play.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "import random\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(42)\n",
    "random.seed(42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.datasets import make_moons\n",
    "X, y = make_moons(n_samples=1000, noise=0.1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x177b899a0>"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 500x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(5,5))\n",
    "plt.scatter(X[:,0], X[:,1], c=y, s=20, cmap='jet')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "rindinces = np.random.permutation(X.shape[0])\n",
    "X = X[rindinces]\n",
    "y = y[rindinces]\n",
    "\n",
    "def standard_normalize(X):\n",
    "    mean = np.mean(X, axis=0)\n",
    "    std = np.std(X, axis=0)\n",
    "    X_standardized = (X - mean) / std\n",
    "    return X_standardized\n",
    "\n",
    "standard_normalize(X)\n",
    "\n",
    "TRAIN_SIZE = 80\n",
    "X_train, X_test = X[:TRAIN_SIZE], X[TRAIN_SIZE:]\n",
    "y_train, y_test = y[:TRAIN_SIZE], y[TRAIN_SIZE:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train = y_train.reshape(-1, 1)\n",
    "y_test = y_test.reshape(-1, 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['Layer1(2, 20)', 'Layer2(20, 10)', 'Layer3(10, 1)']"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from neural_network import NeuralNetwork, Layer, Activation\n",
    "nn = NeuralNetwork(2)\n",
    "\n",
    "nn.add(Layer(20))\n",
    "nn.add(Layer(10))\n",
    "nn.add(Layer(1, activation=Activation.sigmoid))\n",
    "nn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss at step 1: 0.672712\n",
      "Loss at step 2: 0.227291\n",
      "Loss at step 3: 0.215421\n",
      "Loss at step 4: 0.205165\n",
      "Loss at step 5: 0.196262\n",
      "Loss at step 6: 0.188277\n",
      "Loss at step 7: 0.181131\n",
      "Loss at step 8: 0.174942\n",
      "Loss at step 9: 0.169469\n",
      "Loss at step 10: 0.164511\n",
      "Loss at step 11: 0.159929\n",
      "Loss at step 12: 0.155562\n",
      "Loss at step 13: 0.151413\n",
      "Loss at step 14: 0.147520\n",
      "Loss at step 15: 0.143930\n",
      "Loss at step 16: 0.140589\n",
      "Loss at step 17: 0.137506\n",
      "Loss at step 18: 0.134657\n",
      "Loss at step 19: 0.132012\n",
      "Loss at step 20: 0.129541\n",
      "Loss at step 21: 0.127227\n",
      "Loss at step 22: 0.125033\n",
      "Loss at step 23: 0.122956\n",
      "Loss at step 24: 0.121006\n",
      "Loss at step 25: 0.119171\n",
      "Loss at step 26: 0.117440\n",
      "Loss at step 27: 0.115815\n",
      "Loss at step 28: 0.114289\n",
      "Loss at step 29: 0.112837\n",
      "Loss at step 30: 0.111453\n",
      "Loss at step 31: 0.110143\n",
      "Loss at step 32: 0.108900\n",
      "Loss at step 33: 0.107719\n",
      "Loss at step 34: 0.106594\n",
      "Loss at step 35: 0.105523\n",
      "Loss at step 36: 0.104500\n",
      "Loss at step 37: 0.103523\n",
      "Loss at step 38: 0.102595\n",
      "Loss at step 39: 0.101721\n",
      "Loss at step 40: 0.100884\n",
      "Loss at step 41: 0.100078\n",
      "Loss at step 42: 0.099304\n",
      "Loss at step 43: 0.098549\n",
      "Loss at step 44: 0.097795\n",
      "Loss at step 45: 0.097072\n",
      "Loss at step 46: 0.096383\n",
      "Loss at step 47: 0.095719\n",
      "Loss at step 48: 0.095075\n",
      "Loss at step 49: 0.094449\n",
      "Loss at step 50: 0.093844\n",
      "Loss at step 51: 0.093255\n",
      "Loss at step 52: 0.092687\n",
      "Loss at step 53: 0.092144\n",
      "Loss at step 54: 0.091612\n",
      "Loss at step 55: 0.091076\n",
      "Loss at step 56: 0.090552\n",
      "Loss at step 57: 0.090031\n",
      "Loss at step 58: 0.089460\n",
      "Loss at step 59: 0.088912\n",
      "Loss at step 60: 0.088375\n",
      "Loss at step 61: 0.087840\n",
      "Loss at step 62: 0.087294\n",
      "Loss at step 63: 0.086775\n",
      "Loss at step 64: 0.086268\n",
      "Loss at step 65: 0.085775\n",
      "Loss at step 66: 0.085292\n",
      "Loss at step 67: 0.084817\n",
      "Loss at step 68: 0.084350\n",
      "Loss at step 69: 0.083889\n",
      "Loss at step 70: 0.083433\n",
      "Loss at step 71: 0.082986\n",
      "Loss at step 72: 0.082546\n",
      "Loss at step 73: 0.082114\n",
      "Loss at step 74: 0.081688\n",
      "Loss at step 75: 0.081268\n",
      "Loss at step 76: 0.080856\n",
      "Loss at step 77: 0.080459\n",
      "Loss at step 78: 0.080067\n",
      "Loss at step 79: 0.079681\n",
      "Loss at step 80: 0.079300\n",
      "Loss at step 81: 0.078925\n",
      "Loss at step 82: 0.078550\n",
      "Loss at step 83: 0.078160\n",
      "Loss at step 84: 0.077770\n",
      "Loss at step 85: 0.077348\n",
      "Loss at step 86: 0.076933\n",
      "Loss at step 87: 0.076527\n",
      "Loss at step 88: 0.076145\n",
      "Loss at step 89: 0.075770\n",
      "Loss at step 90: 0.075415\n",
      "Loss at step 91: 0.075075\n",
      "Loss at step 92: 0.074744\n",
