\n",
      "727 1.0 0.984375\n",
      "728 1.0 0.9375\n",
      "729 1.0 0.890625\n",
      "730 1.0 0.90625\n",
      "731 1.0 0.96875\n",
      "732 1.0 1.0\n",
      "733 1.0 0.953125\n",
      "734 1.0 0.890625\n",
      "735 1.0 0.984375\n",
      "736 1.0 0.953125\n",
      "737 1.0 0.90625\n",
      "738 1.0 0.9375\n",
      "739 1.0 0.984375\n",
      "740 1.0 0.921875\n",
      "741 1.0 0.921875\n",
      "742 1.0 0.9375\n",
      "743 1.0 0.90625\n",
      "744 1.0 0.921875\n",
      "745 1.0 0.921875\n",
      "746 1.0 0.953125\n",
      "747 1.0 0.90625\n",
      "748 1.0 0.96875\n",
      "749 1.0 0.96875\n",
      "750 1.0 0.921875\n",
      "751 1.0 0.921875\n",
      "752 1.0 0.890625\n",
      "753 1.0 0.9375\n",
      "754 1.0 0.953125\n",
      "755 1.0 0.96875\n",
      "756 1.0 0.984375\n",
      "757 1.0 0.9375\n",
      "758 1.0 0.90625\n",
      "759 1.0 0.96875\n",
      "760 1.0 0.96875\n",
      "761 1.0 0.96875\n",
      "762 1.0 0.890625\n",
      "763 1.0 0.984375\n",
      "764 1.0 0.96875\n",
      "765 1.0 0.921875\n",
      "766 1.0 0.921875\n",
      "767 1.0 0.96875\n",
      "768 1.0 0.90625\n",
      "769 1.0 0.921875\n",
      "770 1.0 0.96875\n",
      "771 1.0 0.9375\n",
      "772 1.0 0.96875\n",
      "773 1.0 0.90625\n",
      "774 1.0 0.921875\n",
      "775 1.0 0.9375\n",
      "776 1.0 0.875\n",
      "777 1.0 0.90625\n",
      "778 1.0 0.9375\n",
      "779 1.0 0.953125\n",
      "780 1.0 0.875\n",
      "781 1.0 0.953125\n",
      "782 1.0 0.984375\n",
      "783 1.0 0.875\n",
      "784 1.0 0.9375\n",
      "785 1.0 0.9375\n",
      "786 1.0 0.984375\n",
      "787 1.0 0.953125\n",
      "788 1.0 0.9375\n",
      "789 1.0 0.953125\n",
      "790 1.0 0.96875\n",
      "791 1.0 0.96875\n",
      "792 1.0 0.984375\n",
      "793 1.0 0.90625\n",
      "794 1.0 0.9375\n",
      "795 1.0 0.921875\n",
      "796 1.0 0.921875\n",
      "797 1.0 0.953125\n",
      "798 1.0 0.921875\n",
      "799 1.0 0.984375\n",
      "800 1.0 0.984375\n",
      "801 1.0 0.921875\n",
      "802 1.0 0.953125\n",
      "803 1.0 0.953125\n",
      "804 1.0 0.96875\n",
      "805 1.0 0.90625\n",
      "806 1.0 0.90625\n",
      "807 1.0 0.921875\n",
      "808 1.0 0.9375\n",
      "809 1.0 0.921875\n",
      "810 1.0 0.96875\n",
      "811 1.0 0.953125\n",
      "812 1.0 0.9375\n",
      "813 1.0 0.9375\n",
      "814 1.0 0.90625\n",
      "815 1.0 0.9375\n",
      "816 1.0 0.90625\n",
      "817 1.0 0.9375\n",
      "818 1.0 0.96875\n",
      "819 1.0 0.921875\n",
      "820 1.0 0.984375\n",
      "821 1.0 0.9375\n",
      "822 1.0 0.9375\n",
      "823 1.0 0.90625\n",
      "824 1.0 0.984375\n",
      "825 1.0 0.984375\n",
      "826 1.0 0.90625\n",
      "827 1.0 0.96875\n",
      "828 1.0 0.921875\n",
      "829 1.0 0.984375\n",
      "830 1.0 0.9375\n",
      "831 1.0 0.9375\n",
      "832 1.0 0.9375\n",
      "833 1.0 0.90625\n",
      "834 1.0 0.953125\n",
      "835 1.0 0.921875\n",
      "836 1.0 0.890625\n",
      "837 1.0 0.921875\n",
      "838 1.0 0.9375\n",
      "839 1.0 0.953125\n",
      "840 1.0 0.9375\n",
      "841 1.0 0.984375\n",
      "842 1.0 0.96875\n",
      "843 1.0 0.921875\n",
      "844 1.0 0.9375\n",
      "845 1.0 0.953125\n",
      "846 1.0 0.953125\n",
      "847 1.0 0.9375\n",
      "848 1.0 0.921875\n",
      "849 1.0 0.9375\n",
      "850 1.0 0.953125\n",
      "851 1.0 1.0\n",
      "852 1.0 0.953125\n",
      "853 1.0 0.9375\n",
      "854 1.0 0.984375\n",
      "855 1.0 0.9375\n",
      "856 1.0 0.921875\n",
      "857 1.0 0.96875\n",
      "858 1.0 0.921875\n",
      "859 1.0 0.953125\n",
      "860 1.0 0.9375\n",
      "861 1.0 0.9375\n",
      "862 1.0 0.90625\n",
      "863 1.0 0.921875\n",
      "864 1.0 0.984375\n",
      "865 1.0 0.90625\n",
