Lab02_keras.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Lab02_keras.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/jeonggunlee/DeepLearningBasics/blob/master/Lab02_keras.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yzPDC28jN3Ub",
        "colab_type": "text"
      },
      "source": [
        "\n",
        "##누구나 즐기는 딥러닝: 오픈소스 Keras를 활용하여!!!\n",
        "---\n",
        "이정근 교수\n",
        "\n",
        "소프트웨어융합대학\n",
        "\n",
        "jeonggun.lee@hallym.ac.kr\n",
        "2020년 8월\n",
        "\n",
        "---\n",
        "###단순 선형 회귀\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "x19kceaPZlMN",
        "colab_type": "text"
      },
      "source": [
        "## 단순 선형 회귀\n",
        "\n",
        "**outcome = data * 2 + 1**에 대한 식의 패턴을 학습방식을 통하여 획득하는 방법을 Keras를 활용하여 공부해보도록 합시다!\n",
        "\n",
        "입력데이터:\n",
        "data = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]\n",
        "\n",
        "출력데이터:\n",
        "outcome = [3, 5, 7, 9, 11, 13, 15, 17, 19, 21]\n",
        "\n",
        "위와 같이 주어진 데이터를 이용하여 입력과 출력에 대한 패턴(알고리즘)을 찾아낼 수 있을까요 ?"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_N2AKFIiNohu",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        },
        "outputId": "86a5d7d8-cdfa-4c68-d3a2-9c23b99aa053"
      },
      "source": [
        "import numpy as np\n",
        "\n",
        "data = np.array([1, 2, 3, 4, 5, 6, 7, 8, 9, 10])\n",
        "outcome = np.array([3, 5, 7, 9, 11, 13, 15, 17, 19, 21])\n",
        "print(data)\n",
        "print(outcome)"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[ 1  2  3  4  5  6  7  8  9 10]\n",
            "[ 3  5  7  9 11 13 15 17 19 21]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "H6ReFgRlCnsp",
        "colab_type": "text"
      },
      "source": [
        "입력데이터와 출력 outcome과의 관계를 그래프를 통해서 살펴볼까요 ?\n",
        "\n",
        "그래프 사용을 위해서 matplotlib를 활용하여 봅시다!"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "X6lVnn7qZgZf",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "outputId": "5caa642c-8885-41b3-e2ce-66503b3762d7"
      },
      "source": [
        "import matplotlib.pylab as plt\n",
        "plt.plot(data, outcome, 'rs--')\n",
        "plt.show()"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAD4CAYAAADiry33AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAceklEQVR4nO3de5RU5ZX38e/mZgQhgLbdICBJRA2CgDaIeAmCEEQMaLxgvKCSQRMd9TUux8skvKMZlzPGC94g3ARFMU4QYQKoyKshRpnYIAqCM6CigDTdiIKiUaH3+8dTPTRtFTRd1XWqTv0+a9XqqqdO1dmWi1+ffs6p/Zi7IyIi8dUo6gJERKRhKehFRGJOQS8iEnMKehGRmFPQi4jEXJOoC0jmkEMO8c6dO0ddhohI3li6dOkWdy9K9lxOBn3nzp0pKyuLugwRkbxhZh+kek5TNyIiMaegFxGJOQW9iEjMKehFRGJOQS8iEnMKehGRKJWUgNm3byUlGduFgl5EJEqbN+/feD0o6EVEYk5BLyIScwp6EZGoLFiQld0o6EVEojBzJgwdmpVdKehFRLLFHbZsCfeHD4e774bi4uTbphqvBwW9iEg2bNoE55wDffvCF19A8+Zw441QXh5+AdS+lZdnbNcKehGRhuQOU6fCD38Izz0HV14JzZpltYScbFMsIhILn3wC558PL74Ip54KkybBkUdmvQwd0YuINJRWraBJExg/Hl56KZKQhzoEvZl1NLOXzGyVmb1tZtclxtua2UIzW5P42SbF60cltlljZqMy/R8gIpJTVq2CYcOgogIaN4b58+Gqq6BRdMfVddnzTuBX7t4V6AtcbWZdgZuBRe7eBViUeLwHM2sLjAVOAPoAY1P9QhARyWtffw2//S306gVLlsA774Rxs2jrog5B7+6b3H1Z4v5nwGrgMGA4MD2x2XRgRJKX/xhY6O5b3f0TYCEwJBOFi4jkjLIy6N0bfv3rcGXNqlVhTj5H7NfJWDPrDPQC/gsodvdNiafKgWQXfR4GrK/xeENiLNl7jwHGAHTq1Gl/yhIRida994br4+fMgZ/8JOpqvqXOQW9mBwGzgOvdfbvV+HPE3d3MPJ1C3H0iMBGgtLQ0rfcSEWlwf/5z+FLT0UfDgw+G+fjWraOuKqk6nR0ws6aEkH/C3Z9JDG82s3aJ59sBFUleuhHoWONxh8SYiEh+2r4dfvEL6N8f/uVfwtjBB+dsyEPdrroxYAqw2t3vrfHUXKD6KppRwJwkL38eGGxmbRInYQcnxkRE8s/8+XDMMTBxItxwA0yeHHVFdVKXI/qTgEuAAWa2PHEbCtwFDDKzNcDpiceYWamZTQZw963AHcDridvtiTERkfwycyaceSZ897vw6qtwzz3QokXUVdWJuefedHhpaamXlZVFXYaIFLrqJmRFRaE/zYQJcM01WW9hUBdmttTdS5M9p2/Giogks3EjjBgBJ564uwnZDTfkZMjvi4JeRKQm99CTpmtXWLgwnHjNw3CvSU3NRKQwlZQkX4C7aVP45ptwVc2kSXDEEVkvLdMU9CJSmJKFPISQnzgRfv7znGhfkAkKehGR2v7hH6KuIKM0Ry8iEnMKehEpPH/7W9QVZJWCXkQKz/33R11BVinoRaQwvPQSrF4d7j/4IBx6aPLtipM14s1vCnoRibdt28KC3AMGwB13hLGDDw5X3bh/+1ZeHm29DUBBLyLxNXdu+OLT5Mlw441504Qs03R5pYjE0xNPwMUXQ/fu8OyzYQWoAqUjehGJD/fdX4Q65xy4777dy/wVMAW9iMTD+vVw1lnQr19oQnbggXD99XnfpyYTFPQikt+qquD3vw8Lgrz0Elx7LRxwQNRV5RTN0YtI/tq6NUzR/PnPMHBg6FHz/e9HXVXOUdCLSP5q3RpatgxX01xxRWyakGWapm5EJL+89Rb8+MfhevdGjeA//xNGj1bI74WCXkTyw1dfwdixcPzx8MYbsHZt1BXljX1O3ZjZVGAYUOHu3RJjfwCOSmzSGvjU3Xsmee064DNgF7Az1XqGIiJ7tWRJOGpftSpcG3///eHbrVIndZmjnwY8BDxWPeDuF1TfN7N7gG17ef1p7r6lvgWKiPDAA/DZZzBvHgwdGnU1eWefQe/ui82sc7LnzMyA84EBmS1LRAreokXQrl1oYfDQQ9CkCbRqFXVVeSndOfpTgM3uvibF8w68YGZLzWzM3t7IzMaYWZmZlVVWVqZZlojkrU8/Dcv4nX46/Pa3YaxtW4V8GtIN+guBmXt5/mR3Pw44A7jazE5NtaG7T3T3UncvLSoqSrMsEclLzz4bjuCnTYObb4YpU6KuKBbqfR29mTUBzgGOT7WNu29M/Kwws9lAH2BxffcpIjFRUpJ6ce4ePcIlk8enjBbZT+l8Yep04B1335DsSTNrATRy988S9wcDt6exPxGJi1QhD/D669C0afZqKQD7nLoxs5nAa8BRZrbBzEYnnhpJrWkbM2tvZvMTD4uBV8zsTeBvwDx3fy5zpYtILCnkM64uV91cmGL8siRjHwFDE/ffA3qkWZ+IxE1VVdQVFBx9M1ZEsmfrVujfP+oqCo6CXkSyp3VraNMm6ioKjoJeRBrW8uXhmvjqJmRz5kBxcfJtU41LWhT0ItIw/v53uO02KC2FlSvh3Xd3P1deHpb9q30rL4+u3hhT0ItI5v31r9CzJ9x5J1xySWhGdtJJUVdVsLTwiIhk3iOPhCP655+HwYOjrqbgKehFJDNeeAE6dNjdhKxpUzjooKirEjR1IyLp+uQTuPzysOrTnXeGsTZtFPI5REEvIvX3zDPhCP7xx+HWW8ParZJzNHUjIvUzY0Y40dqrFyxYEE6+Sk5S0ItI3VVfAtmuHfz0p7BtG4wZo/40OU5TNyJSN+vWhXn4k06CHTvgwAPh6qsV8nlAQS8ie1dVBQ8+CN26wWuvwY03hpCXvKGpGxFJbetWOOssePVVGDIEJkyAww+PuirZTwp6EUmtdevQf+axx+Dii8Es6oqkHjR1IyJ7WrYMBgyATZtCE7JnnglX1yjk85aCXkSCL7+EW26BPn1g9Wp4//2oK5IMUdCLCLzySrgO/q67YNSo0ISsX7+oq5IMqcuasVPNrMLMVtYY+79mttHMliduQ1O8doiZ/beZrTWzmzNZuIjUQ0lJmIKpfRs0CL7+GhYuhClTtDhIzNTlZOw04CHgsVrj97n771K9yMwaAw8Dg4ANwOtmNtfdV9WzVhFJ1+bNycf//ndYsUL9aWJqn0f07r4Y2FqP9+4DrHX399z9a+ApYHg93kdEskEhH1vpzNFfY2ZvJaZ2kv2ddxiwvsbjDYmxpMxsjJmVmVlZZWVlGmWJSFLuUVcgEalv0I8HfgD0BDYB96RbiLtPdPdSdy8tKipK9+1EpLYZM6KuQCJSr6B3983uvsvdq4BJhGma2jYCHWs87pAYE5FscYeNiX92550XbS0SmXoFvZm1q/HwbGBlks1eB7qY2ffMrBkwEphbn/2JSD28/35Yxu/kk0MTsu98J3zLNZlU4xIL+7zqxsxmAv2BQ8xsAzAW6G9mPQEH1gFXJrZtD0x296HuvtPMrgGeBxoDU9397Qb5rxCR3XbtCkv53XorNG4M//7vu5uQlZdHW5tEwjwHT9CUlpZ6WVlZ1GWI5J+PP4Zhw2DJEhg6NDQh69hx36+TvGdmS929NNlzamomEidt2kD79uHE689+pv40AqgFgkj+e/11+NGP4KOPQhOyWbPgoosU8vK/FPQi+eqLL+Cmm6BvX1i7Fj78MOqKJEcp6EXy0csvQ48ecPfdMHp0aELWt2/UVUmO0hy9SD6aMiUs8bdoUegdL7IXCnqRfDFvXljGr1u3sIZr06bQokXUVUke0NSNSK7bsiUs4zdsWLgmHsISfwp5qSMFvUiucoennoIf/hCefhrGjoXJk6OuSvKQpm5EctXjj4fVnnr3DnPy3btHXZHkKQW9SC6pqgpNyDp2hPPPDwuCjB4dWhmI1JOmbkRyxdq1MHAgnHLK7iZkY8Yo5CVtCnqRqO3aBffcA8ceC8uWwT//MzRvHnVVEiOauhGJ0scfwxlnhDYGZ50F48fDYSkXYhOpFx3Ri2RDSUnoPVP7dswx0LkzzJwJc+Yo5KVB6IheJBs2b049/vTT2a1FCo6O6EVEYk5BLyIScwp6EZGY22fQm9lUM6sws5U1xu42s3fM7C0zm21mrVO8dp2ZrTCz5WamtQGlsMydCytWRF2FSJ2O6KcBQ2qNLQS6ufuxwP8At+zl9ae5e89UaxmKxE5FBYwcCcOHh37xAMXFybdNNS6SQfsMendfDGytNfaCu+9MPFwCdGiA2kTyizs88QR07QqzZ8Mdd+xuQlZeHp6vfSsvj7ZmKQiZmKO/AliQ4jkHXjCzpWY2JgP7Esldjz0W2gl36QJvvBG+4dqsWdRViaR3Hb2Z3QbsBJ5IscnJ7r7RzA4FFprZO4m/EJK91xhgDECnTp3SKUske6qqYMMG6NQJLrgAdu6Eyy5TfxrJKfU+ojezy4BhwEXu7sm2cfeNiZ8VwGygT6r3c/eJ7l7q7qVFRUX1LUske9asCcv41WxCpk6TkoPqFfRmNgS4CfiJu3+RYpsWZtay+j4wGFiZbFuRvLJzZzjJeuyxsHx5WBBETcgkh+1z6sbMZgL9gUPMbAMwlnCVzQGE6RiAJe5+lZm1Bya7+1CgGJideL4J8KS7P9cg/xUi2bJlS2hCVlYGI0bAww9D+/ZRVyWyV/sMene/MMnwlBTbfgQMTdx/D+iRVnUiuaZtW/jBD+Cmm+Dcc0NjMpEcp2/GiuzLa69Bv35h5adGjcI6ruedp5CXvKGgF0llxw64/no46aQQ8hs3Rl2RSL0o6EWSefFF6NYNxo2DX/4SVq6EPikvGhPJaepHL5LMjBnhy06LF4fLJ0XymIJepNqzz8L3vx8um3zgAWjaFA48MOqqRNKmqRuRzZvh/PPh7LPh3nvDWKtWCnmJDQW9FC53ePzx0IRszhz413+FSZOirkok4zR1I4XrscdCX5p+/WDKFDj66KgrEmkQCnqJt5KS5AtzFxfDBx+EpmSXXqr+NBJrCnqJt2QhXz1+wAFw+eXZrUckApqjFxGJOQW9iEjMKeglnpIvkSBSkBT0Ej9//Sv07RtWfhIRBb3EyOefw7XXhpYFmzfDpk3h6ppkUo2LxJCuupF4eOEFGDMGPvwQrrkG7rwTDjoIysujrkwkcgp6iYcnnwwtC/7yl9BWWET+l4Je8tesWXDEEdCjR2hC1qxZWKBbRPagOXrJP5s2wU9/Gpbyu//+MNaqlUJeJIU6Bb2ZTTWzCjNbWWOsrZktNLM1iZ9tUrx2VGKbNWY2KlOFSwFyh0cfDU3I5s2Du+5SEzKROqjrEf00YEitsZuBRe7eBViUeLwHM2sLjAVOAPoAY1P9QhDZp2nT4IoroHt3eOst+Kd/giaafRTZlzoFvbsvBrbWGh4OTE/cnw6MSPLSHwML3X2ru38CLOTbvzBEUtu1C9atC/d/9jOYPh1efhmOPDLKqkTySjpz9MXuvilxvxxIdmHyYcD6Go83JMa+xczGmFmZmZVVVlamUZbExurVcOqp4bZjR2hCduml0EinlkT2R0b+xbi7A2l959zdJ7p7qbuXFhUVZaIsyVfffBMWAenZE955J9xv3jzqqkTyVjoTnJvNrJ27bzKzdkBFkm02Av1rPO4AvJzGPiXuKith0CB4882wvN8DD+hbrCJpSueIfi5QfRXNKGBOkm2eBwabWZvESdjBiTGRPVU3ITvkkHCydfZs+MMfFPIiGVDXyytnAq8BR5nZBjMbDdwFDDKzNcDpiceYWamZTQZw963AHcDridvtiTGR3RYvhj59QhMys7CO64hk5/ZFpD7qNHXj7hemeGpgkm3LgJ/XeDwVmFqv6iTetm+HW26BRx6B730v9KXp0CHqqkRiR5cvSDQWLIBu3WD8eLj+elixAkpLo65KJJb0bROJxn/8B7RsCa++GnrHi0iDUdBLwygpSb4w98EHw5YtMG5caEJ2wAHZr02kwCjopWEkC3mAjz8OP1u2zF4tIgVOc/QiIjGnoBcRiTlN3Uhm7doFH3wQdRUiUoOO6CVz3n47LOP3ox9FXYmI1KCgl/R9/TXcfjv06gVr18K//Vvq1gVqaSCSdZq6kfRUVsLAgeELTxdeGC6bLCoKveNFJCfoiF7qp2YTsl69YO5cePLJEPIiklMU9LL/Xn45tCtYvz40IZs+Hc46K+qqRCQFBb3U3bZtcNVVcNpp8OmnUJFsCQIRyTUKeqmbefPgmGNg0iT41a/CnPzxx0ddlYjUgU7GSt3MmgVt2sAzz4Te8SKSNxT0kpx7WOHpqKPCydZx40IDsmbNoq5MRPaTpm7k2zZsgOHDw+WSDz0Uxlq2VMiL5CkFvexWVQUTJ4a5+BdfhHvuCY9FJK/VO+jN7CgzW17jtt3Mrq+1TX8z21Zjm9+kX7I0mGnT4Morw6WTK1bADTdA48ZRVyUiaar3HL27/zfQE8DMGgMbgdlJNv2Luw+r736kge3aBe+/D0ccARdfDM2bwwUXhOvjRSQWMjV1MxB4193VtjCfrFgBJ54I/fvDjh1hDn7kSIW8SMxkKuhHAjNTPHeimb1pZgvM7JhUb2BmY8yszMzKKisrM1SWJPXVVzB2LBx3HKxbF+bimzePuioRaSDm1T1L6vsGZs2Aj4Bj3H1zredaAVXu/rmZDQXGuXuXfb1naWmpl5WVpVWXpFBRAQMGhJbCF10E998f+tWISF4zs6XuXprsuUwc0Z8BLKsd8gDuvt3dP0/cnw80NTOlShSqf6EXFUHv3vCnP8GMGQp5kQKQiS9MXUiKaRszKwE2u7ubWR/CL5aPM7BPSaWkJPnC3E2awHvvQceO8Oij2a9LRCKTVtCbWQtgEHBljbGrANx9AnAu8Asz2wl8CYz0dOeKZO+ShTzAzp2wZUsIehEpKGnP0TcEzdGnYW9XzOTg/2sRyYyGnqMXEZEcpqZmceEeVngSEalFR/RxsH49DBsWvtkqIlKLgj6fVVXB+PGhCdnLL4dr4ouLk2+balxEYk9Bn8+mTYNf/hJOOAFWroTrroPy8jCNU/tWXh51tSISEc3R55udO0MTsi5dwlTNQQfBeeepP42IpKQj+nzy5pvQt29YnLu6Cdn55yvkRWSvFPT54Kuv4Ne/Dn3i168Pc/FqQiYidaSpm1xXURHaCK9eDZdeCvfeCwcfHHVVIpJHdESfq2o2IevXDxYsgOnTFfIist8U9Llo4ULo0QM+/DDMv0+eDEOGRF2ViOQpBX0u+eQTGD0aBg8O8/Jbt0ZdkYjEgII+V8yeDV27humZW24JV9j07Bl1VSISAzoZmyvmzQu95OfNC0v8iYhkiII+Ku7w+OPQrVsI9nHjwnXxTZtGXZmIxIymbqLwwQdwxhkwahRMmBDGWrRQyItIg1DQZ1NVFTz8cDiKf+UVeOCB3UEvItJANHWTTY8+CtdcA4MGwcSJ0Llz1BWJSAFQ0De0b74JTciOPBIuuQRatYJzz1V/GhHJmrSD3szWAZ8Bu4CdtdcsNDMDxgFDgS+Ay9x9Wbr7zUklJckX527UCLZvD/Pw552X/bpEpKBl6oj+NHffkuK5M4AuidsJwPjEz/hJFvIQ5uZbtMhuLSIiCdk4GTsceMyDJUBrM2uXhf2KiAiZCXoHXjCzpWY2JsnzhwHrazzekBjbg5mNMbMyMyurrKzMQFlZVt2ETEQkx2Qi6E929+MIUzRXm9mp9XkTd5/o7qXuXlpUVJSBsrLo+efh2GOjrkJEJKm0g97dNyZ+VgCzgT61NtkIdKzxuENiLP9t3QqXXRY6S+7cGXU1IiJJpRX0ZtbCzFpW3wcGAytrbTYXuNSCvsA2d9+Uzn5zwqxZoQnZjBlw223wxhtQXJx821TjIiJZkO5VN8XA7HAFJU2AJ939OTO7CsDdJwDzCZdWriVcXnl5mvvMDc8/D+3bw3PP7e4yWV4ebU0iIkmY5+BJxNLSUi8rK4u6jD25w7Rp0L17WLt1xw444ABoou+ciUj0zGxp7e8xVVOvm7p4//2wGMgVV8CkSWGsRQuFvIjkBQX93uzaFRqPdesGS5bAI4/A+PFRVyUisl90SLo306bBddeFlsITJkCnTlFXJCKy3xT0tX3zDbz7Lhx9NFx6KbRtCyNGqAmZiOQtTd3UtGwZ9O4NAwaEk61Nm8LZZyvkRSSvKegBvvwSbr4Z+vSBioowF68mZCISE5q62bwZTjkF1qyB0aPhd7+D1q2jrkpEJGMKN+irqkKf+EMPhdNOC1fTDBwYdVUiIhlXmFM38+eHSybXrQvz77//vUJeRGKrsIJ+y5awnN+ZZ4aA37496opERBpc4QT900+HJmRPPQW/+U24wkathUWkABTOHP2iRXD44eFn9+5RVyMikjXxDXp3mDo1HLX37g333QfNmqk/jYgUnHikXklJ6oW5r7wyBH3z5tmtSUQkR8Rjjj5VyEP48pOISAGLR9DvTaP4/yeKiOyNUlBEJOYU9CIiMVfvoDezjmb2kpmtMrO3zey6JNv0N7NtZrY8cftNeuWKiMj+Sueqm53Ar9x9mZm1BJaa2UJ3X1Vru7+4+7A09rNvxcXJT8gWFzfobkVE8kG9g97dNwGbEvc/M7PVwGFA7aBveOXlWd+liEi+yMgcvZl1BnoB/5Xk6RPN7E0zW2Bmx+zlPcaYWZmZlVVWVmaiLBERIQNBb2YHAbOA6929dpewZcDh7t4DeBB4NtX7uPtEdy9199KioqJ0yxIRkYS0gt7MmhJC/gl3f6b28+6+3d0/T9yfDzQ1s0PS2aeIiOyfdK66MWAKsNrd702xTUliO8ysT2J/H9d3nyIisv/SuermJOASYIWZLU+M3Qp0AnD3CcC5wC/MbCfwJTDS3T2NfYqIyH6yXMxdM6sEPoi6jjQdAmyJuogcoc9iT/o89qTPY7d0PovD3T3pCc6cDPo4MLMydy+Nuo5coM9iT/o89qTPY7eG+izUAkFEJOYU9CIiMaegbzgToy4gh+iz2JM+jz3p89itQT4LzdGLiMScjuhFRGJOQS8iEnMK+gyqS4/+QmRmjc3sDTP7U9S1RMnMWpvZH83sHTNbbWYnRl1TlMzs/yT+naw0s5lm9p2oa8omM5tqZhVmtrLGWFszW2hmaxI/22RiXwr6zKru0d8V6AtcbWZdI64pF1wHrI66iBwwDnjO3Y8GelDAn4mZHQZcC5S6ezegMTAy2qqybhowpNbYzcAid+8CLEo8TpuCPoPcfZO7L0vc/4zwD/mwaKuKlpl1AM4EJkddS5TM7LvAqYT+ULj71+7+abRVRa4JcKCZNQGaAx9FXE9WuftiYGut4eHA9MT96cCITOxLQd9A9tGjv5DcD9wEVEVdSMS+B1QCjyamsSabWYuoi4qKu28Efgd8SFjAaJu7vxBtVTmhOLGoE0A5kJFl8hT0DWAfPfoLhpkNAyrcfWnUteSAJsBxwHh37wXsIEN/luejxNzzcMIvwPZACzO7ONqqckuiAWRGrn9X0GfYvnr0F5iTgJ+Y2TrgKWCAmc2ItqTIbAA2uHv1X3h/JAR/oTodeN/dK939G+AZoF/ENeWCzWbWDiDxsyITb6qgz6C69OgvJO5+i7t3cPfOhBNt/8/dC/Kozd3LgfVmdlRiaCBRrK+cOz4E+ppZ88S/m4EU8MnpGuYCoxL3RwFzMvGmCvrMqu7RP8DMliduQ6MuSnLGPwJPmNlbQE/gzojriUziL5s/EpYbXUHIooJqhWBmM4HXgKPMbIOZjQbuAgaZ2RrCXz13ZWRfaoEgIhJvOqIXEYk5Bb2ISMwp6EVEYk5BLyIScwp6EZGYU9CLiMScgl5EJOb+P9X7GVr3BI8EAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "djyKWgmRCzvR",
        "colab_type": "text"
      },
      "source": [
        "자, 이제 학습을 위해서 모델을 만들어 봅시다!\n",
        "\n",
        "가장 쉽게 모델을 만드는 방법은 **Sequential**을 사용하는 것입니다.\n",
        "\n",
        "차곡차곡, 입력부터 층을 만들어 나가는 방법이지요!\n",
        "\n",
        "층의 속성도 여러가지가 있습니다만, 여기서는 가장 간단한 **Dense** type만 사용하겠습니다.\n",
        "\n",
        "첫 모델에는 활성화 함수 (activation function)은 사용하지 않겠습니다. 이유를 아시나요 ? ^^\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "O6Q8DFp7Pu7M",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from keras.models import Sequential\n",