      "Loss at step 93: 0.074427\n",
      "Loss at step 94: 0.074121\n",
      "Loss at step 95: 0.073838\n",
      "Loss at step 96: 0.073558\n",
      "Loss at step 97: 0.073283\n",
      "Loss at step 98: 0.073010\n",
      "Loss at step 99: 0.072741\n",
      "Loss at step 100: 0.072472\n",
      "Loss at step 101: 0.072194\n",
      "Loss at step 102: 0.071922\n",
      "Loss at step 103: 0.071668\n",
      "Loss at step 104: 0.071417\n",
      "Loss at step 105: 0.071168\n",
      "Loss at step 106: 0.070920\n",
      "Loss at step 107: 0.070676\n",
      "Loss at step 108: 0.070437\n",
      "Loss at step 109: 0.070202\n",
      "Loss at step 110: 0.069971\n",
      "Loss at step 111: 0.069741\n",
      "Loss at step 112: 0.069513\n",
      "Loss at step 113: 0.069288\n",
      "Loss at step 114: 0.069064\n",
      "Loss at step 115: 0.068844\n",
      "Loss at step 116: 0.068628\n",
      "Loss at step 117: 0.068417\n",
      "Loss at step 118: 0.068212\n",
      "Loss at step 119: 0.068010\n",
      "Loss at step 120: 0.067812\n",
      "Loss at step 121: 0.067615\n",
      "Loss at step 122: 0.067420\n",
      "Loss at step 123: 0.067228\n",
      "Loss at step 124: 0.067040\n",
      "Loss at step 125: 0.066856\n",
      "Loss at step 126: 0.066672\n",
      "Loss at step 127: 0.066490\n",
      "Loss at step 128: 0.066310\n",
      "Loss at step 129: 0.066134\n",
      "Loss at step 130: 0.065959\n",
      "Loss at step 131: 0.065787\n",
      "Loss at step 132: 0.065616\n",
      "Loss at step 133: 0.065447\n",
      "Loss at step 134: 0.065281\n",
      "Loss at step 135: 0.065115\n",
      "Loss at step 136: 0.064952\n",
      "Loss at step 137: 0.064791\n",
      "Loss at step 138: 0.064632\n",
      "Loss at step 139: 0.064474\n",
      "Loss at step 140: 0.064317\n",
      "Loss at step 141: 0.064163\n",
      "Loss at step 142: 0.064010\n",
      "Loss at step 143: 0.063859\n",
      "Loss at step 144: 0.063709\n",
      "Loss at step 145: 0.063560\n",
      "Loss at step 146: 0.063413\n",
      "Loss at step 147: 0.063269\n",
      "Loss at step 148: 0.063125\n",
      "Loss at step 149: 0.062983\n",
      "Loss at step 150: 0.062843\n",
      "Loss at step 151: 0.062702\n",
      "Loss at step 152: 0.062563\n",
      "Loss at step 153: 0.062426\n",
      "Loss at step 154: 0.062289\n",
      "Loss at step 155: 0.062154\n",
      "Loss at step 156: 0.062020\n",
      "Loss at step 157: 0.061887\n",
      "Loss at step 158: 0.061755\n",
      "Loss at step 159: 0.061624\n",
      "Loss at step 160: 0.061495\n",
      "Loss at step 161: 0.061367\n",
      "Loss at step 162: 0.061240\n",
      "Loss at step 163: 0.061120\n",
      "Loss at step 164: 0.061001\n",
      "Loss at step 165: 0.060883\n",
      "Loss at step 166: 0.060768\n",
      "Loss at step 167: 0.060656\n",
      "Loss at step 168: 0.060547\n",
      "Loss at step 169: 0.060446\n",
      "Loss at step 170: 0.060346\n",
      "Loss at step 171: 0.060246\n",
      "Loss at step 172: 0.060146\n",
      "Loss at step 173: 0.060047\n",
      "Loss at step 174: 0.059948\n",
      "Loss at step 175: 0.059850\n",
      "Loss at step 176: 0.059751\n",
      "Loss at step 177: 0.059653\n",
      "Loss at step 178: 0.059556\n",
      "Loss at step 179: 0.059459\n",
      "Loss at step 180: 0.059365\n",
      "Loss at step 181: 0.059271\n",
      "Loss at step 182: 0.059176\n",
      "Loss at step 183: 0.059071\n",
      "Loss at step 184: 0.058967\n",
      "Loss at step 185: 0.058864\n",
      "Loss at step 186: 0.058762\n",
      "Loss at step 187: 0.058661\n",
      "Loss at step 188: 0.058561\n",
      "Loss at step 189: 0.058463\n",
      "Loss at step 190: 0.058365\n",
      "Loss at step 191: 0.058268\n",
      "Loss at step 192: 0.058172\n",
      "Loss at step 193: 0.058078\n",
      "Loss at step 194: 0.057989\n",
      "Loss at step 195: 0.057900\n",
      "Loss at step 196: 0.057812\n",
      "Loss at step 197: 0.057725\n",
      "Loss at step 198: 0.057638\n",
      "Loss at step 199: 0.057552\n",
      "Loss at step 200: 0.057465\n",
      "Loss at step 201: 0.057380\n",
      "Loss at step 202: 0.057295\n",
      "Loss at step 203: 0.057210\n",
      "Loss at step 204: 0.057126\n",
      "Loss at step 205: 0.057042\n",
      "Loss at step 206: 0.056959\n",
      "Loss at step 207: 0.056876\n",
      "Loss at step 208: 0.056794\n",
      "Loss at step 209: 0.056713\n",
      "Loss at step 210: 0.056633\n",
      "Loss at step 211: 0.056554\n",
      "Loss at step 212: 0.056475\n",
      "Loss at step 213: 0.056396\n",
      "Loss at step 214: 0.056318\n",
      "Loss at step 215: 0.056239\n",
      "Loss at step 216: 0.056162\n",
      "Loss at step 217: 0.056085\n",
      "Loss at step 218: 0.056008\n",
      "Loss at step 219: 0.055931\n",
      "Loss at step 220: 0.055855\n",
      "Loss at step 221: 0.055780\n",
      "Loss at step 222: 0.055704\n",
      "Loss at step 223: 0.055629\n",
      "Loss at step 224: 0.055554\n",
      "Loss at step 225: 0.055480\n",