      "866 1.0 0.96875\n",
      "867 1.0 0.953125\n",
      "868 1.0 0.890625\n",
      "869 1.0 0.96875\n",
      "870 1.0 0.96875\n",
      "871 1.0 0.9375\n",
      "872 1.0 0.90625\n",
      "873 1.0 0.953125\n",
      "874 1.0 0.9375\n",
      "875 1.0 0.984375\n",
      "876 1.0 0.953125\n",
      "877 1.0 0.984375\n",
      "878 1.0 0.921875\n",
      "879 1.0 0.984375\n",
      "880 1.0 0.984375\n",
      "881 1.0 0.96875\n",
      "882 1.0 0.953125\n",
      "883 1.0 0.90625\n",
      "884 1.0 0.9375\n",
      "885 1.0 0.953125\n",
      "886 1.0 0.953125\n",
      "887 1.0 0.953125\n",
      "888 1.0 0.984375\n",
      "889 1.0 0.921875\n",
      "890 1.0 0.921875\n",
      "891 1.0 0.984375\n",
      "892 1.0 0.921875\n",
      "893 1.0 0.921875\n",
      "894 1.0 0.96875\n",
      "895 1.0 0.953125\n",
      "896 1.0 0.890625\n",
      "897 1.0 0.9375\n",
      "898 1.0 0.890625\n",
      "899 1.0 0.953125\n",
      "900 1.0 0.96875\n",
      "901 1.0 0.9375\n",
      "902 1.0 0.921875\n",
      "903 1.0 0.921875\n",
      "904 1.0 0.96875\n",
      "905 1.0 0.90625\n",
      "906 1.0 0.953125\n",
      "907 1.0 0.9375\n",
      "908 1.0 0.9375\n",
      "909 1.0 0.9375\n",
      "910 1.0 0.9375\n",
      "911 1.0 0.953125\n",
      "912 1.0 0.890625\n",
      "913 1.0 0.96875\n",
      "914 1.0 0.9375\n",
      "915 1.0 0.953125\n",
      "916 1.0 0.984375\n",
      "917 1.0 0.9375\n",
      "918 1.0 0.953125\n",
      "919 1.0 0.96875\n",
      "920 1.0 0.96875\n",
      "921 1.0 0.9375\n",
      "922 1.0 1.0\n",
      "923 1.0 0.9375\n",
      "924 1.0 0.9375\n",
      "925 1.0 0.9375\n",
      "926 1.0 0.9375\n",
      "927 1.0 0.921875\n",
      "928 1.0 0.9375\n",
      "929 1.0 0.96875\n",
      "930 1.0 0.96875\n",
      "931 1.0 0.96875\n",
      "932 1.0 0.9375\n",
      "933 1.0 0.953125\n",
      "934 1.0 0.953125\n",
      "935 1.0 0.96875\n",
      "936 1.0 0.875\n",
      "937 1.0 0.90625\n",
      "938 1.0 0.96875\n",
      "939 1.0 0.96875\n",
      "940 1.0 0.953125\n",
      "941 1.0 0.96875\n",
      "942 1.0 0.890625\n",
      "943 1.0 0.9375\n",
      "944 1.0 0.953125\n",
      "945 1.0 0.953125\n",
      "946 1.0 0.90625\n",
      "947 1.0 0.921875\n",
      "948 1.0 0.9375\n",
      "949 1.0 0.984375\n",
      "950 1.0 0.921875\n",
      "951 1.0 0.96875\n",
      "952 1.0 0.9375\n",
      "953 1.0 0.953125\n",
      "954 1.0 0.953125\n",
      "955 1.0 0.9375\n",
      "956 1.0 0.875\n",
      "957 1.0 0.9375\n",
      "958 1.0 0.921875\n",
      "959 1.0 0.96875\n",
      "960 1.0 0.96875\n",
      "961 1.0 1.0\n",
      "962 1.0 1.0\n",
      "963 1.0 1.0\n",
      "964 1.0 0.90625\n",
      "965 1.0 0.9375\n",
      "966 1.0 0.953125\n",
      "967 1.0 0.9375\n",
      "968 1.0 0.984375\n",
      "969 1.0 0.921875\n",
      "970 1.0 0.96875\n",
      "971 1.0 0.96875\n",
      "972 1.0 0.90625\n",
      "973 1.0 0.96875\n",
      "974 1.0 0.953125\n",
      "975 1.0 0.953125\n",
      "976 1.0 0.9375\n",
      "977 1.0 0.921875\n",
      "978 1.0 0.90625\n",
      "979 1.0 0.96875\n",
      "980 1.0 0.9375\n",
      "981 1.0 0.9375\n",
      "982 1.0 0.9375\n",
      "983 1.0 0.9375\n",
      "984 1.0 0.953125\n",
      "985 1.0 0.921875\n",
      "986 1.0 0.953125\n",
      "987 1.0 0.9375\n",
      "988 1.0 0.984375\n",
      "989 1.0 0.953125\n",
      "990 1.0 0.96875\n",
      "991 1.0 0.984375\n",
      "992 1.0 0.921875\n",
      "993 1.0 0.96875\n",
      "994 1.0 0.953125\n",
      "995 1.0 0.96875\n",
      "996 1.0 0.9375\n",
      "997 1.0 0.890625\n",
      "998 1.0 0.984375\n",
      "999 1.0 0.96875\n"
     ]
    }
   ],
   "source": [
    "# define and fit a discriminator model\n",
    "from numpy import zeros\n",
    "from numpy import ones\n",