        "from keras.layers import Dense\n",
        "\n",
        "model = Sequential()\n",
        "model.add(Dense(1, input_dim=1))"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nsWM2GcZD2F5",
        "colab_type": "text"
      },
      "source": [
        "모델이 만들어 졌으면, 학습 및 최적화 방법에 대한 옵션을 기술하고, 모델을 Compile하겠습니다!"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ECujJ4BJQHTz",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# from keras import optimizers\n",
        "model.compile(loss='mean_squared_error', optimizer='adam')   # keras.optimizers.Adam(learning_rate=0.001)"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "A565ACc5EAtR",
        "colab_type": "text"
      },
      "source": [
        "자, 이제 학습을 진행해볼까요 ? 첫 학습을 위해서 **epochs**는 10번, **batch_size**는 1로 설정하여 진행하겠습니다."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FF6qIa0bQL6b",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "af7305e6-9a14-408d-910d-07946f4953ca"
      },
      "source": [
        "model.fit(data, outcome, epochs=1000, batch_size=1)"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 475.8527\n",
            "Epoch 2/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 473.2504\n",
            "Epoch 3/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 470.9315\n",
            "Epoch 4/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 468.6291\n",
            "Epoch 5/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 465.9984\n",
            "Epoch 6/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 463.9283\n",
            "Epoch 7/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 461.7060\n",
            "Epoch 8/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 458.9204\n",
            "Epoch 9/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 456.4214\n",
            "Epoch 10/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 454.2605\n",
            "Epoch 11/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 451.9039\n",
            "Epoch 12/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 449.7944\n",
            "Epoch 13/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 447.1524\n",
            "Epoch 14/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 445.1161\n",
            "Epoch 15/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 442.6039\n",
            "Epoch 16/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 440.4099\n",
            "Epoch 17/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 437.8669\n",
            "Epoch 18/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 435.8148\n",
            "Epoch 19/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 433.5076\n",
            "Epoch 20/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 431.1761\n",
            "Epoch 21/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 428.9161\n",
            "Epoch 22/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 426.5427\n",
            "Epoch 23/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 424.3986\n",
            "Epoch 24/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 422.1821\n",
            "Epoch 25/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 420.0020\n",
            "Epoch 26/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 417.9572\n",
            "Epoch 27/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 415.3922\n",
            "Epoch 28/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 413.6180\n",
            "Epoch 29/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 411.3093\n",
            "Epoch 30/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 408.7874\n",
            "Epoch 31/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 406.9422\n",
            "Epoch 32/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 404.4621\n",
            "Epoch 33/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 402.5219\n",
            "Epoch 34/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 400.4158\n",
            "Epoch 35/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 398.0890\n",
            "Epoch 36/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 396.2667\n",
            "Epoch 37/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 394.0217\n",
            "Epoch 38/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 391.8496\n",
            "Epoch 39/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 389.7091\n",
            "Epoch 40/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 387.5480\n",
            "Epoch 41/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 385.4619\n",
            "Epoch 42/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 383.4421\n",
            "Epoch 43/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 381.5735\n",
            "Epoch 44/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 379.3450\n",
            "Epoch 45/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 377.3698\n",
            "Epoch 46/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 375.1612\n",
            "Epoch 47/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 373.1220\n",
            "Epoch 48/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 371.3160\n",
            "Epoch 49/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 369.0662\n",
            "Epoch 50/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 367.0267\n",
            "Epoch 51/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 365.1685\n",
            "Epoch 52/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 363.1587\n",
            "Epoch 53/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 361.0590\n",
            "Epoch 54/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 359.2608\n",
            "Epoch 55/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 357.2827\n",
            "Epoch 56/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 355.0571\n",
            "Epoch 57/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 353.4460\n",
            "Epoch 58/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 351.4068\n",
            "Epoch 59/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 349.2935\n",
            "Epoch 60/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 347.3548\n",
            "Epoch 61/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 345.7328\n",
            "Epoch 62/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 343.4305\n",
            "Epoch 63/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 341.7654\n",
            "Epoch 64/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 339.8135\n",
            "Epoch 65/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 337.7458\n",
            "Epoch 66/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 335.8428\n",
            "Epoch 67/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 334.2375\n",
            "Epoch 68/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 332.1272\n",
            "Epoch 69/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 330.6478\n",
            "Epoch 70/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 328.4651\n",
            "Epoch 71/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 326.8396\n",
            "Epoch 72/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 324.9181\n",
            "Epoch 73/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 322.8676\n",
            "Epoch 74/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 321.1146\n",
            "Epoch 75/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 319.3608\n",
            "Epoch 76/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 317.4316\n",
            "Epoch 77/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 315.7673\n",
            "Epoch 78/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 313.7983\n",
            "Epoch 79/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 312.2939\n",
            "Epoch 80/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 310.3501\n",
            "Epoch 81/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 308.7562\n",
            "Epoch 82/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 306.6936\n",
            "Epoch 83/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 305.1281\n",
            "Epoch 84/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 303.2305\n",
            "Epoch 85/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 301.4789\n",
            "Epoch 86/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 299.6868\n",
            "Epoch 87/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 298.1301\n",
            "Epoch 88/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 296.5446\n",
            "Epoch 89/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 294.5798\n",
            "Epoch 90/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 293.1342\n",
            "Epoch 91/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 291.0724\n",
            "Epoch 92/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 289.4785\n",
            "Epoch 93/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 287.9344\n",
            "Epoch 94/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 286.0593\n",
            "Epoch 95/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 284.4674\n",
            "Epoch 96/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 282.8621\n",
            "Epoch 97/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 281.1066\n",
            "Epoch 98/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 279.5932\n",
            "Epoch 99/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 277.9141\n",
            "Epoch 100/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 276.1844\n",
            "Epoch 101/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 274.5178\n",
            "Epoch 102/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 272.9781\n",
            "Epoch 103/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 271.3509\n",
            "Epoch 104/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 269.6766\n",
            "Epoch 105/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 268.1902\n",
            "Epoch 106/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 266.5132\n",
            "Epoch 107/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 265.0292\n",
            "Epoch 108/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 263.3456\n",
            "Epoch 109/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 261.7183\n",
            "Epoch 110/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 260.0681\n",
            "Epoch 111/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 258.5497\n",
            "Epoch 112/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 257.2444\n",
            "Epoch 113/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 255.5660\n",
            "Epoch 114/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 253.8511\n",
            "Epoch 115/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 252.5193\n",
            "Epoch 116/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 250.8577\n",
            "Epoch 117/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 249.4575\n",
            "Epoch 118/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 247.8817\n",
            "Epoch 119/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 246.3331\n",
            "Epoch 120/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 244.8705\n",
            "Epoch 121/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 243.3645\n",
            "Epoch 122/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 241.7931\n",
            "Epoch 123/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 240.3678\n",
            "Epoch 124/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 238.9525\n",
            "Epoch 125/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 237.4526\n",
            "Epoch 126/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 235.8464\n",
            "Epoch 127/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 234.4673\n",
            "Epoch 128/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 233.1945\n",
            "Epoch 129/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 231.5168\n",
            "Epoch 130/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 230.2049\n",
            "Epoch 131/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 228.7411\n",
