gpu_test.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.layers import Dense, Input, Dropout\n",
    "from keras.models import Model\n",
    "from keras.regularizers import l2\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "N = 100000\n",
    "\n",
    "x = np.random.uniform(-10, 10, size=N)\n",
    "y = np.sin(x)*np.sin(3*x) + np.random.normal(size=N)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f9f87dfe358>"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pd.DataFrame({'x': x, 'y': y}).plot(x='x', y='y', kind='scatter', alpha=0.01)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/500\n",
      "100000/100000 [==============================] - 4s 43us/step - loss: 354.1005 - mean_squared_error: 354.1005\n",
      "Epoch 2/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 3.3142 - mean_squared_error: 3.3142\n",
      "Epoch 3/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 2.9989 - mean_squared_error: 2.9989\n",
      "Epoch 4/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 2.0170 - mean_squared_error: 2.0170\n",
      "Epoch 5/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.3250 - mean_squared_error: 1.3250\n",
      "Epoch 6/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.2343 - mean_squared_error: 1.2343\n",
      "Epoch 7/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.2510 - mean_squared_error: 1.2510\n",
      "Epoch 8/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.2030 - mean_squared_error: 1.2030\n",
      "Epoch 9/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1934 - mean_squared_error: 1.1934\n",
      "Epoch 10/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1863 - mean_squared_error: 1.1863\n",
      "Epoch 11/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1823 - mean_squared_error: 1.1823\n",
      "Epoch 12/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1803 - mean_squared_error: 1.1803\n",
      "Epoch 13/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.1747 - mean_squared_error: 1.1747\n",
      "Epoch 14/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1715 - mean_squared_error: 1.1715\n",
      "Epoch 15/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.1709 - mean_squared_error: 1.1709\n",
      "Epoch 16/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1668 - mean_squared_error: 1.1668\n",
      "Epoch 17/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1657 - mean_squared_error: 1.1657\n",
      "Epoch 18/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1659 - mean_squared_error: 1.1659\n",
      "Epoch 19/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.1652 - mean_squared_error: 1.1652\n",
      "Epoch 20/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1638 - mean_squared_error: 1.1638\n",
      "Epoch 21/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1631 - mean_squared_error: 1.1631\n",
      "Epoch 22/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.1612 - mean_squared_error: 1.1612\n",
      "Epoch 23/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1606 - mean_squared_error: 1.1606\n",
      "Epoch 24/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1604 - mean_squared_error: 1.1604\n",
      "Epoch 25/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.1598 - mean_squared_error: 1.1598\n",
      "Epoch 26/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1557 - mean_squared_error: 1.1557\n",
      "Epoch 27/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1585 - mean_squared_error: 1.1585\n",
      "Epoch 28/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.1545 - mean_squared_error: 1.1545\n",
      "Epoch 29/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1518 - mean_squared_error: 1.1518\n",
      "Epoch 30/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1501 - mean_squared_error: 1.1501\n",
      "Epoch 31/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1486 - mean_squared_error: 1.1486\n",
      "Epoch 32/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1489 - mean_squared_error: 1.1489\n",
      "Epoch 33/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1427 - mean_squared_error: 1.1427\n",
      "Epoch 34/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1398 - mean_squared_error: 1.1398\n",
      "Epoch 35/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1418 - mean_squared_error: 1.1418\n",
      "Epoch 36/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1366 - mean_squared_error: 1.1366\n",
      "Epoch 37/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1371 - mean_squared_error: 1.1371\n",
      "Epoch 38/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1300 - mean_squared_error: 1.1300\n",
      "Epoch 39/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1318 - mean_squared_error: 1.1318\n",
      "Epoch 40/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1287 - mean_squared_error: 1.1287\n",
      "Epoch 41/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1290 - mean_squared_error: 1.1290\n",
      "Epoch 42/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1228 - mean_squared_error: 1.1228\n",
      "Epoch 43/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1279 - mean_squared_error: 1.1279\n",
      "Epoch 44/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1251 - mean_squared_error: 1.1251\n",
      "Epoch 45/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1193 - mean_squared_error: 1.1193\n",
      "Epoch 46/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1187 - mean_squared_error: 1.1187\n",
      "Epoch 47/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1185 - mean_squared_error: 1.1185\n",