      "Loss at step 226: 0.055406\n",
      "Loss at step 227: 0.055332\n",
      "Loss at step 228: 0.055258\n",
      "Loss at step 229: 0.055184\n",
      "Loss at step 230: 0.055111\n",
      "Loss at step 231: 0.055038\n",
      "Loss at step 232: 0.054965\n",
      "Loss at step 233: 0.054892\n",
      "Loss at step 234: 0.054820\n",
      "Loss at step 235: 0.054748\n",
      "Loss at step 236: 0.054677\n",
      "Loss at step 237: 0.054606\n",
      "Loss at step 238: 0.054535\n",
      "Loss at step 239: 0.054464\n",
      "Loss at step 240: 0.054393\n",
      "Loss at step 241: 0.054322\n",
      "Loss at step 242: 0.054251\n",
      "Loss at step 243: 0.054180\n",
      "Loss at step 244: 0.054110\n",
      "Loss at step 245: 0.054040\n",
      "Loss at step 246: 0.053969\n",
      "Loss at step 247: 0.053900\n",
      "Loss at step 248: 0.053830\n",
      "Loss at step 249: 0.053761\n",
      "Loss at step 250: 0.053692\n",
      "Loss at step 251: 0.053623\n",
      "Loss at step 252: 0.053555\n",
      "Loss at step 253: 0.053488\n",
      "Loss at step 254: 0.053420\n",
      "Loss at step 255: 0.053353\n",
      "Loss at step 256: 0.053286\n",
      "Loss at step 257: 0.053220\n",
      "Loss at step 258: 0.053153\n",
      "Loss at step 259: 0.053087\n",
      "Loss at step 260: 0.053021\n",
      "Loss at step 261: 0.052956\n",
      "Loss at step 262: 0.052890\n",
      "Loss at step 263: 0.052825\n",
      "Loss at step 264: 0.052761\n",
      "Loss at step 265: 0.052697\n",
      "Loss at step 266: 0.052633\n",
      "Loss at step 267: 0.052569\n",
      "Loss at step 268: 0.052505\n",
      "Loss at step 269: 0.052443\n",
      "Loss at step 270: 0.052381\n",
      "Loss at step 271: 0.052319\n",
      "Loss at step 272: 0.052257\n",
      "Loss at step 273: 0.052195\n",
      "Loss at step 274: 0.052134\n",
      "Loss at step 275: 0.052073\n",
      "Loss at step 276: 0.052012\n",
      "Loss at step 277: 0.051951\n",
      "Loss at step 278: 0.051891\n",
      "Loss at step 279: 0.051831\n",
      "Loss at step 280: 0.051771\n",
      "Loss at step 281: 0.051712\n",
      "Loss at step 282: 0.051652\n",
      "Loss at step 283: 0.051597\n",
      "Loss at step 284: 0.051541\n",
      "Loss at step 285: 0.051485\n",
      "Loss at step 286: 0.051429\n",
      "Loss at step 287: 0.051374\n",
      "Loss at step 288: 0.051318\n",
      "Loss at step 289: 0.051263\n",
      "Loss at step 290: 0.051207\n",
      "Loss at step 291: 0.051152\n",
      "Loss at step 292: 0.051096\n",
      "Loss at step 293: 0.051041\n",
      "Loss at step 294: 0.050986\n",
      "Loss at step 295: 0.050931\n",
      "Loss at step 296: 0.050876\n",
      "Loss at step 297: 0.050821\n",
      "Loss at step 298: 0.050766\n",
      "Loss at step 299: 0.050715\n",
      "Loss at step 300: 0.050663\n",
      "Loss at step 301: 0.050611\n",
      "Loss at step 302: 0.050561\n",
      "Loss at step 303: 0.050511\n",
      "Loss at step 304: 0.050461\n",
      "Loss at step 305: 0.050410\n",
      "Loss at step 306: 0.050360\n",
      "Loss at step 307: 0.050311\n",
      "Loss at step 308: 0.050260\n",
      "Loss at step 309: 0.050210\n",
      "Loss at step 310: 0.050161\n",
      "Loss at step 311: 0.050110\n",
      "Loss at step 312: 0.050060\n",
      "Loss at step 313: 0.050010\n",
      "Loss at step 314: 0.049961\n",
      "Loss at step 315: 0.049913\n",
      "Loss at step 316: 0.049862\n",
      "Loss at step 317: 0.049812\n",
      "Loss at step 318: 0.049761\n",
      "Loss at step 319: 0.049713\n",
      "Loss at step 320: 0.049662\n",
      "Loss at step 321: 0.049612\n",
      "Loss at step 322: 0.049561\n",
      "Loss at step 323: 0.049513\n",
      "Loss at step 324: 0.049463\n",
      "Loss at step 325: 0.049415\n",
      "Loss at step 326: 0.049363\n",
      "Loss at step 327: 0.049315\n",
      "Loss at step 328: 0.049265\n",
      "Loss at step 329: 0.049217\n",
      "Loss at step 330: 0.049165\n",
      "Loss at step 331: 0.049117\n",
      "Loss at step 332: 0.049067\n",
      "Loss at step 333: 0.049019\n",
      "Loss at step 334: 0.048970\n",
      "Loss at step 335: 0.048922\n",
      "Loss at step 336: 0.048878\n",
      "Loss at step 337: 0.048830\n",
      "Loss at step 338: 0.048785\n",
      "Loss at step 339: 0.048738\n",
      "Loss at step 340: 0.048693\n",
      "Loss at step 341: 0.048645\n",
      "Loss at step 342: 0.048599\n",
      "Loss at step 343: 0.048553\n",
      "Loss at step 344: 0.048507\n",
      "Loss at step 345: 0.048461\n",
      "Loss at step 346: 0.048416\n",
      "Loss at step 347: 0.048371\n",
      "Loss at step 348: 0.048324\n",
      "Loss at step 349: 0.048281\n",
      "Loss at step 350: 0.048236\n",
      "Loss at step 351: 0.048193\n",
      "Loss at step 352: 0.048152\n",
      "Loss at step 353: 0.048112\n",
      "Loss at step 354: 0.048074\n",
      "Loss at step 355: 0.048035\n",
      "Loss at step 356: 0.047999\n",
      "Loss at step 357: 0.047960\n",