            "Epoch 132/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 227.2524\n",
            "Epoch 133/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 225.8506\n",
            "Epoch 134/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 224.4296\n",
            "Epoch 135/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 223.0592\n",
            "Epoch 136/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 221.7211\n",
            "Epoch 137/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 220.2066\n",
            "Epoch 138/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 218.8552\n",
            "Epoch 139/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 217.4847\n",
            "Epoch 140/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 216.1119\n",
            "Epoch 141/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 214.6852\n",
            "Epoch 142/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 213.4487\n",
            "Epoch 143/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 211.9239\n",
            "Epoch 144/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 210.6598\n",
            "Epoch 145/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 209.2579\n",
            "Epoch 146/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 208.0912\n",
            "Epoch 147/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 206.6802\n",
            "Epoch 148/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 205.2142\n",
            "Epoch 149/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 203.9270\n",
            "Epoch 150/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 202.6406\n",
            "Epoch 151/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 201.4000\n",
            "Epoch 152/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 200.0937\n",
            "Epoch 153/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 198.7552\n",
            "Epoch 154/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 197.4861\n",
            "Epoch 155/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 196.1219\n",
            "Epoch 156/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 194.8049\n",
            "Epoch 157/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 193.5878\n",
            "Epoch 158/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 192.4179\n",
            "Epoch 159/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 191.1284\n",
            "Epoch 160/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 189.8665\n",
            "Epoch 161/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 188.6214\n",
            "Epoch 162/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 187.2411\n",
            "Epoch 163/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 186.0749\n",
            "Epoch 164/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 184.8801\n",
            "Epoch 165/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 183.5434\n",
            "Epoch 166/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 182.3552\n",
            "Epoch 167/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 181.1126\n",
            "Epoch 168/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 180.1822\n",
            "Epoch 169/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 178.7738\n",
            "Epoch 170/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 177.4336\n",
            "Epoch 171/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 176.3920\n",
            "Epoch 172/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 175.2658\n",
            "Epoch 173/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 174.0512\n",
            "Epoch 174/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 172.7823\n",
            "Epoch 175/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 171.7480\n",
            "Epoch 176/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 170.4176\n",
            "Epoch 177/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 169.2995\n",
            "Epoch 178/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 168.2595\n",
            "Epoch 179/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 166.9047\n",
            "Epoch 180/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 165.9438\n",
            "Epoch 181/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 164.6196\n",
            "Epoch 182/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 163.6015\n",
            "Epoch 183/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 162.4264\n",
            "Epoch 184/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 161.3885\n",
            "Epoch 185/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 160.2058\n",
            "Epoch 186/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 159.1216\n",
            "Epoch 187/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 157.9919\n",
            "Epoch 188/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 156.9733\n",
            "Epoch 189/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 155.7892\n",
            "Epoch 190/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 154.6935\n",
            "Epoch 191/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 153.6806\n",
            "Epoch 192/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 152.5374\n",
            "Epoch 193/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 151.5491\n",
            "Epoch 194/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 150.3883\n",
            "Epoch 195/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 149.3120\n",
            "Epoch 196/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 148.3504\n",
            "Epoch 197/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 147.2133\n",
            "Epoch 198/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 146.1237\n",
            "Epoch 199/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 145.1259\n",
            "Epoch 200/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 144.1407\n",
            "Epoch 201/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 143.0743\n",
            "Epoch 202/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 142.1034\n",
            "Epoch 203/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 140.9769\n",
            "Epoch 204/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 139.9994\n",
            "Epoch 205/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 139.0390\n",
            "Epoch 206/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 137.9433\n",
            "Epoch 207/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 137.0588\n",
            "Epoch 208/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 135.9100\n",
            "Epoch 209/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 134.9732\n",
            "Epoch 210/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 134.0535\n",
            "Epoch 211/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 133.0892\n",
            "Epoch 212/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 132.1192\n",
            "Epoch 213/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 131.0943\n",
            "Epoch 214/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 130.0704\n",
            "Epoch 215/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 129.2707\n",
            "Epoch 216/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 128.1143\n",
            "Epoch 217/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 127.3117\n",
            "Epoch 218/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 126.2325\n",
            "Epoch 219/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 125.3741\n",
            "Epoch 220/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 124.3699\n",
            "Epoch 221/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 123.5916\n",
            "Epoch 222/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 122.5510\n",
            "Epoch 223/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 121.6086\n",
            "Epoch 224/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 120.8730\n",
            "Epoch 225/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 119.7525\n",
            "Epoch 226/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 118.9365\n",
            "Epoch 227/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 117.9975\n",
            "Epoch 228/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 117.0200\n",
            "Epoch 229/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 116.3005\n",
            "Epoch 230/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 115.3185\n",
            "Epoch 231/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 114.3735\n",
            "Epoch 232/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 113.5789\n",
            "Epoch 233/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 112.7113\n",
            "Epoch 234/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 111.7931\n",
            "Epoch 235/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 110.9716\n",
            "Epoch 236/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 110.0738\n",
            "Epoch 237/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 109.2698\n",
            "Epoch 238/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 108.3973\n",
            "Epoch 239/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 107.6021\n",
            "Epoch 240/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 106.6011\n",
            "Epoch 241/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 105.8211\n",
            "Epoch 242/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 105.0956\n",
            "Epoch 243/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 104.2893\n",
            "Epoch 244/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 103.3147\n",
            "Epoch 245/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 102.6240\n",
            "Epoch 246/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 101.7323\n",
            "Epoch 247/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 100.9058\n",
            "Epoch 248/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 100.2073\n",
            "Epoch 249/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 99.3019\n",
            "Epoch 250/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 98.4800\n",
            "Epoch 251/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 97.7473\n",
            "Epoch 252/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 97.0759\n",
            "Epoch 253/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 96.1471\n",
            "Epoch 254/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 95.3328\n",
            "Epoch 255/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 94.6561\n",
            "Epoch 256/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 93.8559\n",
            "Epoch 257/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 93.0603\n",
            "Epoch 258/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 92.3820\n",
            "Epoch 259/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 91.5682\n",
            "Epoch 260/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 90.8459\n",
            "Epoch 261/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 90.0901\n",
            "Epoch 262/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 89.2642\n",
            "Epoch 263/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 88.6131\n",
            "Epoch 264/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 87.8155\n",
            "Epoch 265/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 87.2129\n",
            "Epoch 266/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 86.3620\n",
            "Epoch 267/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 85.6639\n",
            "Epoch 268/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 84.8835\n",
            "Epoch 269/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 84.3248\n",
            "Epoch 270/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 83.4889\n",
            "Epoch 271/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 82.8479\n",
            "Epoch 272/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 82.1729\n",
            "Epoch 273/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 81.4460\n",