      "Epoch 48/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1185 - mean_squared_error: 1.1185\n",
      "Epoch 49/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1155 - mean_squared_error: 1.1155\n",
      "Epoch 50/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1148 - mean_squared_error: 1.1148\n",
      "Epoch 51/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1163 - mean_squared_error: 1.1163\n",
      "Epoch 52/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1111 - mean_squared_error: 1.1111\n",
      "Epoch 53/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1134 - mean_squared_error: 1.1134\n",
      "Epoch 54/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1104 - mean_squared_error: 1.1104\n",
      "Epoch 55/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1098 - mean_squared_error: 1.1098\n",
      "Epoch 56/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1084 - mean_squared_error: 1.1084\n",
      "Epoch 57/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.1057 - mean_squared_error: 1.1057\n",
      "Epoch 58/500\n",
      "100000/100000 [==============================] - 3s 28us/step - loss: 1.1090 - mean_squared_error: 1.1090\n",
      "Epoch 59/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1079 - mean_squared_error: 1.1079\n",
      "Epoch 60/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1024 - mean_squared_error: 1.1024\n",
      "Epoch 61/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.1028 - mean_squared_error: 1.1028\n",
      "Epoch 62/500\n",
      "100000/100000 [==============================] - 3s 29us/step - loss: 1.1066 - mean_squared_error: 1.1066\n",
      "Epoch 63/500\n",
      "100000/100000 [==============================] - 3s 29us/step - loss: 1.1060 - mean_squared_error: 1.1060\n",
      "Epoch 64/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1005 - mean_squared_error: 1.1005\n",
      "Epoch 65/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.1021 - mean_squared_error: 1.1021\n",
      "Epoch 66/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.1006 - mean_squared_error: 1.1006\n",
      "Epoch 67/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.1039 - mean_squared_error: 1.1039\n",
      "Epoch 68/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0973 - mean_squared_error: 1.0973\n",
      "Epoch 69/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0975 - mean_squared_error: 1.0975\n",
      "Epoch 70/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0961 - mean_squared_error: 1.0961\n",
      "Epoch 71/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0939 - mean_squared_error: 1.0939\n",
      "Epoch 72/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0974 - mean_squared_error: 1.0974\n",
      "Epoch 73/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0969 - mean_squared_error: 1.0969\n",
      "Epoch 74/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0984 - mean_squared_error: 1.0984\n",
      "Epoch 75/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0944 - mean_squared_error: 1.0944\n",
      "Epoch 76/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0951 - mean_squared_error: 1.0951\n",
      "Epoch 77/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0868 - mean_squared_error: 1.0868\n",
      "Epoch 78/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0954 - mean_squared_error: 1.0954\n",
      "Epoch 79/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0897 - mean_squared_error: 1.0897\n",
      "Epoch 80/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0926 - mean_squared_error: 1.0926\n",
      "Epoch 81/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0913 - mean_squared_error: 1.0913\n",
      "Epoch 82/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0896 - mean_squared_error: 1.0896\n",
      "Epoch 83/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0901 - mean_squared_error: 1.0901\n",
      "Epoch 84/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0878 - mean_squared_error: 1.0878\n",
      "Epoch 85/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0914 - mean_squared_error: 1.0914\n",
      "Epoch 86/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0908 - mean_squared_error: 1.0908\n",
      "Epoch 87/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0836 - mean_squared_error: 1.0836\n",
      "Epoch 88/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0882 - mean_squared_error: 1.0882\n",
      "Epoch 89/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0820 - mean_squared_error: 1.0820\n",
      "Epoch 90/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0875 - mean_squared_error: 1.0875\n",
      "Epoch 91/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0868 - mean_squared_error: 1.0868\n",
      "Epoch 92/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0849 - mean_squared_error: 1.0849\n",
      "Epoch 93/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0850 - mean_squared_error: 1.0850\n",
      "Epoch 94/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0838 - mean_squared_error: 1.0838\n",
      "Epoch 95/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0851 - mean_squared_error: 1.0851\n",
      "Epoch 96/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0765 - mean_squared_error: 1.0765\n",
      "Epoch 97/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0843 - mean_squared_error: 1.0843\n",
      "Epoch 98/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0792 - mean_squared_error: 1.0792\n",
      "Epoch 99/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0828 - mean_squared_error: 1.0828\n",