      "Loss at step 358: 0.047923\n",
      "Loss at step 359: 0.047888\n",
      "Loss at step 360: 0.047851\n",
      "Loss at step 361: 0.047818\n",
      "Loss at step 362: 0.047779\n",
      "Loss at step 363: 0.047745\n",
      "Loss at step 364: 0.047708\n",
      "Loss at step 365: 0.047674\n",
      "Loss at step 366: 0.047636\n",
      "Loss at step 367: 0.047601\n",
      "Loss at step 368: 0.047565\n",
      "Loss at step 369: 0.047531\n",
      "Loss at step 370: 0.047494\n",
      "Loss at step 371: 0.047458\n",
      "Loss at step 372: 0.047426\n",
      "Loss at step 373: 0.047388\n",
      "Loss at step 374: 0.047354\n",
      "Loss at step 375: 0.047317\n",
      "Loss at step 376: 0.047283\n",
      "Loss at step 377: 0.047247\n",
      "Loss at step 378: 0.047211\n",
      "Loss at step 379: 0.047176\n",
      "Loss at step 380: 0.047141\n",
      "Loss at step 381: 0.047108\n",
      "Loss at step 382: 0.047070\n",
      "Loss at step 383: 0.047036\n",
      "Loss at step 384: 0.047000\n",
      "Loss at step 385: 0.046966\n",
      "Loss at step 386: 0.046931\n",
      "Loss at step 387: 0.046897\n",
      "Loss at step 388: 0.046862\n",
      "Loss at step 389: 0.046826\n",
      "Loss at step 390: 0.046792\n",
      "Loss at step 391: 0.046757\n",
      "Loss at step 392: 0.046722\n",
      "Loss at step 393: 0.046689\n",
      "Loss at step 394: 0.046652\n",
      "Loss at step 395: 0.046619\n",
      "Loss at step 396: 0.046584\n",
      "Loss at step 397: 0.046548\n",
      "Loss at step 398: 0.046516\n",
      "Loss at step 399: 0.046479\n",
      "Loss at step 400: 0.046446\n",
      "Loss at step 401: 0.046411\n",
      "Loss at step 402: 0.046376\n",
      "Loss at step 403: 0.046343\n",
      "Loss at step 404: 0.046307\n",
      "Loss at step 405: 0.046274\n",
      "Loss at step 406: 0.046239\n",
      "Loss at step 407: 0.046204\n",
      "Loss at step 408: 0.046171\n",
      "Loss at step 409: 0.046136\n",
      "Loss at step 410: 0.046103\n",
      "Loss at step 411: 0.046071\n",
      "Loss at step 412: 0.046034\n",
      "Loss at step 413: 0.046000\n",
      "Loss at step 414: 0.045969\n",
      "Loss at step 415: 0.045932\n",
      "Loss at step 416: 0.045900\n",
      "Loss at step 417: 0.045865\n",
      "Loss at step 418: 0.045832\n",
      "Loss at step 419: 0.045801\n",
      "Loss at step 420: 0.045765\n",
      "Loss at step 421: 0.045732\n",
      "Loss at step 422: 0.045702\n",
      "Loss at step 423: 0.045665\n",
      "Loss at step 424: 0.045633\n",
      "Loss at step 425: 0.045598\n",
      "Loss at step 426: 0.045566\n",
      "Loss at step 427: 0.045534\n",
      "Loss at step 428: 0.045499\n",
      "Loss at step 429: 0.045466\n",
      "Loss at step 430: 0.045435\n",
      "Loss at step 431: 0.045400\n",
      "Loss at step 432: 0.045368\n",
      "Loss at step 433: 0.045338\n",
      "Loss at step 434: 0.045301\n",
      "Loss at step 435: 0.045269\n",
      "Loss at step 436: 0.045239\n",
      "Loss at step 437: 0.045204\n",
      "Loss at step 438: 0.045172\n",
      "Loss at step 439: 0.045138\n",
      "Loss at step 440: 0.045106\n",
      "Loss at step 441: 0.045074\n",
      "Loss at step 442: 0.045040\n",
      "Loss at step 443: 0.045008\n",
      "Loss at step 444: 0.044977\n",
      "Loss at step 445: 0.044944\n",
      "Loss at step 446: 0.044912\n",
      "Loss at step 447: 0.044880\n",
      "Loss at step 448: 0.044847\n",
      "Loss at step 449: 0.044816\n",
      "Loss at step 450: 0.044786\n",
      "Loss at step 451: 0.044754\n",
      "Loss at step 452: 0.044724\n",
      "Loss at step 453: 0.044692\n",
      "Loss at step 454: 0.044661\n",
      "Loss at step 455: 0.044631\n",
      "Loss at step 456: 0.044601\n",
      "Loss at step 457: 0.044569\n",
      "Loss at step 458: 0.044540\n",
      "Loss at step 459: 0.044509\n",
      "Loss at step 460: 0.044482\n",
      "Loss at step 461: 0.044448\n",
      "Loss at step 462: 0.044419\n",
      "Loss at step 463: 0.044390\n",
      "Loss at step 464: 0.044358\n",
      "Loss at step 465: 0.044329\n",
      "Loss at step 466: 0.044300\n",
      "Loss at step 467: 0.044267\n",
      "Loss at step 468: 0.044238\n",
      "Loss at step 469: 0.044210\n",
      "Loss at step 470: 0.044178\n",
      "Loss at step 471: 0.044149\n",
      "Loss at step 472: 0.044121\n",
      "Loss at step 473: 0.044089\n",
      "Loss at step 474: 0.044060\n",
      "Loss at step 475: 0.044031\n",
      "Loss at step 476: 0.044004\n",
      "Loss at step 477: 0.043971\n",
      "Loss at step 478: 0.043943\n",
      "Loss at step 479: 0.043915\n",
      "Loss at step 480: 0.043883\n",
      "Loss at step 481: 0.043855\n",
      "Loss at step 482: 0.043826\n",
      "Loss at step 483: 0.043795\n",
      "Loss at step 484: 0.043767\n",
      "Loss at step 485: 0.043738\n",
      "Loss at step 486: 0.043711\n",
      "Loss at step 487: 0.043678\n",
      "Loss at step 488: 0.043650\n",
      "Loss at step 489: 0.043622\n",
      "Loss at step 490: 0.043591\n",