    "from numpy import hstack\n",
    "from numpy.random import rand\n",
    "from numpy.random import randn\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "\n",
    "# define the standalone discriminator model\n",
    "def define_discriminator(n_inputs=2):\n",
    "    model = Sequential()\n",
    "    model.add(Dense(25, activation='relu', kernel_initializer='he_uniform', input_dim=n_inputs))\n",
    "    model.add(Dense(30, activation='relu'))\n",
    "    model.add(Dense(1, activation='sigmoid'))\n",
    "    # compile model\n",
    "    model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "    return model\n",
    "\n",
    "# generate n real samples with class labels\n",
    "def generate_real_samples(n):\n",
    "\t# generate inputs in [-0.5, 0.5]\n",
    "\tX1 = rand(n) - 0.5\n",
    "\t# generate outputs X^2\n",
    "\tX2 = X1 * X1 * X1\n",
    "\t# stack arrays\n",
    "\tX1 = X1.reshape(n, 1)\n",
    "\tX2 = X2.reshape(n, 1)\n",
    "\tX = hstack((X1, X2))\n",
    "\t# generate class labels\n",
    "\ty = ones((n, 1))\n",
    "\treturn X, y\n",
    "\n",
    "# generate n fake samples with class labels\n",
    "def generate_fake_samples(n):\n",
    "\t# generate inputs in [-1, 1]\n",
    "\tX1 = -1 + rand(n) * 2\n",
    "\t# generate outputs in [-1, 1]\n",
    "\tX2 = -1 + rand(n) * 2\n",
    "\t# stack arrays\n",
    "\tX1 = X1.reshape(n, 1)\n",
    "\tX2 = X2.reshape(n, 1)\n",
    "\tX = hstack((X1, X2))\n",
    "\t# generate class labels\n",
    "\ty = zeros((n, 1))\n",
    "\treturn X, y\n",
    "\n",
    "# train the discriminator model\n",
    "def train_discriminator(model, n_epochs=1000, n_batch=128):\n",
    "\thalf_batch = int(n_batch / 2)\n",
    "\t# run epochs manually\n",
    "\tfor i in range(n_epochs):\n",
    "\t\t# generate real examples\n",
    "\t\tX_real, y_real = generate_real_samples(half_batch)\n",
    "\t\t# update model\n",
    "\t\tmodel.train_on_batch(X_real, y_real)\n",
    "\t\t# generate fake examples\n",
    "\t\tX_fake, y_fake = generate_fake_samples(half_batch)\n",
    "\t\t# update model\n",
    "\t\tmodel.train_on_batch(X_fake, y_fake)\n",
    "\t\t# evaluate the model\n",
    "\t\t_, acc_real = model.evaluate(X_real, y_real, verbose=2)\n",
    "\t\t_, acc_fake = model.evaluate(X_fake, y_fake, verbose=2)\n",
    "\t\tprint(i, acc_real, acc_fake)\n",
    "\n",
    "# define the discriminator model\n",
    "model = define_discriminator()\n",
    "# fit the model\n",
    "train_discriminator(model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_37\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_67 (Dense)             (None, 15)                90        \n",
      "_________________________________________________________________\n",
      "dense_68 (Dense)             (None, 30)                480       \n",
      "_________________________________________________________________\n",
      "dense_69 (Dense)             (None, 2)                 62        \n",
      "=================================================================\n",
      "Total params: 632\n",
      "Trainable params: 632\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# define the generator model\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "from keras.utils.vis_utils import plot_model\n",
    "\n",
    "# define the standalone generator model\n",
    "def define_generator(latent_dim, n_outputs=2):\n",
    "    model = Sequential()\n",
    "    model.add(Dense(15, activation='relu', kernel_initializer='he_uniform', input_dim=latent_dim))\n",
    "    model.add(Dense(30, activation='relu'))\n",
    "    model.add(Dense(n_outputs, activation='tanh'))\n",
    "    return model\n",
    "\n",
    "# define the discriminator model\n",
    "model = define_generator(5)\n",