      "Epoch 100/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0841 - mean_squared_error: 1.0841\n",
      "Epoch 101/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0789 - mean_squared_error: 1.0789\n",
      "Epoch 102/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0791 - mean_squared_error: 1.0791\n",
      "Epoch 103/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0800 - mean_squared_error: 1.0800\n",
      "Epoch 104/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0778 - mean_squared_error: 1.0778\n",
      "Epoch 105/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0742 - mean_squared_error: 1.0742\n",
      "Epoch 106/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0784 - mean_squared_error: 1.0784\n",
      "Epoch 107/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0755 - mean_squared_error: 1.0755\n",
      "Epoch 108/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0802 - mean_squared_error: 1.0802\n",
      "Epoch 109/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0745 - mean_squared_error: 1.0745\n",
      "Epoch 110/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0731 - mean_squared_error: 1.0731\n",
      "Epoch 111/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0760 - mean_squared_error: 1.0760\n",
      "Epoch 112/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0780 - mean_squared_error: 1.0780\n",
      "Epoch 113/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0746 - mean_squared_error: 1.0746\n",
      "Epoch 114/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0728 - mean_squared_error: 1.0728\n",
      "Epoch 115/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0682 - mean_squared_error: 1.0682\n",
      "Epoch 116/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0734 - mean_squared_error: 1.0734\n",
      "Epoch 117/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0671 - mean_squared_error: 1.0671\n",
      "Epoch 118/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0781 - mean_squared_error: 1.0781\n",
      "Epoch 119/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0705 - mean_squared_error: 1.0705\n",
      "Epoch 120/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0692 - mean_squared_error: 1.0692\n",
      "Epoch 121/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0729 - mean_squared_error: 1.0729\n",
      "Epoch 122/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0676 - mean_squared_error: 1.0676\n",
      "Epoch 123/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0697 - mean_squared_error: 1.0697\n",
      "Epoch 124/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0683 - mean_squared_error: 1.0683\n",
      "Epoch 125/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0680 - mean_squared_error: 1.0680\n",
      "Epoch 126/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0715 - mean_squared_error: 1.0715\n",
      "Epoch 127/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0651 - mean_squared_error: 1.0651\n",
      "Epoch 128/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0691 - mean_squared_error: 1.0691\n",
      "Epoch 129/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0659 - mean_squared_error: 1.0659\n",
      "Epoch 130/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0696 - mean_squared_error: 1.0696\n",
      "Epoch 131/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0636 - mean_squared_error: 1.0636\n",
      "Epoch 132/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0666 - mean_squared_error: 1.0666\n",
      "Epoch 133/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0670 - mean_squared_error: 1.0670\n",
      "Epoch 134/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0646 - mean_squared_error: 1.0646\n",
      "Epoch 135/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0629 - mean_squared_error: 1.0629\n",
      "Epoch 136/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0647 - mean_squared_error: 1.0647\n",
      "Epoch 137/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0613 - mean_squared_error: 1.0613\n",
      "Epoch 138/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0640 - mean_squared_error: 1.0640\n",
      "Epoch 139/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0640 - mean_squared_error: 1.0640\n",
      "Epoch 140/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0586 - mean_squared_error: 1.0586\n",
      "Epoch 141/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0637 - mean_squared_error: 1.0637\n",
      "Epoch 142/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0628 - mean_squared_error: 1.0628\n",
      "Epoch 143/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0613 - mean_squared_error: 1.0613\n",
      "Epoch 144/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0608 - mean_squared_error: 1.0608\n",
      "Epoch 145/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0632 - mean_squared_error: 1.0632\n",
      "Epoch 146/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0563 - mean_squared_error: 1.0563\n",
      "Epoch 147/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0586 - mean_squared_error: 1.0586\n",
      "Epoch 148/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0586 - mean_squared_error: 1.0586\n",
      "Epoch 149/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0578 - mean_squared_error: 1.0578\n",
      "Epoch 150/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0597 - mean_squared_error: 1.0597\n",
      "Epoch 151/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0615 - mean_squared_error: 1.0615\n",
      "Epoch 152/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0594 - mean_squared_error: 1.0594\n",