      "Loss at step 491: 0.043564\n",
      "Loss at step 492: 0.043534\n",
      "Loss at step 493: 0.043504\n",
      "Loss at step 494: 0.043475\n",
      "Loss at step 495: 0.043447\n",
      "Loss at step 496: 0.043421\n",
      "Loss at step 497: 0.043388\n",
      "Loss at step 498: 0.043360\n",
      "Loss at step 499: 0.043332\n",
      "Loss at step 500: 0.043301\n",
      "Loss at step 501: 0.043273\n",
      "Loss at step 502: 0.043245\n",
      "Loss at step 503: 0.043219\n",
      "Loss at step 504: 0.043186\n",
      "Loss at step 505: 0.043158\n",
      "Loss at step 506: 0.043130\n",
      "Loss at step 507: 0.043100\n",
      "Loss at step 508: 0.043073\n",
      "Loss at step 509: 0.043044\n",
      "Loss at step 510: 0.043018\n",
      "Loss at step 511: 0.042985\n",
      "Loss at step 512: 0.042958\n",
      "Loss at step 513: 0.042931\n",
      "Loss at step 514: 0.042900\n",
      "Loss at step 515: 0.042873\n",
      "Loss at step 516: 0.042845\n",
      "Loss at step 517: 0.042818\n",
      "Loss at step 518: 0.042787\n",
      "Loss at step 519: 0.042760\n",
      "Loss at step 520: 0.042732\n",
      "Loss at step 521: 0.042702\n",
      "Loss at step 522: 0.042674\n",
      "Loss at step 523: 0.042647\n",
      "Loss at step 524: 0.042621\n",
      "Loss at step 525: 0.042589\n",
      "Loss at step 526: 0.042562\n",
      "Loss at step 527: 0.042536\n",
      "Loss at step 528: 0.042505\n",
      "Loss at step 529: 0.042478\n",
      "Loss at step 530: 0.042450\n",
      "Loss at step 531: 0.042423\n",
      "Loss at step 532: 0.042394\n",
      "Loss at step 533: 0.042367\n",
      "Loss at step 534: 0.042339\n",
      "Loss at step 535: 0.042313\n",
      "Loss at step 536: 0.042282\n",
      "Loss at step 537: 0.042255\n",
      "Loss at step 538: 0.042227\n",
      "Loss at step 539: 0.042198\n",
      "Loss at step 540: 0.042171\n",
      "Loss at step 541: 0.042145\n",
      "Loss at step 542: 0.042118\n",
      "Loss at step 543: 0.042088\n",
      "Loss at step 544: 0.042061\n",
      "Loss at step 545: 0.042035\n",
      "Loss at step 546: 0.042008\n",
      "Loss at step 547: 0.041977\n",
      "Loss at step 548: 0.041951\n",
      "Loss at step 549: 0.041926\n",
      "Loss at step 550: 0.041903\n",
      "Loss at step 551: 0.041873\n",
      "Loss at step 552: 0.041848\n",
      "Loss at step 553: 0.041825\n",
      "Loss at step 554: 0.041796\n",
      "Loss at step 555: 0.041771\n",
      "Loss at step 556: 0.041746\n",
      "Loss at step 557: 0.041722\n",
      "Loss at step 558: 0.041694\n",
      "Loss at step 559: 0.041669\n",
      "Loss at step 560: 0.041645\n",
      "Loss at step 561: 0.041621\n",
      "Loss at step 562: 0.041592\n",
      "Loss at step 563: 0.041568\n",
      "Loss at step 564: 0.041543\n",
      "Loss at step 565: 0.041519\n",
      "Loss at step 566: 0.041490\n",
      "Loss at step 567: 0.041466\n",
      "Loss at step 568: 0.041441\n",
      "Loss at step 569: 0.041418\n",
      "Loss at step 570: 0.041389\n",
      "Loss at step 571: 0.041365\n",
      "Loss at step 572: 0.041339\n",
      "Loss at step 573: 0.041316\n",
      "Loss at step 574: 0.041287\n",
      "Loss at step 575: 0.041263\n",
      "Loss at step 576: 0.041237\n",
      "Loss at step 577: 0.041214\n",
      "Loss at step 578: 0.041185\n",
      "Loss at step 579: 0.041161\n",
      "Loss at step 580: 0.041138\n",
      "Loss at step 581: 0.041117\n",
      "Loss at step 582: 0.041089\n",
      "Loss at step 583: 0.041069\n",
      "Loss at step 584: 0.041046\n",
      "Loss at step 585: 0.041026\n",
      "Loss at step 586: 0.040999\n",
      "Loss at step 587: 0.040976\n",
      "Loss at step 588: 0.040956\n",
      "Loss at step 589: 0.040933\n",
      "Loss at step 590: 0.040907\n",
      "Loss at step 591: 0.040887\n",
      "Loss at step 592: 0.040864\n",
      "Loss at step 593: 0.040843\n",
      "Loss at step 594: 0.040821\n",
      "Loss at step 595: 0.040794\n",
      "Loss at step 596: 0.040773\n",
      "Loss at step 597: 0.040750\n",
      "Loss at step 598: 0.040733\n",
      "Loss at step 599: 0.040703\n",
      "Loss at step 600: 0.040682\n",
      "Loss at step 601: 0.040661\n",
      "Loss at step 602: 0.040638\n",
      "Loss at step 603: 0.040613\n",
      "Loss at step 604: 0.040593\n",
      "Loss at step 605: 0.040570\n",
      "Loss at step 606: 0.040549\n",
      "Loss at step 607: 0.040529\n",
      "Loss at step 608: 0.040500\n",
      "Loss at step 609: 0.040480\n",
      "Loss at step 610: 0.040459\n",
      "Loss at step 611: 0.040435\n",
      "Loss at step 612: 0.040413\n",
      "Loss at step 613: 0.040390\n",
      "Loss at step 614: 0.040367\n",
      "Loss at step 615: 0.040346\n",
      "Loss at step 616: 0.040326\n",
      "Loss at step 617: 0.040304\n",
      "Loss at step 618: 0.040286\n",
      "Loss at step 619: 0.040268\n",
      "Loss at step 620: 0.040251\n",
      "Loss at step 621: 0.040229\n",
      "Loss at step 622: 0.040212\n",