    "# summarize the model\n",
    "model.summary()\n",
    "# plot the model\n",
    "#plot_model(model, to_file='generator_plot.png', show_shapes=True, show_layer_names=True)\n",
    "#plot_model('model')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# define and use the generator model\n",
    "\n",
    "# generate points in latent space as input for the generator\n",
    "def generate_latent_points(latent_dim, n):\n",
    "\t# generate points in the latent space\n",
    "\tx_input = randn(latent_dim * n)\n",
    "\t# reshape into a batch of inputs for the network\n",
    "\tx_input = x_input.reshape(n, latent_dim)\n",
    "\treturn x_input\n",
    "\n",
    "# use the generator to generate n fake examples and plot the results\n",
    "def generate_fake_samples(generator, latent_dim, n):\n",
    "\t# generate points in latent space\n",
    "\tx_input = generate_latent_points(latent_dim, n)\n",
    "\t# predict outputs\n",
    "\tX = generator.predict(x_input)\n",
    "\t# plot the results\n",
    "\tpyplot.scatter(X[:, 0], X[:, 1])\n",
    "\tpyplot.show()\n",
    "\n",
    "# size of the latent space\n",
    "latent_dim = 5\n",
    "# define the discriminator model\n",
    "model = define_generator(latent_dim)\n",
    "# generate and plot generated samples\n",
    "generate_fake_samples(model, latent_dim, 100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(100, 5)"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "generate_latent_points(5,100).shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [],
   "source": [
    "# use the generator to generate n fake examples, with class labels\n",
    "def generate_fake_samples(generator, latent_dim, n):\n",
    "\t# generate points in latent space\n",
    "\tx_input = generate_latent_points(latent_dim, n)\n",
    "\t# predict outputs\n",
    "\tX = generator.predict(x_input)\n",
    "\t# create class labels\n",
    "\ty = zeros((n, 1))\n",
    "\treturn X, y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_41\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "sequential_40 (Sequential)   (None, 2)                 632       \n",
      "_________________________________________________________________\n",
      "sequential_39 (Sequential)   (None, 1)                 886       \n",
      "=================================================================\n",
      "Total params: 1,518\n",
      "Trainable params: 632\n",
      "Non-trainable params: 886\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# define the combined generator and discriminator model, for updating the generator\n",
    "def define_gan(generator, discriminator):\n",
    "\t# make weights in the discriminator not trainable\n",
    "\tdiscriminator.trainable = False\n",
    "\t# connect them\n",
    "\tmodel = Sequential()\n",
    "\t# add generator\n",
    "\tmodel.add(generator)\n",
    "\t# add the discriminator\n",
    "\tmodel.add(discriminator)\n",
    "\t# compile model\n",
    "\tmodel.compile(loss='binary_crossentropy', optimizer='adam')\n",
    "\treturn model\n",
    " \n",
    "# create the discriminator\n",
    "discriminator = define_discriminator()\n",
    "# create the generator\n",
    "generator = define_generator(latent_dim)\n",
    "# create the gan\n",
    "gan_model = define_gan(generator, discriminator)\n",
    "# summarize gan model\n",
    "gan_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"# train the composite model\\ndef train_gan(gan_model, latent_dim, n_epochs=10000, n_batch=128):\\n\\t# manually enumerate epochs\\n\\tfor i in range(n_epochs):\\n\\t\\t# prepare points in latent space as input for the generator\\n\\t\\tx_gan = generate_latent_points(latent_dim, n_batch)\\n\\t\\t# create inverted labels for the fake samples\\n\\t\\ty_gan = ones((n_batch, 1))\\n\\t\\t# update the generator via the discriminator's error\\n\\t\\tgan_model.train_on_batch(x_gan, y_gan)\""