      "Epoch 153/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0573 - mean_squared_error: 1.0573\n",
      "Epoch 154/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0554 - mean_squared_error: 1.0554\n",
      "Epoch 155/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0572 - mean_squared_error: 1.0572\n",
      "Epoch 156/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0582 - mean_squared_error: 1.0582\n",
      "Epoch 157/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0517 - mean_squared_error: 1.0517\n",
      "Epoch 158/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0613 - mean_squared_error: 1.0613\n",
      "Epoch 159/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0560 - mean_squared_error: 1.0560\n",
      "Epoch 160/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0555 - mean_squared_error: 1.0555\n",
      "Epoch 161/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0526 - mean_squared_error: 1.0526\n",
      "Epoch 162/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0583 - mean_squared_error: 1.0583\n",
      "Epoch 163/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0539 - mean_squared_error: 1.0539\n",
      "Epoch 164/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0518 - mean_squared_error: 1.0518\n",
      "Epoch 165/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0543 - mean_squared_error: 1.0543\n",
      "Epoch 166/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0521 - mean_squared_error: 1.0521\n",
      "Epoch 167/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0554 - mean_squared_error: 1.0554\n",
      "Epoch 168/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0521 - mean_squared_error: 1.0521\n",
      "Epoch 169/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0562 - mean_squared_error: 1.0562\n",
      "Epoch 170/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0539 - mean_squared_error: 1.0539\n",
      "Epoch 171/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0512 - mean_squared_error: 1.0512\n",
      "Epoch 172/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0511 - mean_squared_error: 1.0511\n",
      "Epoch 173/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0515 - mean_squared_error: 1.0515\n",
      "Epoch 174/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0524 - mean_squared_error: 1.0524\n",
      "Epoch 175/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0534 - mean_squared_error: 1.0534\n",
      "Epoch 176/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0471 - mean_squared_error: 1.0471\n",
      "Epoch 177/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0518 - mean_squared_error: 1.0518\n",
      "Epoch 178/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0513 - mean_squared_error: 1.0513\n",
      "Epoch 179/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0513 - mean_squared_error: 1.0513\n",
      "Epoch 180/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0503 - mean_squared_error: 1.0503\n",
      "Epoch 181/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0505 - mean_squared_error: 1.0505\n",
      "Epoch 182/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0495 - mean_squared_error: 1.0495\n",
      "Epoch 183/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0483 - mean_squared_error: 1.0483\n",
      "Epoch 184/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0538 - mean_squared_error: 1.0538\n",
      "Epoch 185/500\n",
      "100000/100000 [==============================] - 3s 25us/step - loss: 1.0504 - mean_squared_error: 1.0504\n",
      "Epoch 186/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0456 - mean_squared_error: 1.0456\n",
      "Epoch 187/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0482 - mean_squared_error: 1.0482\n",
      "Epoch 188/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0506 - mean_squared_error: 1.0506\n",
      "Epoch 189/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0458 - mean_squared_error: 1.0458\n",
      "Epoch 190/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0490 - mean_squared_error: 1.0490\n",
      "Epoch 191/500\n",
      "100000/100000 [==============================] - 3s 28us/step - loss: 1.0505 - mean_squared_error: 1.0505\n",
      "Epoch 192/500\n",
      "100000/100000 [==============================] - 3s 29us/step - loss: 1.0459 - mean_squared_error: 1.0459\n",
      "Epoch 193/500\n",
      "100000/100000 [==============================] - 3s 28us/step - loss: 1.0474 - mean_squared_error: 1.0474\n",
      "Epoch 194/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0477 - mean_squared_error: 1.0477\n",
      "Epoch 195/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0463 - mean_squared_error: 1.0463\n",
      "Epoch 196/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0431 - mean_squared_error: 1.0431\n",
      "Epoch 197/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0451 - mean_squared_error: 1.0451\n",
      "Epoch 198/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0457 - mean_squared_error: 1.0457\n",
      "Epoch 199/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0505 - mean_squared_error: 1.0505\n",
      "Epoch 200/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0448 - mean_squared_error: 1.0448\n",
      "Epoch 201/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0465 - mean_squared_error: 1.0465\n",
      "Epoch 202/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0475 - mean_squared_error: 1.0475\n",
      "Epoch 203/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0481 - mean_squared_error: 1.0481\n",