      "Loss at step 623: 0.040193\n",
      "Loss at step 624: 0.040174\n",
      "Loss at step 625: 0.040154\n",
      "Loss at step 626: 0.040136\n",
      "Loss at step 627: 0.040117\n",
      "Loss at step 628: 0.040097\n",
      "Loss at step 629: 0.040076\n",
      "Loss at step 630: 0.040060\n",
      "Loss at step 631: 0.040040\n",
      "Loss at step 632: 0.040021\n",
      "Loss at step 633: 0.040005\n",
      "Loss at step 634: 0.039981\n",
      "Loss at step 635: 0.039963\n",
      "Loss at step 636: 0.039944\n",
      "Loss at step 637: 0.039926\n",
      "Loss at step 638: 0.039902\n",
      "Loss at step 639: 0.039886\n",
      "Loss at step 640: 0.039866\n",
      "Loss at step 641: 0.039848\n",
      "Loss at step 642: 0.039824\n",
      "Loss at step 643: 0.039806\n",
      "Loss at step 644: 0.039788\n",
      "Loss at step 645: 0.039769\n",
      "Loss at step 646: 0.039745\n",
      "Loss at step 647: 0.039728\n",
      "Loss at step 648: 0.039707\n",
      "Loss at step 649: 0.039689\n",
      "Loss at step 650: 0.039672\n",
      "Loss at step 651: 0.039647\n",
      "Loss at step 652: 0.039629\n",
      "Loss at step 653: 0.039609\n",
      "Loss at step 654: 0.039590\n",
      "Loss at step 655: 0.039569\n",
      "Loss at step 656: 0.039550\n",
      "Loss at step 657: 0.039533\n",
      "Loss at step 658: 0.039510\n",
      "Loss at step 659: 0.039493\n",
      "Loss at step 660: 0.039469\n",
      "Loss at step 661: 0.039452\n",
      "Loss at step 662: 0.039430\n",
      "Loss at step 663: 0.039412\n",
      "Loss at step 664: 0.039391\n",
      "Loss at step 665: 0.039371\n",
      "Loss at step 666: 0.039353\n",
      "Loss at step 667: 0.039331\n",
      "Loss at step 668: 0.039315\n",
      "Loss at step 669: 0.039296\n",
      "Loss at step 670: 0.039281\n",
      "Loss at step 671: 0.039266\n",
      "Loss at step 672: 0.039251\n",
      "Loss at step 673: 0.039232\n",
      "Loss at step 674: 0.039217\n",
      "Loss at step 675: 0.039202\n",
      "Loss at step 676: 0.039182\n",
      "Loss at step 677: 0.039168\n",
      "Loss at step 678: 0.039152\n",
      "Loss at step 679: 0.039138\n",
      "Loss at step 680: 0.039118\n",
      "Loss at step 681: 0.039103\n",
      "Loss at step 682: 0.039086\n",
      "Loss at step 683: 0.039071\n",
      "Loss at step 684: 0.039051\n",
      "Loss at step 685: 0.039037\n",
      "Loss at step 686: 0.039018\n",
      "Loss at step 687: 0.039006\n",
      "Loss at step 688: 0.038983\n",
      "Loss at step 689: 0.038967\n",
      "Loss at step 690: 0.038953\n",
      "Loss at step 691: 0.038937\n",
      "Loss at step 692: 0.038917\n",
      "Loss at step 693: 0.038899\n",
      "Loss at step 694: 0.038885\n",
      "Loss at step 695: 0.038871\n",
      "Loss at step 696: 0.038848\n",
      "Loss at step 697: 0.038831\n",
      "Loss at step 698: 0.038815\n",
      "Loss at step 699: 0.038796\n",
      "Loss at step 700: 0.038782\n",
      "Loss at step 701: 0.038763\n",
      "Loss at step 702: 0.038746\n",
      "Loss at step 703: 0.038734\n",
      "Loss at step 704: 0.038710\n",
      "Loss at step 705: 0.038694\n",
      "Loss at step 706: 0.038677\n",
      "Loss at step 707: 0.038663\n",
      "Loss at step 708: 0.038641\n",
      "Loss at step 709: 0.038624\n",
      "Loss at step 710: 0.038610\n",
      "Loss at step 711: 0.038594\n",
      "Loss at step 712: 0.038572\n",
      "Loss at step 713: 0.038558\n",
      "Loss at step 714: 0.038538\n",
      "Loss at step 715: 0.038526\n",
      "Loss at step 716: 0.038502\n",
      "Loss at step 717: 0.038487\n",
      "Loss at step 718: 0.038471\n",
      "Loss at step 719: 0.038453\n",
      "Loss at step 720: 0.038432\n",
      "Loss at step 721: 0.038418\n",
      "Loss at step 722: 0.038399\n",
      "Loss at step 723: 0.038385\n",
      "Loss at step 724: 0.038361\n",
      "Loss at step 725: 0.038346\n",
      "Loss at step 726: 0.038330\n",
      "Loss at step 727: 0.038313\n",
      "Loss at step 728: 0.038293\n",
      "Loss at step 729: 0.038277\n",
      "Loss at step 730: 0.038259\n",
      "Loss at step 731: 0.038244\n",
      "Loss at step 732: 0.038226\n",
      "Loss at step 733: 0.038204\n",
      "Loss at step 734: 0.038187\n",
      "Loss at step 735: 0.038172\n",
      "Loss at step 736: 0.038155\n",
      "Loss at step 737: 0.038132\n",
      "Loss at step 738: 0.038117\n",
      "Loss at step 739: 0.038101\n",
      "Loss at step 740: 0.038081\n",
      "Loss at step 741: 0.038068\n",
      "Loss at step 742: 0.038045\n",
      "Loss at step 743: 0.038030\n",
      "Loss at step 744: 0.038010\n",
      "Loss at step 745: 0.037994\n",
      "Loss at step 746: 0.037975\n",
      "Loss at step 747: 0.037956\n",
      "Loss at step 748: 0.037941\n",
      "Loss at step 749: 0.037920\n",
      "Loss at step 750: 0.037908\n",
      "Loss at step 751: 0.037885\n",
      "Loss at step 752: 0.037872\n",
      "Loss at step 753: 0.037858\n",
      "Loss at step 754: 0.037842\n",