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "'''# train the composite model\n",
    "def train_gan(gan_model, latent_dim, n_epochs=10000, n_batch=128):\n",
    "\t# manually enumerate epochs\n",
    "\tfor i in range(n_epochs):\n",
    "\t\t# prepare points in latent space as input for the generator\n",
    "\t\tx_gan = generate_latent_points(latent_dim, n_batch)\n",
    "\t\t# create inverted labels for the fake samples\n",
    "\t\ty_gan = ones((n_batch, 1))\n",
    "\t\t# update the generator via the discriminator's error\n",
    "\t\tgan_model.train_on_batch(x_gan, y_gan)'''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [],
   "source": [
    "# train a generative adversarial network on a one-dimensional function\n",
    "\n",
    "# evaluate the discriminator and plot real and fake points\n",
    "def summarize_performance(epoch, generator, discriminator, latent_dim, n=100):\n",
    "\t# prepare real samples\n",
    "\tx_real, y_real = generate_real_samples(n)\n",
    "\t# evaluate discriminator on real examples\n",
    "\t_, acc_real = discriminator.evaluate(x_real, y_real, verbose=0)\n",
    "\t# prepare fake examples\n",
    "\tx_fake, y_fake = generate_fake_samples(generator, latent_dim, n)\n",
    "\t# evaluate discriminator on fake examples\n",
    "\t_, acc_fake = discriminator.evaluate(x_fake, y_fake, verbose=0)\n",
    "\t# summarize discriminator performance\n",
    "\tprint(epoch, acc_real, acc_fake)\n",
    "\t# scatter plot real and fake data points\n",
    "\tpyplot.scatter(x_real[:, 0], x_real[:, 1], color='red')\n",
    "\tpyplot.scatter(x_fake[:, 0], x_fake[:, 1], color='blue')\n",
    "\tpyplot.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1999 0.550000011920929 0.9399999976158142\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3999 0.5699999928474426 0.550000011920929\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5999 0.3400000035762787 0.44999998807907104\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "7999 0.8999999761581421 0.9200000166893005\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "9999 0.5199999809265137 0.4099999964237213\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAD8CAYAAAB6paOMAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nO3df5DcdZ3n8ed7ehJIJ6BkJiJFmG4subqDrStdRyzP8kctKGhtEasOPcMkG0EdkxEvVdZ6i8xdeYU3qGvduhSahEHAmPn6++pWapeVgyiWVp0ck113FS0kkOkhQkmYIEUYMJmZ9/3x/fZMd8/3O909/WP6x+tR1dX9/dX9+U7g+/5+fr2/5u6IiEj36lnrAoiIyNpSIBAR6XIKBCIiXU6BQESkyykQiIh0OQUCEZEuV5dAYGZXm9ljZnbUzG6K2f4OM/snM5szs2tLts2b2S+i1731KI+IiFTOap1HYGYp4LfAu4HjwCPAdnf/dcE+WeBc4C+Be939+wXbTrn7ppoKISIiq9Zbh++4HDjq7k8CmNm3gW3AYiBw96lo20Idfk9EROqoHoHgQuCpguXjwFuqOP5sM5sE5oAvuPvflTugv7/fs9lsVYUUEel2R44cec7dt5Sur0cgsJh11bQ3Dbj702b2OuBHZvZLd39i2Y+YDQPDAAMDA0xOTq6utCIiXcrMcnHr69FZfBy4qGB5K/B0pQe7+9PR+5PAQ8AbE/Ybd/dBdx/csmVZQBMRkVWqRyB4BLjEzC42s/XAh4CKRv+Y2Xlmdlb0uR94GwV9CyIi0ng1BwJ3nwNuBO4HfgN8190fNbNbzOwaADN7s5kdBz4A3GFmj0aH/ztg0sz+BfgxYR+BAoGISBPVPHx0LQwODrr6CEREqmNmR9x9sHS9ZhaLiHQ5BQIRkS6nQCAi0uUUCEREupwCgYhIl1MgEBHpcgoEIiJdToFARKTLKRCIiHQ5BQIRkS6nQCAi0uUUCEREWlUQQDYLPT3hexA05Gfq8WAaERGptyCA4WGYnQ2Xc7lwGWBoqK4/pRqBiEgrGh1dCgJ5s7Ph+jpTIBARaUXT09Wtr4ECgYhIKxoYqG59DRQIRERa0dgYpNPF69LpcH2dKRCIiLSioSEYH4dMBszC9/HxuncUg0YNiYi0rqGhhlz4S6lGICLS5RQIRES6nAKBiEiXUyAQEVlLQQD9/WGHsFn4uUGpJJKos1hEZK0EAVx/PZw5s7RuZgZuuCH83ISOYlCNQERk7ezdWxwE8k6fbkgqiSQKBCIiayEIwrv/JA1IJZFEgUBEZC3s3r3y9gakkkhSl0BgZleb2WNmdtTMborZ/g4z+yczmzOza0u27TKzx6PXrnqUR0SkpQUBnDqVvH39+oakkkhScyAwsxTwVeC9wKXAdjO7tGS3aeDDwDdLjt0MfBZ4C3A58FkzO6/WMomItLSPf3zl7Xff3bSOYqhPjeBy4Ki7P+nup4FvA9sKd3D3KXf/V2Ch5NirgAfc/aS7Pw88AFxdhzKJiLSmkRF46aWiVWlewFhYfKU/1rwgAPUJBBcCTxUsH4/WNfpYEZH2c+BA0WKaF3iZcwBbfL388vLEo41Uj0BgMeu83sea2bCZTZrZ5IkTJyounIhIywgC8OJL3FIQKPbyy00qE/UJBMeBiwqWtwJP1/tYdx9390F3H9yyZcuqCioisqaaODegGvUIBI8Al5jZxWa2HvgQcG+Fx94PvMfMzos6id8TrRMR6TxNnBtQjZoDgbvPATcSXsB/A3zX3R81s1vM7BoAM3uzmR0HPgDcYWaPRseeBD5HGEweAW6J1omIdJ6YuQEbeJG4FvENG5pQnoi5V9qc3zoGBwd9cnJyrYshIlKdIIDhYZidXVqXTpM+8zwvn1m/uGrDhuJd6sXMjrj7YOl6zSwWEWmWhMdPzp5ejzuLr0YEgZUoEIiINEoQQDYLPT3hexCEwWBqChYWwvcmThxLojTUIiKNUNoMlMuFy9ASF/9CqhGIiDTC6OjyNp7Z2ZYcQqpAICLSCElDRVtwCKkCgYhIIySlkW5ieulKKRCIiNRLYefwqVOwbl3x9nS6qemlK6VAICJSD/nO4VwuHAM6MxMOEe3rKxoq2modxaBAICJSH6OjBLPbyHKMHubJcozg9H+ETZtaaqhoHA0fFRGpgyD3NoYZZ5aNAOTIMsydkBumNS//S1QjEBGpg709ty8GgbxZNjKa+uIalahyCgQiIjUIAujvh5mF+KfsTs+3/rO21DQkIrJKQQDXXw9nzkD8c7ZgIBO/vpWoRiAiskof/3g+CCRrwdGiyygQiIisQhAsewb9Mn19LTtQqIgCgYhINaJJY6M7psruetttjS9OPaiPQESkUgUZRadZOVVEu9QGQDUCEZHKFWQUHSA5edy6de1TGwAFAhGRygRBmD4iMsbNpFneSbBxI9xzT/vUBkCBQESkvHyTUIEhvsU4HyPDFMYCmQxMTIS55topCIACgYjIyoKAYOc/kp19dCmHENuBMBhMpS9jYeJbrZxKqCx1FouIJAhGfsbe/Vcxw3XkJ4wt5hAiDAStmlG0GqoRiIjECAIY3v+nzNBP6azhWTYyyq1hauk2DwKgQCAiEmt07ylmSSdun2agPaYNV0CBQEQkxvRMchAAGOib7YjaACgQiIjEWmmeQHr9HGO3bWpiaRpLgUBEJMZY39/EzBNw+myG8bt7O6UyANQpEJjZ1Wb2mJkdNbObYrafZWbfibY/bGbZaH3WzF42s19ErwP1KI+ISNUKHzyfzTL0wTnG1924NE+AKSbW3cBzh37YUUEA6hAIzCwFfBV4L3ApsN3MLi3Z7SPA8+7+euDLQOEje55w9zdEr921lkdEpGojI7Bz59KD53M5OHiQoY9uYCrzLhasl6nMuxi658qO6RcoVI95BJcDR939SQAz+zawDfh1wT7bgP8eff4+8BUza/2nNYhI5wsCOHAgDACFZmfhvvvCh853uHo0DV0IPFWwfDxaF7uPu88BLwB90baLzeyfzewnZvb2pB8xs2EzmzSzyRMnTtSh2CIihInkSoNA3nRyh3EnqUcgiLuzL/2rJu3zDDDg7m8EPgV808zOjfsRdx9390F3H9yyZUtNBRYRWbTSxX5g5VTTnaIegeA4cFHB8lbg6aR9zKwXeBVw0t3/6O4zAO5+BHgC+Dd1KJOISGWSLvZmHTNhrJx6BIJHgEvM7GIzWw98CLi3ZJ97gV3R52uBH7m7m9mWqLMZM3sdcAnwZB3KJCKSqGiA0KlfEaz7cPEOZrB7d0d2DMepubPY3efM7EbgfiAF3O3uj5rZLcCku98L3AUcMrOjwEnCYAHwDuAWM5sD5oHd7n6y1jKJiCQpeMgYALmZTQyvvxP6zmHo5FfCGsLYWNcEAQDzpE6SFjY4OOiTk5NrXQwRaSNBEPYLFzxbpkgm0/kDhMzsiLsPlq5XGmoR6XiltYA4XTJAKJZSTIhIZ4o6AgIbYteOuRWDAHTNAKFYCgQi0nmimcJB7j8wzDjzZRo/0umuGSAUS4FARDpHEEB/P+zfD+6MciuzbFzxkEymIx4yVhP1EYhIZyjpCAjYTo5M4u7ptAJAnmoEItIZ9u5dDAIj3M5OJohPahBSEFiiQCAi7W9kBGZmgLAmcIARvMzlTUFgiQKBiLS3fPbQyCi3lg0CqVSjC9VeFAhEpL2VZA+dpvw40OHhRhao/SgQiEh7K5kJttKzhs1gzx7Yt6/RhWovCgQi0h7yQ0PNwld/f7iuZCbYGDcve9ZwPgAsLCgIxFEgEJHWFwRwww2LHcIAzMwwsuMFeqefwFiglzOMcDtDfItxhslsmsEsnCdw6JACwEo0j0BEWtcKmeJGuJ397AEPh4jO08t+PgEbz2HfHb0MDfUtO0biKfuoiLSmK6+Ew4cTN/dyJjZ1RCoFc3ONLFj7Sso+qqYhEWktQQCbNiUGgYDtZDnGPPFjQOfnG1m4zqSmIRFpHWXyRQdsZ5g7V8wfpDkC1VMgEJG1Udr0c8UVcPToig8NqCSJnOYIVE9NQyLSXCMj4XjO0qafw4eTHx8WWWmyWCqlOQKrpRqBiDReuedEVmjg7BPkXjl/2fpueMxkI6lGICKNk58EtmNH5UEgnV6+rqcH9uxh7GvnL9vc7Q+VqQcFAhFpjHzHb+EksKRdo5FAPcyT3fB7gr5PEnAd2dRT9LBA9qJ5grftY2goTB+dybA4WUzppGuneQQiUrt808/0dJjyYWysoqaggO3s5TZm6GelZweAHiRTD5pHICKNkb/zz+XCLKC5HMH1D5LNPRTe4XOMEW5fuuPnGAHbF4eCzrCFckEAwsFEo6ONP51upBqBiFRuZATGxwnmP8gon2eaAQZSv+N983/Hffw50wywmRle5FxOc1bBgU7hxT7NS2xgNgoClTMLE8fJ6qhGICLLBAFks2FfbDYbLi+uNIPeXgK7jmzvcXpsgez+/8LI/N8yzJ3kyOAYufmt7OcT5Mji9DDDlpIgAKV3/LNsrDoIwLJEo1InGj4q0qVKJ/HmcjC84yXgHxgibNsP5j8YzuSdDydx5cgmPAayfNNOrcw0OqhRVCMQ6SBxd/jBlXeTtVx4R29TBHYd9PYyuntm2STeWTayi28stuXv5bZlM3nLPQayUn198SNF45jB7t3qKG4Yd6/5BVwNPAYcBW6K2X4W8J1o+8NAtmDbZ6L1jwFXVfJ7b3rTm1ykKSYm3DMZd7Pwfc8e974+97Bb1L2nJ3zv6wtfZsWfM5nwO+K+r/SYjRuXvrfgNWFDntn03PKvm5goKstEaqenOVV0+Pqe076Ol4vWpTnlE2x3Yz7u50peCxXsU/0rnQ6LH/fnzWTCfVKp8L30TyirB0x63DU8bmU1LyAFPAG8DlgP/Atwack+I8CB6POHgO9Eny+N9j8LuDj6nlS531QgkLpJuhKZuW/aVP+rXjpddv8JtnuGY27Mex/PLr+Qp90n9vzUff36ouMyHKu4SBmOVbV/6atcEFm3rjgWFv5ZdWFfO40MBG8F7i9Y/gzwmZJ97gfeGn3uBZ4jbFQs2rdwv5VeCgRdLn/xrvW2scILczUX7gzHfILtMVfezFKZy3xX6V198p36gvfx7OLvVXaHv3Qhr+63lpbTnPI93O4ZptxY0IW+jTQyEFwLfK1geSfwlZJ9fgVsLVh+AugHvgLsKFh/F3Btwu8MA5PA5MDAQIP/XNKSSppClr3yd96VWuHCXNHFPdqvl9OxF+kUZ3wPt0dXXgtfZa66q7tLn3eY9xRnKj4mw7Go/Nd5xnLRd8Q3A/XxrGdSTy1