      "Epoch 204/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0454 - mean_squared_error: 1.0454\n",
      "Epoch 205/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0461 - mean_squared_error: 1.0461\n",
      "Epoch 206/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0459 - mean_squared_error: 1.0459\n",
      "Epoch 207/500\n",
      "100000/100000 [==============================] - 3s 26us/step - loss: 1.0460 - mean_squared_error: 1.0460\n",
      "Epoch 208/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0428 - mean_squared_error: 1.0428\n",
      "Epoch 209/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0470 - mean_squared_error: 1.0470\n",
      "Epoch 210/500\n",
      "100000/100000 [==============================] - 3s 27us/step - loss: 1.0468 - mean_squared_error: 1.0468\n",
      "Epoch 211/500\n",
      " 57344/100000 [================>.............] - ETA: 1s - loss: 1.0425 - mean_squared_error: 1.0425"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-45-1bc8510a9011>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     17\u001b[0m \u001b[0mmodel\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mModel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mlayers\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0moutputs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mlayers\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     18\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcompile\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'RMSprop'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mloss\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'mean_squared_error'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmetrics\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'mean_squared_error'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 19\u001b[0;31m \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mepochs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m500\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbatch_size\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m4096\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m/home/akelleh/.local/lib/python3.6/site-packages/keras/engine/training.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, **kwargs)\u001b[0m\n\u001b[1;32m   1035\u001b[0m                                         \u001b[0minitial_epoch\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0minitial_epoch\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1036\u001b[0m                                         \u001b[0msteps_per_epoch\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0msteps_per_epoch\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1037\u001b[0;31m                                         validation_steps=validation_steps)\n\u001b[0m\u001b[1;32m   1038\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1039\u001b[0m     def evaluate(self, x=None, y=None,\n",
      "\u001b[0;32m/home/akelleh/.local/lib/python3.6/site-packages/keras/engine/training_arrays.py\u001b[0m in \u001b[0;36mfit_loop\u001b[0;34m(model, f, ins, out_labels, batch_size, epochs, verbose, callbacks, val_f, val_ins, shuffle, callback_metrics, initial_epoch, steps_per_epoch, validation_steps)\u001b[0m\n\u001b[1;32m    197\u001b[0m                     \u001b[0mins_batch\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mins_batch\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtoarray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    198\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 199\u001b[0;31m                 \u001b[0mouts\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mf\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mins_batch\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    200\u001b[0m                 \u001b[0mouts\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mto_list\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mouts\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    201\u001b[0m                 \u001b[0;32mfor\u001b[0m \u001b[0ml\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mo\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mzip\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mout_labels\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mouts\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/home/akelleh/.local/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, inputs)\u001b[0m\n\u001b[1;32m   2664\u001b[0m                 \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_legacy_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2665\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2666\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2667\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2668\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mpy_any\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mis_tensor\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0minputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/home/akelleh/.local/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py\u001b[0m in \u001b[0;36m_call\u001b[0;34m(self, inputs)\u001b[0m\n\u001b[1;32m   2634\u001b[0m                                 \u001b[0msymbol_vals\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2635\u001b[0m                                 session)\n\u001b[0;32m-> 2636\u001b[0;31m         \u001b[0mfetched\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_callable_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0marray_vals\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2637\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mfetched\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0moutputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2638\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/home/akelleh/.local/lib/python3.6/site-packages/tensorflow/python/client/session.