      "Loss at step 755: 0.037822\n",
      "Loss at step 756: 0.037808\n",
      "Loss at step 757: 0.037791\n",
      "Loss at step 758: 0.037775\n",
      "Loss at step 759: 0.037757\n",
      "Loss at step 760: 0.037746\n",
      "Loss at step 761: 0.037729\n",
      "Loss at step 762: 0.037712\n",
      "Loss at step 763: 0.037694\n",
      "Loss at step 764: 0.037680\n",
      "Loss at step 765: 0.037662\n",
      "Loss at step 766: 0.037648\n",
      "Loss at step 767: 0.037633\n",
      "Loss at step 768: 0.037613\n",
      "Loss at step 769: 0.037598\n",
      "Loss at step 770: 0.037580\n",
      "Loss at step 771: 0.037568\n",
      "Loss at step 772: 0.037542\n",
      "Loss at step 773: 0.037527\n",
      "Loss at step 774: 0.037507\n",
      "Loss at step 775: 0.037491\n",
      "Loss at step 776: 0.037467\n",
      "Loss at step 777: 0.037452\n",
      "Loss at step 778: 0.037437\n",
      "Loss at step 779: 0.037415\n",
      "Loss at step 780: 0.037402\n",
      "Loss at step 781: 0.037375\n",
      "Loss at step 782: 0.037360\n",
      "Loss at step 783: 0.037342\n",
      "Loss at step 784: 0.037323\n",
      "Loss at step 785: 0.037306\n",
      "Loss at step 786: 0.037280\n",
      "Loss at step 787: 0.037265\n",
      "Loss at step 788: 0.037248\n",
      "Loss at step 789: 0.037227\n",
      "Loss at step 790: 0.037215\n",
      "Loss at step 791: 0.037187\n",
      "Loss at step 792: 0.037172\n",
      "Loss at step 793: 0.037151\n",
      "Loss at step 794: 0.037136\n",
      "Loss at step 795: 0.037110\n",
      "Loss at step 796: 0.037094\n",
      "Loss at step 797: 0.037075\n",
      "Loss at step 798: 0.037056\n",
      "Loss at step 799: 0.037038\n",
      "Loss at step 800: 0.037015\n",
      "Loss at step 801: 0.036996\n",
      "Loss at step 802: 0.036978\n",
      "Loss at step 803: 0.036958\n",
      "Loss at step 804: 0.036945\n",
      "Loss at step 805: 0.036917\n",
      "Loss at step 806: 0.036901\n",
      "Loss at step 807: 0.036883\n",
      "Loss at step 808: 0.036864\n",
      "Loss at step 809: 0.036838\n",
      "Loss at step 810: 0.036822\n",
      "Loss at step 811: 0.036803\n",
      "Loss at step 812: 0.036783\n",
      "Loss at step 813: 0.036768\n",
      "Loss at step 814: 0.036742\n",
      "Loss at step 815: 0.036726\n",
      "Loss at step 816: 0.036705\n",
      "Loss at step 817: 0.036688\n",
      "Loss at step 818: 0.036667\n",
      "Loss at step 819: 0.036646\n",
      "Loss at step 820: 0.036630\n",
      "Loss at step 821: 0.036607\n",
      "Loss at step 822: 0.036591\n",
      "Loss at step 823: 0.036566\n",
      "Loss at step 824: 0.036550\n",
      "Loss at step 825: 0.036528\n",
      "Loss at step 826: 0.036511\n",
      "Loss at step 827: 0.036491\n",
      "Loss at step 828: 0.036475\n",
      "Loss at step 829: 0.036449\n",
      "Loss at step 830: 0.036431\n",
      "Loss at step 831: 0.036412\n",
      "Loss at step 832: 0.036391\n",
      "Loss at step 833: 0.036368\n",
      "Loss at step 834: 0.036353\n",
      "Loss at step 835: 0.036333\n",
      "Loss at step 836: 0.036314\n",
      "Loss at step 837: 0.036294\n",
      "Loss at step 838: 0.036270\n",
      "Loss at step 839: 0.036252\n",
      "Loss at step 840: 0.036236\n",
      "Loss at step 841: 0.036214\n",
      "Loss at step 842: 0.036197\n",
      "Loss at step 843: 0.036173\n",
      "Loss at step 844: 0.036154\n",
      "Loss at step 845: 0.036138\n",
      "Loss at step 846: 0.036115\n",
      "Loss at step 847: 0.036100\n",
      "Loss at step 848: 0.036072\n",
      "Loss at step 849: 0.036056\n",
      "Loss at step 850: 0.036037\n",
      "Loss at step 851: 0.036017\n",
      "Loss at step 852: 0.035999\n",
      "Loss at step 853: 0.035975\n",
      "Loss at step 854: 0.035959\n",
      "Loss at step 855: 0.035939\n",
      "Loss at step 856: 0.035918\n",
      "Loss at step 857: 0.035899\n",
      "Loss at step 858: 0.035881\n",
      "Loss at step 859: 0.035859\n",
      "Loss at step 860: 0.035841\n",
      "Loss at step 861: 0.035821\n",
      "Loss at step 862: 0.035800\n",
      "Loss at step 863: 0.035782\n",
      "Loss at step 864: 0.035758\n",
      "Loss at step 865: 0.035742\n",
      "Loss at step 866: 0.035720\n",
      "Loss at step 867: 0.035701\n",
      "Loss at step 868: 0.035680\n",
      "Loss at step 869: 0.035658\n",
      "Loss at step 870: 0.035640\n",
      "Loss at step 871: 0.035621\n",
      "Loss at step 872: 0.035602\n",
      "Loss at step 873: 0.035580\n",
      "Loss at step 874: 0.035566\n",
      "Loss at step 875: 0.035541\n",
      "Loss at step 876: 0.035525\n",
      "Loss at step 877: 0.035502\n",
      "Loss at step 878: 0.035485\n",
      "Loss at step 879: 0.035463\n",
      "Loss at step 880: 0.035441\n",
      "Loss at step 881: 0.035423\n",
      "Loss at step 882: 0.035404\n",
      "Loss at step 883: 0.035387\n",
      "Loss at step 884: 0.035364\n",
      "Loss at step 885: 0.035349\n",
      "Loss at step 886: 0.035322\n",
      "Loss at step 887: 0.035305\n",