            "Epoch 274/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 80.6978\n",
            "Epoch 275/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 80.1113\n",
            "Epoch 276/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 79.3392\n",
            "Epoch 277/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 78.6937\n",
            "Epoch 278/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 78.0476\n",
            "Epoch 279/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 77.3114\n",
            "Epoch 280/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 76.6871\n",
            "Epoch 281/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 76.0249\n",
            "Epoch 282/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 75.3894\n",
            "Epoch 283/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 74.6706\n",
            "Epoch 284/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 73.9901\n",
            "Epoch 285/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 73.3511\n",
            "Epoch 286/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 72.8076\n",
            "Epoch 287/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 72.1535\n",
            "Epoch 288/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 71.4821\n",
            "Epoch 289/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 70.8079\n",
            "Epoch 290/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 70.3164\n",
            "Epoch 291/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 69.5538\n",
            "Epoch 292/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 68.9522\n",
            "Epoch 293/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 68.3203\n",
            "Epoch 294/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 67.7806\n",
            "Epoch 295/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 67.1939\n",
            "Epoch 296/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 66.5389\n",
            "Epoch 297/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 65.9431\n",
            "Epoch 298/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 65.3303\n",
            "Epoch 299/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 64.7221\n",
            "Epoch 300/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 64.1395\n",
            "Epoch 301/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 63.5344\n",
            "Epoch 302/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 63.0594\n",
            "Epoch 303/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 62.3952\n",
            "Epoch 304/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 61.8297\n",
            "Epoch 305/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 61.3153\n",
            "Epoch 306/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 60.6966\n",
            "Epoch 307/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 60.1416\n",
            "Epoch 308/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 59.6128\n",
            "Epoch 309/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 59.0133\n",
            "Epoch 310/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 58.4620\n",
            "Epoch 311/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 57.9118\n",
            "Epoch 312/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 57.4605\n",
            "Epoch 313/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 56.8121\n",
            "Epoch 314/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 56.2941\n",
            "Epoch 315/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 55.7850\n",
            "Epoch 316/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 55.2233\n",
            "Epoch 317/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 54.7961\n",
            "Epoch 318/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 54.1950\n",
            "Epoch 319/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 53.6233\n",
            "Epoch 320/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 53.2185\n",
            "Epoch 321/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 52.6718\n",
            "Epoch 322/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 52.1688\n",
            "Epoch 323/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 51.6433\n",
            "Epoch 324/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 51.0844\n",
            "Epoch 325/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 50.6929\n",
            "Epoch 326/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 50.1983\n",
            "Epoch 327/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 49.6638\n",
            "Epoch 328/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 49.1181\n",
            "Epoch 329/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 48.7354\n",
            "Epoch 330/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 48.2001\n",
            "Epoch 331/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 47.7139\n",
            "Epoch 332/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 47.2340\n",
            "Epoch 333/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 46.8637\n",
            "Epoch 334/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 46.3154\n",
            "Epoch 335/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 45.8276\n",
            "Epoch 336/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 45.4420\n",
            "Epoch 337/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 44.9785\n",
            "Epoch 338/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 44.4802\n",
            "Epoch 339/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 44.0157\n",
            "Epoch 340/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 43.5967\n",
            "Epoch 341/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 43.1179\n",
            "Epoch 342/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 42.7487\n",
            "Epoch 343/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 42.3268\n",
            "Epoch 344/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 41.8463\n",
            "Epoch 345/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 41.3857\n",
            "Epoch 346/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 40.9638\n",
            "Epoch 347/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 40.5824\n",
            "Epoch 348/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 40.1277\n",
            "Epoch 349/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 39.7685\n",
            "Epoch 350/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 39.2933\n",
            "Epoch 351/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 38.8945\n",
            "Epoch 352/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 38.5173\n",
            "Epoch 353/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 38.0596\n",
            "Epoch 354/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 37.6871\n",
            "Epoch 355/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 37.2959\n",
            "Epoch 356/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 36.8905\n",
            "Epoch 357/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 36.5326\n",
            "Epoch 358/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 36.0828\n",
            "Epoch 359/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 35.7611\n",
            "Epoch 360/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 35.3384\n",
            "Epoch 361/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 34.9260\n",
            "Epoch 362/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 34.6123\n",
            "Epoch 363/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 34.2359\n",
            "Epoch 364/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 33.8109\n",
            "Epoch 365/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 33.5282\n",
            "Epoch 366/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 33.1039\n",
            "Epoch 367/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 32.7263\n",
            "Epoch 368/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 32.4038\n",
            "Epoch 369/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 31.9887\n",
            "Epoch 370/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 31.7254\n",
            "Epoch 371/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 31.3381\n",
            "Epoch 372/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 30.9622\n",
            "Epoch 373/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 30.6091\n",
            "Epoch 374/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 30.3165\n",
            "Epoch 375/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 29.9950\n",
            "Epoch 376/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 29.6390\n",
            "Epoch 377/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 29.2767\n",
            "Epoch 378/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 28.9364\n",
            "Epoch 379/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 28.6104\n",
            "Epoch 380/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 28.3040\n",
            "Epoch 381/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 27.9759\n",
            "Epoch 382/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 27.6629\n",
            "Epoch 383/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 27.3704\n",
            "Epoch 384/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 27.0303\n",
            "Epoch 385/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 26.7329\n",
            "Epoch 386/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 26.4158\n",
            "Epoch 387/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 26.0795\n",
            "Epoch 388/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 25.8126\n",
            "Epoch 389/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 25.5111\n",
            "Epoch 390/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 25.1848\n",
            "Epoch 391/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 24.9085\n",
            "Epoch 392/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 24.6256\n",
            "Epoch 393/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 24.3445\n",
            "Epoch 394/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 24.0243\n",
            "Epoch 395/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 23.7980\n",
            "Epoch 396/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 23.4548\n",
            "Epoch 397/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 23.2224\n",
            "Epoch 398/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 22.9119\n",
            "Epoch 399/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 22.6459\n",
            "Epoch 400/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 22.3537\n",
            "Epoch 401/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 22.1312\n",
            "Epoch 402/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 21.8114\n",
            "Epoch 403/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 21.6034\n",
            "Epoch 404/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 21.2927\n",
            "Epoch 405/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 21.0645\n",
            "Epoch 406/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 20.7907\n",
            "Epoch 407/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 20.5503\n",
            "Epoch 408/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 20.2889\n",
            "Epoch 409/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 20.0642\n",
            "Epoch 410/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 19.7670\n",
            "Epoch 411/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 19.5818\n",
            "Epoch 412/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 19.2927\n",
            "Epoch 413/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 19.0629\n",
            "Epoch 414/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 18.8428\n",
            "Epoch 415/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 18.5690\n",
            "Epoch 416/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 18.3604\n",