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   1380\u001b[0m           ret = tf_session.TF_SessionRunCallable(\n\u001b[1;32m   1381\u001b[0m               \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_session\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_session\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_handle\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstatus\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1382\u001b[0;31m               run_metadata_ptr)\n\u001b[0m\u001b[1;32m   1383\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mrun_metadata\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1384\u001b[0m           \u001b[0mproto_data\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtf_session\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mTF_GetBuffer\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrun_metadata_ptr\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "layer_size = 4096\n",
    "num_hidden_layers = 3\n",
    "activation = 'relu'\n",
    "dropout = 0.05\n",
    "alpha_k = 0.#0001\n",
    "alpha_b = 0.#001\n",
    "\n",
    "layers = [Input(shape=(1,))]\n",
    "for _ in range(num_hidden_layers):\n",
    "    layers.append(Dense(layer_size, \n",
    "                        activation=activation, \n",
    "                        kernel_regularizer=l2(alpha_k), \n",
    "                        bias_regularizer=l2(alpha_b))(layers[-1]))\n",
    "    layers.append(Dropout(dropout)(layers[-1]))\n",
    "layers.append(Dense(1, activation='linear')(layers[-1]))\n",
    "\n",
    "model = Model(inputs=layers[0], outputs=layers[-1])\n",
    "model.compile('RMSprop', loss='mean_squared_error', metrics=['mean_squared_error'])\n",
    "model.fit(x, y, epochs=500, batch_size=4096)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f9f87f02ef0>"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "x_pred = np.random.uniform(-10, 10, size=10000)\n",
    "\n",
    "y_pred = model.predict(x_pred).T[0]\n",
    "y_true = np.sin(x_pred)*np.sin(3*x_pred)\n",
    "\n",
    "pd.DataFrame({'x': x_pred, 'y': y_pred}).plot(x='x', y='y', kind='scatter')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_25 (InputLayer)        (None, 1)                 0         \n",
      "_________________________________________________________________\n",
      "dense_174 (Dense)            (None, 2048)              4096      \n",
      "_________________________________________________________________\n",
      "dropout_128 (Dropout)        (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "dense_175 (Dense)            (None, 2048)              4196352   \n",
      "_________________________________________________________________\n",
      "dropout_129 (Dropout)        (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "dense_176 (Dense)            (None, 2048)              4196352   \n",
      "_________________________________________________________________\n",
      "dropout_130 (Dropout)        (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "dense_177 (Dense)            (None, 2048)              4196352   \n",
      "_________________________________________________________________\n",
      "dropout_131 (Dropout)        (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "dense_178 (Dense)            (None, 2048)              4196352   \n",
      "_________________________________________________________________\n",
      "dropout_132 (Dropout)        (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "dense_179 (Dense)            (None, 2048)              4196352   \n",
      "_________________________________________________________________\n",
      "dropout_133 (Dropout)        (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "dense_180 (Dense)            (None, 2048)              4196352   \n",
      "_________________________________________________________________\n",
      "dropout_134 (Dropout)        (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "dense_181 (Dense)            (None, 2048)              4196352   \n",
      "_________________________________________________________________\n",
      "dropout_135 (Dropout)        (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "dense_182 (Dense)            (None, 1)                 2049      \n",
      "=================================================================\n",
      "Total params: 29,380,609\n",
      "Trainable params: 29,380,609\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "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": [
    "#from keras import backend as K\n",
    "K.tensorflow_backend._get_available_gpus()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "GPU = True\n",
    "CPU = True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from keras import backend as K\n",
    "\n",
    "num_cores = 8\n",
    "\n",
    "if GPU:\n",
    "    num_GPU = 1\n",
    "    num_CPU = 1\n",
    "if CPU:\n",
    "    num_CPU = 1\n",
    "    num_GPU = 0\n",
    "\n",
    "config = tf.ConfigProto(intra_op_parallelism_threads=num_cores,\\\n",
    "        inter_op_parallelism_threads=num_cores, allow_soft_placement=True,\\\n",
    "        device_count = {'CPU' : num_CPU, 'GPU' : num_GPU})\n",
    "session = tf.Session(config=config)\n",
    "K.set_session(session)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}