      "Loss at step 888: 0.035287\n",
      "Loss at step 889: 0.035266\n",
      "Loss at step 890: 0.035251\n",
      "Loss at step 891: 0.035223\n",
      "Loss at step 892: 0.035207\n",
      "Loss at step 893: 0.035188\n",
      "Loss at step 894: 0.035171\n",
      "Loss at step 895: 0.035149\n",
      "Loss at step 896: 0.035135\n",
      "Loss at step 897: 0.035107\n",
      "Loss at step 898: 0.035090\n",
      "Loss at step 899: 0.035073\n",
      "Loss at step 900: 0.035052\n",
      "Loss at step 901: 0.035036\n",
      "Loss at step 902: 0.035010\n",
      "Loss at step 903: 0.034994\n",
      "Loss at step 904: 0.034976\n",
      "Loss at step 905: 0.034954\n",
      "Loss at step 906: 0.034935\n",
      "Loss at step 907: 0.034916\n",
      "Loss at step 908: 0.034896\n",
      "Loss at step 909: 0.034877\n",
      "Loss at step 910: 0.034861\n",
      "Loss at step 911: 0.034837\n",
      "Loss at step 912: 0.034820\n",
      "Loss at step 913: 0.034799\n",
      "Loss at step 914: 0.034778\n",
      "Loss at step 915: 0.034762\n",
      "Loss at step 916: 0.034742\n",
      "Loss at step 917: 0.034722\n",
      "Loss at step 918: 0.034702\n",
      "Loss at step 919: 0.034686\n",
      "Loss at step 920: 0.034661\n",
      "Loss at step 921: 0.034643\n",
      "Loss at step 922: 0.034626\n",
      "Loss at step 923: 0.034606\n",
      "Loss at step 924: 0.034586\n",
      "Loss at step 925: 0.034563\n",
      "Loss at step 926: 0.034546\n",
      "Loss at step 927: 0.034530\n",
      "Loss at step 928: 0.034507\n",
      "Loss at step 929: 0.034490\n",
      "Loss at step 930: 0.034475\n",
      "Loss at step 931: 0.034448\n",
      "Loss at step 932: 0.034432\n",
      "Loss at step 933: 0.034411\n",
      "Loss at step 934: 0.034395\n",
      "Loss at step 935: 0.034375\n",
      "Loss at step 936: 0.034355\n",
      "Loss at step 937: 0.034334\n",
      "Loss at step 938: 0.034315\n",
      "Loss at step 939: 0.034298\n",
      "Loss at step 940: 0.034281\n",
      "Loss at step 941: 0.034258\n",
      "Loss at step 942: 0.034244\n",
      "Loss at step 943: 0.034217\n",
      "Loss at step 944: 0.034200\n",
      "Loss at step 945: 0.034183\n",
      "Loss at step 946: 0.034160\n",
      "Loss at step 947: 0.034144\n",
      "Loss at step 948: 0.034126\n",
      "Loss at step 949: 0.034100\n",
      "Loss at step 950: 0.034082\n",
      "Loss at step 951: 0.034065\n",
      "Loss at step 952: 0.034047\n",
      "Loss at step 953: 0.034026\n",
      "Loss at step 954: 0.034009\n",
      "Loss at step 955: 0.033986\n",
      "Loss at step 956: 0.033967\n",
      "Loss at step 957: 0.033950\n",
      "Loss at step 958: 0.033930\n",
      "Loss at step 959: 0.033910\n",
      "Loss at step 960: 0.033887\n",
      "Loss at step 961: 0.033870\n",
      "Loss at step 962: 0.033855\n",
      "Loss at step 963: 0.033832\n",
      "Loss at step 964: 0.033812\n",
      "Loss at step 965: 0.033794\n",
      "Loss at step 966: 0.033770\n",
      "Loss at step 967: 0.033752\n",
      "Loss at step 968: 0.033733\n",
      "Loss at step 969: 0.033715\n",
      "Loss at step 970: 0.033694\n",
      "Loss at step 971: 0.033678\n",
      "Loss at step 972: 0.033652\n",
      "Loss at step 973: 0.033635\n",
      "Loss at step 974: 0.033619\n",
      "Loss at step 975: 0.033599\n",
      "Loss at step 976: 0.033581\n",
      "Loss at step 977: 0.033566\n",
      "Loss at step 978: 0.033537\n",
      "Loss at step 979: 0.033519\n",
      "Loss at step 980: 0.033502\n",
      "Loss at step 981: 0.033486\n",
      "Loss at step 982: 0.033467\n",
      "Loss at step 983: 0.033453\n",
      "Loss at step 984: 0.033431\n",
      "Loss at step 985: 0.033416\n",
      "Loss at step 986: 0.033403\n",
      "Loss at step 987: 0.033384\n",
      "Loss at step 988: 0.033367\n",
      "Loss at step 989: 0.033346\n",
      "Loss at step 990: 0.033331\n",
      "Loss at step 991: 0.033318\n",
      "Loss at step 992: 0.033298\n",
      "Loss at step 993: 0.033282\n",
      "Loss at step 994: 0.033267\n",
      "Loss at step 995: 0.033245\n",
      "Loss at step 996: 0.033231\n",
      "Loss at step 997: 0.033213\n",
      "Loss at step 998: 0.033197\n",
      "Loss at step 999: 0.033179\n",
      "Loss at step 1000: 0.033160\n"
     ]
    }
   ],
   "source": [
    "nn.train(X_train, y_train, learning_rate=0.05, epochs=1000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "preds = nn.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "binary_preds = (preds >= 0.5).astype(int)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "correct_predictions = (binary_preds == y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9108695652173913"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracy = np.sum(correct_predictions) / len(y_test)\n",
    "accuracy"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "venv",
   "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": "3.9.16"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}