            "Epoch 417/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 18.1442\n",
            "Epoch 418/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 17.8977\n",
            "Epoch 419/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 17.6521\n",
            "Epoch 420/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 17.4474\n",
            "Epoch 421/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 17.2364\n",
            "Epoch 422/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 17.0161\n",
            "Epoch 423/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 16.7878\n",
            "Epoch 424/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 16.5742\n",
            "Epoch 425/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 16.3809\n",
            "Epoch 426/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 16.1378\n",
            "Epoch 427/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 15.9324\n",
            "Epoch 428/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 15.7811\n",
            "Epoch 429/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 15.5415\n",
            "Epoch 430/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 15.3294\n",
            "Epoch 431/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 15.1189\n",
            "Epoch 432/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 14.9472\n",
            "Epoch 433/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 14.7300\n",
            "Epoch 434/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 14.5231\n",
            "Epoch 435/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 14.3591\n",
            "Epoch 436/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 14.1588\n",
            "Epoch 437/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 13.9777\n",
            "Epoch 438/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 13.7741\n",
            "Epoch 439/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 13.6100\n",
            "Epoch 440/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 13.4277\n",
            "Epoch 441/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 13.2449\n",
            "Epoch 442/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 13.0956\n",
            "Epoch 443/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 12.8924\n",
            "Epoch 444/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 12.7145\n",
            "Epoch 445/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 12.5282\n",
            "Epoch 446/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 12.3862\n",
            "Epoch 447/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 12.1978\n",
            "Epoch 448/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 12.0462\n",
            "Epoch 449/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 11.8866\n",
            "Epoch 450/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 11.7146\n",
            "Epoch 451/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 11.5554\n",
            "Epoch 452/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 11.4002\n",
            "Epoch 453/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 11.2439\n",
            "Epoch 454/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 11.0927\n",
            "Epoch 455/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 10.9269\n",
            "Epoch 456/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 10.7822\n",
            "Epoch 457/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 10.6336\n",
            "Epoch 458/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 10.4689\n",
            "Epoch 459/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 10.3411\n",
            "Epoch 460/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 10.2080\n",
            "Epoch 461/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 10.0442\n",
            "Epoch 462/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 9.9244\n",
            "Epoch 463/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 9.7596\n",
            "Epoch 464/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 9.6205\n",
            "Epoch 465/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 9.5078\n",
            "Epoch 466/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 9.3711\n",
            "Epoch 467/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 9.2232\n",
            "Epoch 468/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 9.1127\n",
            "Epoch 469/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 8.9682\n",
            "Epoch 470/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 8.8499\n",
            "Epoch 471/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 8.7045\n",
            "Epoch 472/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 8.5839\n",
            "Epoch 473/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 8.4790\n",
            "Epoch 474/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 8.3454\n",
            "Epoch 475/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 8.2400\n",
            "Epoch 476/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 8.0982\n",
            "Epoch 477/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 8.0088\n",
            "Epoch 478/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 7.8839\n",
            "Epoch 479/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 7.7607\n",
            "Epoch 480/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 7.6517\n",
            "Epoch 481/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 7.5474\n",
            "Epoch 482/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 7.4245\n",
            "Epoch 483/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 7.3326\n",
            "Epoch 484/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 7.2431\n",
            "Epoch 485/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 7.1069\n",
            "Epoch 486/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 7.0095\n",
            "Epoch 487/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 6.8950\n",
            "Epoch 488/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 6.8066\n",
            "Epoch 489/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 6.7171\n",
            "Epoch 490/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 6.6105\n",
            "Epoch 491/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 6.5258\n",
            "Epoch 492/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 6.4092\n",
            "Epoch 493/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 6.3199\n",
            "Epoch 494/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 6.2310\n",
            "Epoch 495/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 6.1488\n",
            "Epoch 496/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 6.0528\n",
            "Epoch 497/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 5.9521\n",
            "Epoch 498/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 5.8772\n",
            "Epoch 499/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 5.7961\n",
            "Epoch 500/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 5.6957\n",
            "Epoch 501/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 5.6171\n",
            "Epoch 502/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 5.5425\n",
            "Epoch 503/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 5.4506\n",
            "Epoch 504/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 5.3706\n",
            "Epoch 505/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 5.2948\n",
            "Epoch 506/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 5.2128\n",
            "Epoch 507/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 5.1468\n",
            "Epoch 508/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 5.0656\n",
            "Epoch 509/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 4.9869\n",
            "Epoch 510/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 4.9110\n",
            "Epoch 511/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 4.8407\n",
            "Epoch 512/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 4.7722\n",
            "Epoch 513/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 4.7118\n",
            "Epoch 514/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 4.6322\n",
            "Epoch 515/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 4.5641\n",
            "Epoch 516/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 4.5030\n",
            "Epoch 517/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 4.4273\n",
            "Epoch 518/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 4.3751\n",
            "Epoch 519/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 4.3091\n",
            "Epoch 520/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 4.2410\n",
            "Epoch 521/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 4.1798\n",
            "Epoch 522/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 4.1159\n",
            "Epoch 523/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 4.0690\n",
            "Epoch 524/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 3.9934\n",
            "Epoch 525/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 3.9397\n",
            "Epoch 526/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 3.8859\n",
            "Epoch 527/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 3.8263\n",
            "Epoch 528/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 3.7711\n",
            "Epoch 529/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 3.7303\n",
            "Epoch 530/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 3.6673\n",
            "Epoch 531/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 3.6096\n",
            "Epoch 532/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 3.5617\n",
            "Epoch 533/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 3.5122\n",
            "Epoch 534/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 3.4669\n",
            "Epoch 535/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 3.4123\n",
            "Epoch 536/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 3.3643\n",
            "Epoch 537/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 3.3152\n",
            "Epoch 538/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 3.2747\n",
            "Epoch 539/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 3.2295\n",
            "Epoch 540/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 3.1766\n",
            "Epoch 541/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 3.1359\n",
            "Epoch 542/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 3.0877\n",
            "Epoch 543/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 3.0469\n",
            "Epoch 544/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 3.0134\n",
            "Epoch 545/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.9660\n",
            "Epoch 546/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.9330\n",
            "Epoch 547/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.8807\n",
            "Epoch 548/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 2.8494\n",
            "Epoch 549/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.8082\n",
            "Epoch 550/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.7707\n",
            "Epoch 551/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 2.7333\n",
            "Epoch 552/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.6994\n",
            "Epoch 553/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 2.6653\n",
            "Epoch 554/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.6271\n",
            "Epoch 555/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.5954\n",
            "Epoch 556/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.5645\n",
            "Epoch 557/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.5304\n",
            "Epoch 558/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.4964\n",
            "Epoch 559/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.4628\n",
            "Epoch 560/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.4273\n",