e9HWRb19JgaAevT5xwwW8wn0qOTZc6T7u7oPuPrhlS/XDzqTNVZKuoIIZR0WdqbmHGOF2+nkWYwFjgX6eZYTbuZ67F4dD5sjyF3ydfp4tmhAFsJsDzLGu5FcMsMVHJ45wezjusYKxjytl10zWA/RET+sq/t9nPX9kHa8UrUvzEmPcDMDQnlcxtnuatL1C3P+O6TTcNrGFqbmtLLgxNaUO244UFx2qeaGmIWmGCppVFu+8I3HN/8tbguLugst3kOY7XCvZN8WZivsIamm3L/y9xZrMFXf5xBV3Rc04BbWbVCr8g6zwp02ldPffaWhg01Av8CRhZ2++s/iykn0+QXFn8Xejz5dR3Fn8JOosljgVNKuEbR6ZxBakSr+imuaVykbVLCydR5lRQ/Ht9tWN3DHmiy70q/3TFsRU6RANCwThd/M+4LeEbf+j0bpbgGuiz2cD3yMcJvr/gNcVHDsaHfcY8N5Kfk+BoI2U3pbH3GLG7bLsemnPLY6iOZtZz7fDG3NL7fDptE/s+Wk9+n8ru+BW1uzvqVRtf7KVukUSYmFVkmoE1X6PtL6GBoJmvxQI2sSePcuvlCUdunG7VP9a8D0bv+4+MVFxC1JdagSZsPzl9qvwxjxRNYObqu0vT/r+1XyPtD4FAmmuiYmiK3zRKJzUU8vmZdX6yt91lwsq9WoeKrxQ7tmzNIq18PuraJ2p6M8ZV7GqoMJV0/dLZ0kKBMo+KnW1mJY+t8BmwhE+M/SRH0nTSO7hiKCkFPjpNOzaBffdt5Q2//WvX/7o3HXr4KMfLd7vfe8rXh4b0+gZaT9J2UeVdE5WJ+ZBJAFDBUnMelaVXXK1UqnwfWysOJFaXl8f3HZb/MU77pkqushLN1GNQKpXmrYSIJ0mu+H35GY2rUmR9uyBffuWiqcLu8hySTUCBQKpXP4Km9D20sN83TJTltq4MYw7mzfDSy/BK9EcqXxWynwQEJFkahqS2sTVAkoMME2ObM0/1dMDGzaEP6U7epHG0/MIpDKjo4tBIGD74vNnN/ECKeYwFsiRqfjrUinIZMI7+kwGJiaWxuTMz8OpU+EjCZXSQKTxFAhkubinm0xPh5uiB47n8/C8xLkskKKaUUGpFBw8GF7kdbEXWXsKBFIs3wSUy4W357lcuLx5MwCj3LrsiVWVsChGZDJhENCFX6R1KBBIsYImoEX55XS66uyY+aafQ4fCuKK7f5HWo85iKRY1AS1z8iQcOsTArqfJzW+t6KsymfDCLyKtTTWCbjQyAr294e16b2+4HAk237jYEVyYd5+BARgaYuzg1ooeOJ5Oh6N9RKT1qUbQbS67DH7966Xl+XnYvx+A4G37GH7xb5iN/rPIkWWYO2HdWQyNXQksNevkJ2yl0/Dyy2Gnb14moyGfIu1EE8q6ycjI4kV/mVSK/lfPxT4ArG/TKzz34tmNLZuINFzShDI1DXWTO+5I3BTMfzDxKZAzp84mCBpUJhFZcwoE3aSw/abEKJ9f8dAyjwIWkTamQNClCmcHZzlGrsyw0KTBRCLS/hQIOlHczGAIM7exfHZwjixmK88KHqhu+oCItBGNGuo0pcnh8jODIewj2LWL0fnls4Pdw9GkcWMHNBRUpLOpRtBpYmYGB7PbyO56Jz07h8i++g+JzUDu4dBPWHrQSyYD4+MaCirSyVQj6BQJzwrINwPNzoc1gNzMpsQ7f80EFulOCgSdYIVnBcQliYtrBlLzj0j3UtNQJ4hLFBdJShKXbwbKJ4VT849I91KNoBOsMLZzIBWfJE7NQCKSpxpBB0hMFJfJxCaJUzOQiBRSjaDNBQErJoorTRKnZwCLSCklnWtz2eyygUIAZPpOMfXcpqaXR0RaV0OSzpnZZjN7wMwej97PS9hvV7TP42a2q2D9Q2b2mJn9Inq9ppbydJsgiA8CANMnFQREpDK19hHcBBx290uAw9FyETPbDHwWeAtwOfDZkoAx5O5viF7P1liezhQE0N8fDvExg/5+gpGfLU4YjqOUECJSqVoDwTbgYPT5IPD+mH2uAh5w95Pu/jzwAHB1jb/bPYIAbriBohzRMzOM7r8oacSoOoNFpCq1BoLz3f0ZgOg9rmnnQuCpguXj0bq8e6Jmof9mK2Q+M7NhM5s0s8kTJ07UWOw2MjoKp08vLuazhq6ULVRzAkSkGmVHDZnZg8BrYzZVmqE+7uKe76Eecvffmdk5wP8CdgLfiPsSdx8HxiHsLK7wt9tfwRyBEW7nACP4CvE7k1EQEJHqlA0E7n5l0jYz+72ZXeDuz5jZBUBcG/9x4F0Fy1uBh6Lv/l30/qKZfZOwDyE2EHStgQHI5QjYXjYIqElIRFaj1qahe4H8KKBdwA9i9rkfeI+ZnRd1Er8HuN/Mes2sH8DM1gF/DvyqxvJ0nrExWL+eUW4tWxNQk5CIrEatE8q+AHzXzD4CTAMfADCzQWC3u3/U3U+a2eeAR6JjbonWbSQMCOuAFPAgcGeN5ek80ZV9ekdyn4DSRYhILTShrJXkU0nHTAFOmjhmBocOqSYgIuU1ZEKZ1FE+lXQuF6YGzT9ZLHrM5NgYy3IGmcHu3QoCIlIbBYJWEASwa9fyVNKzs2ENgfBiPz5enDr60CHYt28NyisiHUVJ59ZaviYwPx+/vWD46NCQ7v5FpP4UCNZSviaQFARAuSJEpOHUNLRWytUEQBMDRKQpFAjWygqPlwQgldLEABFpCgWCtbBS/mgIawIHDyoIiEhTKBA0W75JKIlqAiLSZAoEzbZSk5BqAiKyBhQImq1gOOgyqgmIyBpQIGi2pOGgyh8tImtEgaDZ4nJFaJioiKwhBYJmi8sVoSYhEVlDmlm8FpQrQkRaiGoEjRQEYf7onp7wPcokKiLSSlQjaJSREThwIEwpDUtppUG1ARFpKaoRNEIQwP794E7AdrIco4d5srOPEux9eK1LJyJSRIGgEfbuBSBgO8PcSY4sTg85sgzPfF4tRCLSUhQIGmFmBoBRbmWWjUWbZtmYf9aMiEhLUCBooGniJ4+tNLlYRKTZFAjqKRollO8XcCx2Nz1rRkRaiUYN1UuUVTSY3cYwdy5rEsrTJGIRaTWqEdTL6Cgjs19kB0FCEHBNIhaRlqQaQZ2M5D7NfkYgoTnIzJiaamqRREQqohpBraJ+gXE+TlIQAPULiEjrUo2gFvmnjc3OMk8qcTf1C4hIK1ONoBYFTxtLMZ+4m/oFRKSV1RQIzGyzmT1gZo9H7+cl7PdDM/uDmf19yfqLzezh6PjvmNn6WsrTdAUTAoY5APiyXfbsURAQkdZWa43gJuCwu18CHI6W43wJ2Bmz/ovAl6Pjnwc+UmN5micIwqyikX18kj18lRRzgJNKhUFg3761K6KISCVqDQTbgIPR54PA++N2cvfDwIuF68zMgD8Dvl/u+JaT7xuYL24O2scnmUu/Cp/4JnNzCgIi0h5qDQTnu/szANH7a6o4tg/4g7vPRcvHgQuTdjazYTObNLPJEydOrLrAdVHQN1AklVKHgIi0nbKjhszsQeC1MZtqTZ0WN9ZyeSN7foP7ODAOMDg4mLhfUyQlC1pYUBAQkbZTtkbg7le6+5/EvH4A/N7MLgCI3p+t4refA15tZvlgtBV4utoTaKbFB475HFmOEbC9eAdNFhCRNlRr09C9wK7o8y7gB5Ue6O4O/Bi4djXHN1sQwA03hA8ayz9b4AbuWQoGmiwgIm2q1kDwBeDdZvY48O5oGTMbNLOv5Xcys58C3wOuMLPjZnZVtOmvgE+Z2VHCPoO7aixPw+zdC6dPF687zVns