            "Epoch 561/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.4102\n",
            "Epoch 562/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.3755\n",
            "Epoch 563/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.3373\n",
            "Epoch 564/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.3130\n",
            "Epoch 565/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.2857\n",
            "Epoch 566/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.2603\n",
            "Epoch 567/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.2299\n",
            "Epoch 568/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 2.2015\n",
            "Epoch 569/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.1798\n",
            "Epoch 570/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 2.1512\n",
            "Epoch 571/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.1253\n",
            "Epoch 572/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 2.1021\n",
            "Epoch 573/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.0796\n",
            "Epoch 574/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.0539\n",
            "Epoch 575/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.0388\n",
            "Epoch 576/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 2.0084\n",
            "Epoch 577/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.9856\n",
            "Epoch 578/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.9693\n",
            "Epoch 579/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.9439\n",
            "Epoch 580/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 1.9207\n",
            "Epoch 581/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 1.9048\n",
            "Epoch 582/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.8846\n",
            "Epoch 583/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.8616\n",
            "Epoch 584/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 1.8431\n",
            "Epoch 585/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.8234\n",
            "Epoch 586/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.8079\n",
            "Epoch 587/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.7895\n",
            "Epoch 588/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.7710\n",
            "Epoch 589/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.7531\n",
            "Epoch 590/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.7366\n",
            "Epoch 591/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.7228\n",
            "Epoch 592/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.7026\n",
            "Epoch 593/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.6909\n",
            "Epoch 594/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.6756\n",
            "Epoch 595/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.6578\n",
            "Epoch 596/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.6408\n",
            "Epoch 597/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.6275\n",
            "Epoch 598/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.6174\n",
            "Epoch 599/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.5997\n",
            "Epoch 600/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.5868\n",
            "Epoch 601/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.5740\n",
            "Epoch 602/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.5634\n",
            "Epoch 603/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.5485\n",
            "Epoch 604/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.5372\n",
            "Epoch 605/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 1.5255\n",
            "Epoch 606/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.5109\n",
            "Epoch 607/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 1.4996\n",
            "Epoch 608/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 1.4882\n",
            "Epoch 609/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 1.4791\n",
            "Epoch 610/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.4681\n",
            "Epoch 611/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.4576\n",
            "Epoch 612/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.4460\n",
            "Epoch 613/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.4355\n",
            "Epoch 614/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.4272\n",
            "Epoch 615/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.4205\n",
            "Epoch 616/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.4102\n",
            "Epoch 617/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.3972\n",
            "Epoch 618/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.3911\n",
            "Epoch 619/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.3798\n",
            "Epoch 620/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.3720\n",
            "Epoch 621/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.3632\n",
            "Epoch 622/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.3556\n",
            "Epoch 623/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.3469\n",
            "Epoch 624/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 1.3391\n",
            "Epoch 625/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.3325\n",
            "Epoch 626/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.3244\n",
            "Epoch 627/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 1.3181\n",
            "Epoch 628/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.3094\n",
            "Epoch 629/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.3030\n",
            "Epoch 630/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 1.2986\n",
            "Epoch 631/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.2902\n",
            "Epoch 632/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.2824\n",
            "Epoch 633/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.2789\n",
            "Epoch 634/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.2697\n",
            "Epoch 635/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.2665\n",
            "Epoch 636/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.2576\n",
            "Epoch 637/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 1.2531\n",
            "Epoch 638/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.2476\n",
            "Epoch 639/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.2416\n",
            "Epoch 640/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.2365\n",
            "Epoch 641/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.2311\n",
            "Epoch 642/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.2252\n",
            "Epoch 643/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.2215\n",
            "Epoch 644/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.2159\n",
            "Epoch 645/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.2112\n",
            "Epoch 646/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.2058\n",
            "Epoch 647/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.2014\n",
            "Epoch 648/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1978\n",
            "Epoch 649/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1928\n",
            "Epoch 650/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1871\n",
            "Epoch 651/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1842\n",
            "Epoch 652/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1798\n",
            "Epoch 653/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1763\n",
            "Epoch 654/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 1.1721\n",
            "Epoch 655/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1680\n",
            "Epoch 656/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1634\n",
            "Epoch 657/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1611\n",
            "Epoch 658/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1563\n",
            "Epoch 659/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1539\n",
            "Epoch 660/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1513\n",
            "Epoch 661/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1456\n",
            "Epoch 662/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 1.1420\n",
            "Epoch 663/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1389\n",
            "Epoch 664/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1359\n",
            "Epoch 665/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 1.1326\n",
            "Epoch 666/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1299\n",
            "Epoch 667/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1279\n",
            "Epoch 668/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1231\n",
            "Epoch 669/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1203\n",
            "Epoch 670/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1175\n",
            "Epoch 671/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1149\n",
            "Epoch 672/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1123\n",
            "Epoch 673/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1087\n",
            "Epoch 674/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1057\n",
            "Epoch 675/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1032\n",
            "Epoch 676/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.1007\n",
            "Epoch 677/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0982\n",
            "Epoch 678/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0956\n",
            "Epoch 679/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0932\n",
            "Epoch 680/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0907\n",
            "Epoch 681/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0881\n",
            "Epoch 682/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0850\n",
            "Epoch 683/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0829\n",
            "Epoch 684/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 1.0805\n",
            "Epoch 685/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0786\n",
            "Epoch 686/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0765\n",
            "Epoch 687/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0749\n",
            "Epoch 688/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0710\n",
            "Epoch 689/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 1.0689\n",
            "Epoch 690/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0671\n",
            "Epoch 691/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0654\n",
            "Epoch 692/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0625\n",
            "Epoch 693/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0606\n",
            "Epoch 694/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0582\n",
            "Epoch 695/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 1.0570\n",
            "Epoch 696/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0540\n",
            "Epoch 697/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 1.0515\n",
            "Epoch 698/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0500\n",
            "Epoch 699/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0486\n",
            "Epoch 700/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 1.0463\n",
            "Epoch 701/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0443\n",
            "Epoch 702/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0429\n",
            "Epoch 703/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0394\n",
            "Epoch 704/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0384\n",
            "Epoch 705/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0358\n",
            "Epoch 706/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 1.0336\n",
            "Epoch 707/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0319\n",
            "Epoch 708/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0299\n",