5TaUREhE2pm5r21z+2oMDg765ORk834wCLAd15GUQqIN/4Qi0oXM7Ii7D5au18ziMoKRn5Hd+fa1LoaISMMo19AKggCGD/wps55O3Kevr4kFEhFpANUIVjA6yopBYN06uO22JhZIRKQBFAhWMJ1Lavx3MuS45x71D4tI+1MgSBIEDJCL3ZQhxxRZBQER6QgKBElGRxnjZtK8VLQ6zUuMcXM4ZFREpAOoszjJ9DRDUY1glFuZZoABphnjZobs2zB2aI0LKCJSH6oRRBbTR/RAtv8UAdcBMMS3mOJiFkgxxcUM8S3YvVudAyLSMVQjoOiJkwDkZjYxzB3AQnjhzzMLg4DyS4tIB1GNgPis0rNsZJRbl1akUnDokIKAiHQcBQKSs0pPU5BNVCmmRaRDKRCQnD16gOnyO4mItDkFAsLs0emSCcSLw0QLdxIR6UAKBIQtPuPjkEkdx1ggwxTjfGypo7ivT81CItKxujIQFA0VzYbLQ0MwdfAnLKTPWRomCmFVQQmFRKSDdd3w0WVDRXPhMsBQ/q5/dDTsQR4YCJuEVBsQkQ7WdQ+myWbDi3+pTAampmoqlohIS9ODaSKJQ0UT1ouIdLquCwSJQ0U1OlREulTXBYLYoaJpjQ4Vke7VdYFgcahoJkwdlMmEy+oPFpFu1XWjhiC86OvCLyIS6roagYiIFFMgEBHpcgoEIiJdToFARKTLKRCIiHS5mgKBmW02swfM7PHo/byE/X5oZn8ws78vWf91MztmZr+IXm+opTwiIlK9WmsENwGH3f0S4HC0HOdLwM6EbZ929zdEr1/UWB4REalSrYFgG3Aw+nwQeH/cTu5+GHixxt8SEZEGqDUQnO/uzwBE769ZxXeMmdm/mtmXzeyspJ3MbNjMJs1s8sSJE6str4iIlCgbCMzsQTP7VcxrWx1+/zPAvwXeDGwG/ippR3cfd/dBdx/csmVLHX5aRESgghQT7n5l0jYz+72ZXeDuz5jZBcCz1fx4vjYB/NHM7gH+spLjjhw58pyZxTxVoO31A8+tdSGapFvOtVvOE7rnXNv5PDNxK2vNNXQvsAv4QvT+g2oOLggiRti/8KtKjnP3jqwSmNlk3EMjOlG3nGu3nCd0z7l24nnW2kfwBeDdZvY48O5oGTMbNLOv5Xcys58C3wOuMLPjZnZVtCkws18CvySMsv+jxvKIiEiVaqoRuPsMcEXM+kngowXLb084/s9q+X0REamdZha3lvG1LkATdcu5dst5Qveca8edZ1s+vF5EROpHNQIRkS6nQLCGKs3VFO17rpn9zsy+0swy1ksl52pmbzCz/2tmj0aTDP/TWpR1NczsajN7zMyOmtmyVCtmdpaZfSfa/rCZZZtfytpVcJ6fMrNfR/9+h80sdrhiOyh3rgX7XWtmbmZtO5JIgWBtVZqrCeBzwE+aUqrGqORcZ4G/cPfLgKuBvzWzVzexjKtiZingq8B7gUuB7WZ2acluHwGed/fXA18GvtjcUtauwvP8Z2DQ3f898H3gr5tbyvqo8Fwxs3OA/ww83NwS1pcCwdqqKFeTmb0JOB/4P00qVyOUPVd3/627Px59fppwgmI7zBm5HDjq7k+6+2ng24TnW6jw/L9POJTamljGeih7nu7+Y3efjRZ/DmxtchnrpZJ/Uwhv0P4aeKWZhas3BYK1VTZXk5n1AP8T+HSTy1ZvVeWlMrPLgfXAE00oW60uBJ4qWD4erYvdx93ngBeAvqaUrn4qOc9CHwH+saElapyy52pmbwQucvei9PrtqNaZxVKGmT0IvDZm02iFXzEC3OfuT7X6DWQdzjX/PRcAh4Bd7r5Qj7I1WNw/TOlwvEr2aXUVn4OZ7QAGgXc2tESNs+K5RjdoXwY+3KwCNZICQYPVIVfTW4G3m9kIsAlYb2an3H2l/oQ1UY+8VGZ2LvAPwH919583qKj1dhy4qGB5K/B0wj7HzawXeBVwsjnFq5tKzhMzu5Iw+L/T3f/YpLLVW7lzPQf4E+Ch6AbttcC9ZnZNNKG2rahpaG3lczVBQrpbR3UAAADhSURBVK4mdx9y9wF3zxIm5ftGKwaBCpQ9VzNbD/xvwnP8XhPLVqtHgEvM7OLoHD5EeL6FCs//WuBH3n6TeMqeZ9RccgdwjbtXlYSyxax4ru7+grv3u3s2+n/z54Tn3HZBABQI1lpFuZo6RCXn+kHgHcCHrY0eXxq1+d8I3A/8Bviuuz9qZreY2TXRbncBfWZ2FPgUK48Qa0kVnueXCGuu34v+/UoDYluo8Fw7hmYWi4h0OdUIRES6nAKBiEiXUyAQEelyCgQiIl1OgUBEpMspEIiIdDkFAhGRLqdAICLS5f4/GAFeQlZcY3UAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# train the generator and discriminator\n",
    "def train(g_model, d_model, gan_model, latent_dim, n_epochs=10000, n_batch=128, n_eval=2000):\n",
    "\t# determine half the size of one batch, for updating the discriminator\n",
    "\thalf_batch = int(n_batch / 2)\n",
    "\t# manually enumerate epochs\n",
    "\tfor i in range(n_epochs):\n",
    "\t\t# prepare real samples\n",
    "\t\tx_real, y_real = generate_real_samples(half_batch)\n",
    "\t\t# prepare fake examples\n",
    "\t\tx_fake, y_fake = generate_fake_samples(g_model, latent_dim, half_batch)\n",
    "\t\t# update discriminator\n",
    "\t\td_model.train_on_batch(x_real, y_real)\n",
    "\t\td_model.train_on_batch(x_fake, y_fake)\n",
    "\t\t# prepare points in latent space as input for the generator\n",
    "\t\tx_gan = generate_latent_points(latent_dim, n_batch)\n",
    "\t\t# create inverted labels for the fake samples\n",
    "\t\ty_gan = ones((n_batch, 1))\n",
    "\t\t# update the generator via the discriminator's error\n",
    "\t\tgan_model.train_on_batch(x_gan, y_gan)\n",
    "\t\t# evaluate the model every n_eval epochs\n",
    "\t\tif (i+1) % n_eval == 0:\n",
    "\t\t\tsummarize_performance(i, g_model, d_model, latent_dim)\n",
    "\n",
    "# train model\n",
    "train(generator, discriminator, gan_model, latent_dim)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}