            "Epoch 709/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0283\n",
            "Epoch 710/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0276\n",
            "Epoch 711/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0241\n",
            "Epoch 712/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0229\n",
            "Epoch 713/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0206\n",
            "Epoch 714/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0191\n",
            "Epoch 715/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0168\n",
            "Epoch 716/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0149\n",
            "Epoch 717/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0133\n",
            "Epoch 718/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0116\n",
            "Epoch 719/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 1.0094\n",
            "Epoch 720/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0074\n",
            "Epoch 721/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0056\n",
            "Epoch 722/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0041\n",
            "Epoch 723/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0020\n",
            "Epoch 724/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 1.0004\n",
            "Epoch 725/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9992\n",
            "Epoch 726/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9970\n",
            "Epoch 727/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9949\n",
            "Epoch 728/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9929\n",
            "Epoch 729/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9914\n",
            "Epoch 730/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9894\n",
            "Epoch 731/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9878\n",
            "Epoch 732/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.9857\n",
            "Epoch 733/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9841\n",
            "Epoch 734/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9825\n",
            "Epoch 735/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9807\n",
            "Epoch 736/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9791\n",
            "Epoch 737/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9770\n",
            "Epoch 738/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9750\n",
            "Epoch 739/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9731\n",
            "Epoch 740/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9722\n",
            "Epoch 741/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9696\n",
            "Epoch 742/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9685\n",
            "Epoch 743/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9668\n",
            "Epoch 744/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9643\n",
            "Epoch 745/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9628\n",
            "Epoch 746/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9611\n",
            "Epoch 747/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9592\n",
            "Epoch 748/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9571\n",
            "Epoch 749/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.9558\n",
            "Epoch 750/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9533\n",
            "Epoch 751/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9523\n",
            "Epoch 752/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9500\n",
            "Epoch 753/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.9481\n",
            "Epoch 754/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.9471\n",
            "Epoch 755/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9463\n",
            "Epoch 756/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9429\n",
            "Epoch 757/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9415\n",
            "Epoch 758/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9393\n",
            "Epoch 759/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9373\n",
            "Epoch 760/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9357\n",
            "Epoch 761/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.9338\n",
            "Epoch 762/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9322\n",
            "Epoch 763/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.9303\n",
            "Epoch 764/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9288\n",
            "Epoch 765/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9265\n",
            "Epoch 766/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9249\n",
            "Epoch 767/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9234\n",
            "Epoch 768/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9215\n",
            "Epoch 769/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9193\n",
            "Epoch 770/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9183\n",
            "Epoch 771/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9156\n",
            "Epoch 772/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9140\n",
            "Epoch 773/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9117\n",
            "Epoch 774/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9100\n",
            "Epoch 775/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9087\n",
            "Epoch 776/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9066\n",
            "Epoch 777/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.9046\n",
            "Epoch 778/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9032\n",
            "Epoch 779/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.9006\n",
            "Epoch 780/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8990\n",
            "Epoch 781/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.8971\n",
            "Epoch 782/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8953\n",
            "Epoch 783/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8936\n",
            "Epoch 784/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8913\n",
            "Epoch 785/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8904\n",
            "Epoch 786/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8878\n",
            "Epoch 787/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8858\n",
            "Epoch 788/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.8843\n",
            "Epoch 789/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8822\n",
            "Epoch 790/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8800\n",
            "Epoch 791/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8790\n",
            "Epoch 792/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8767\n",
            "Epoch 793/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8749\n",
            "Epoch 794/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8724\n",
            "Epoch 795/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8708\n",
            "Epoch 796/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8689\n",
            "Epoch 797/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8668\n",
            "Epoch 798/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8651\n",
            "Epoch 799/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8633\n",
            "Epoch 800/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8623\n",
            "Epoch 801/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8602\n",
            "Epoch 802/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.8573\n",
            "Epoch 803/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8555\n",
            "Epoch 804/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8538\n",
            "Epoch 805/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8514\n",
            "Epoch 806/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8499\n",
            "Epoch 807/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8475\n",
            "Epoch 808/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8463\n",
            "Epoch 809/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8439\n",
            "Epoch 810/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.8418\n",
            "Epoch 811/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8401\n",
            "Epoch 812/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.8381\n",
            "Epoch 813/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8360\n",
            "Epoch 814/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8341\n",
            "Epoch 815/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8334\n",
            "Epoch 816/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8302\n",
            "Epoch 817/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8284\n",
            "Epoch 818/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8264\n",
            "Epoch 819/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8245\n",
            "Epoch 820/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8225\n",
            "Epoch 821/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8205\n",
            "Epoch 822/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8188\n",
            "Epoch 823/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8165\n",
            "Epoch 824/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8151\n",
            "Epoch 825/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8135\n",
            "Epoch 826/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8112\n",
            "Epoch 827/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8087\n",
            "Epoch 828/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8068\n",
            "Epoch 829/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.8053\n",
            "Epoch 830/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8040\n",
            "Epoch 831/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.8005\n",
            "Epoch 832/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7988\n",
            "Epoch 833/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7969\n",
            "Epoch 834/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7952\n",
            "Epoch 835/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.7933\n",
            "Epoch 836/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7907\n",
            "Epoch 837/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7890\n",
            "Epoch 838/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7872\n",
            "Epoch 839/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7854\n",
            "Epoch 840/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7835\n",
            "Epoch 841/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7808\n",
            "Epoch 842/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7790\n",
            "Epoch 843/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7771\n",
            "Epoch 844/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7750\n",
            "Epoch 845/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7733\n",
            "Epoch 846/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7706\n",
            "Epoch 847/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7696\n",
            "Epoch 848/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7675\n",
            "Epoch 849/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7651\n",
            "Epoch 850/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7626\n",
            "Epoch 851/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7609\n",
            "Epoch 852/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7591\n",
            "Epoch 853/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7568\n",
            "Epoch 854/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7551\n",
            "Epoch 855/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7528\n",
            "Epoch 856/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7512\n",
            "Epoch 857/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7487\n",
            "Epoch 858/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7473\n",
            "Epoch 859/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7446\n",
            "Epoch 860/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.7440\n",
            "Epoch 861/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7410\n",
            "Epoch 862/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7392\n",
            "Epoch 863/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7376\n",
            "Epoch 864/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7348\n",
            "Epoch 865/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7327\n",
            "Epoch 866/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7303\n",
            "Epoch 867/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7291\n",
            "Epoch 868/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7280\n",
            "Epoch 869/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7245\n",
            "Epoch 870/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.7224\n",
            "Epoch 871/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.7204\n",
            "Epoch 872/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7193\n",
            "Epoch 873/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7163\n",
            "Epoch 874/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.7146\n",
            "Epoch 875/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.7124\n",
            "Epoch 876/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7111\n",
            "Epoch 877/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7094\n",
            "Epoch 878/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7069\n",
            "Epoch 879/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7050\n",
            "Epoch 880/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7037\n",
            "Epoch 881/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.7001\n",
            "Epoch 882/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6994\n",
            "Epoch 883/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6970\n",
            "Epoch 884/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6945\n",
            "Epoch 885/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6926\n",
            "Epoch 886/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6898\n",
            "Epoch 887/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6882\n",
            "Epoch 888/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6858\n",
            "Epoch 889/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6837\n",
            "Epoch 890/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6820\n",
            "Epoch 891/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6796\n",
            "Epoch 892/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.6779\n",
            "Epoch 893/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.6759\n",
            "Epoch 894/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6735\n",
            "Epoch 895/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6720\n",
            "Epoch 896/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6697\n",
            "Epoch 897/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6682\n",
            "Epoch 898/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6659\n",
            "Epoch 899/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.6637\n",
            "Epoch 900/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6625\n",
            "Epoch 901/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6596\n",
            "Epoch 902/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6577\n",
            "Epoch 903/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6555\n",
            "Epoch 904/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.6538\n",
            "Epoch 905/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6513\n",
            "Epoch 906/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.6499\n",
            "Epoch 907/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6474\n",
            "Epoch 908/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6451\n",
            "Epoch 909/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.6435\n",
            "Epoch 910/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6418\n",
            "Epoch 911/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6393\n",
            "Epoch 912/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6380\n",
            "Epoch 913/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6352\n",
            "Epoch 914/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6332\n",
            "Epoch 915/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6310\n",
            "Epoch 916/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6295\n",
            "Epoch 917/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6277\n",
            "Epoch 918/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.6248\n",
            "Epoch 919/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6229\n",
            "Epoch 920/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6212\n",
            "Epoch 921/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6206\n",
            "Epoch 922/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6167\n",
            "Epoch 923/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6154\n",
            "Epoch 924/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6137\n",
            "Epoch 925/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6111\n",
            "Epoch 926/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6096\n",
            "Epoch 927/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6076\n",
            "Epoch 928/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6052\n",
            "Epoch 929/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.6039\n",
            "Epoch 930/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.6006\n",
            "Epoch 931/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5992\n",
            "Epoch 932/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5970\n",
            "Epoch 933/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5950\n",
            "Epoch 934/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5928\n",
            "Epoch 935/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5906\n",
            "Epoch 936/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.5896\n",
            "Epoch 937/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5868\n",
            "Epoch 938/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5846\n",
            "Epoch 939/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5841\n",
            "Epoch 940/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5813\n",
            "Epoch 941/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5800\n",
            "Epoch 942/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5770\n",
            "Epoch 943/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5749\n",
            "Epoch 944/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5732\n",
            "Epoch 945/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5709\n",
            "Epoch 946/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5686\n",
            "Epoch 947/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5670\n",
            "Epoch 948/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5649\n",
            "Epoch 949/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5628\n",
            "Epoch 950/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5607\n",
            "Epoch 951/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5593\n",
            "Epoch 952/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.5570\n",
            "Epoch 953/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5555\n",
            "Epoch 954/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5530\n",
            "Epoch 955/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5515\n",
            "Epoch 956/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5498\n",
            "Epoch 957/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5475\n",
            "Epoch 958/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5450\n",
            "Epoch 959/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5434\n",
            "Epoch 960/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5414\n",
            "Epoch 961/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5395\n",
            "Epoch 962/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5374\n",
            "Epoch 963/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5356\n",
            "Epoch 964/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5332\n",
            "Epoch 965/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5317\n",
            "Epoch 966/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5297\n",
            "Epoch 967/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5275\n",
            "Epoch 968/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5260\n",
            "Epoch 969/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5235\n",
            "Epoch 970/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5218\n",
            "Epoch 971/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5195\n",
            "Epoch 972/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5180\n",
            "Epoch 973/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5159\n",
            "Epoch 974/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5137\n",
            "Epoch 975/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5120\n",
            "Epoch 976/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5098\n",
            "Epoch 977/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5079\n",
            "Epoch 978/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5063\n",
            "Epoch 979/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5052\n",
            "Epoch 980/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5025\n",
            "Epoch 981/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.5018\n",
            "Epoch 982/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4993\n",
            "Epoch 983/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4968\n",
            "Epoch 984/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4946\n",
            "Epoch 985/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4927\n",
            "Epoch 986/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4906\n",
            "Epoch 987/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4888\n",
            "Epoch 988/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4870\n",
            "Epoch 989/1000\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.4856\n",
            "Epoch 990/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4835\n",
            "Epoch 991/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4829\n",
            "Epoch 992/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4791\n",
            "Epoch 993/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4779\n",
            "Epoch 994/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4766\n",
            "Epoch 995/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4735\n",
            "Epoch 996/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4716\n",
            "Epoch 997/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4700\n",
            "Epoch 998/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4683\n",
            "Epoch 999/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4667\n",
            "Epoch 1000/1000\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4660\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tensorflow.python.keras.callbacks.History at 0x7fe2600a7588>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NrykycJCEcS5",
        "colab_type": "text"
      },
      "source": [
        "위의 학습 과정을 보면, epoch이 진행되기는 하지만 loss에 큰 변화가 없어보이네요. \n",
        "\n",
        "여러분, 어떻게 하면 좋을까요 ?\n",
        "\n",
        "일단 학습된 모델을 가지고 새로운 data (20)에 대한 outcome을 예측해보도록 하겠습니다."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zdOx5lZuQfBE",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "58495751-a34b-4416-eb91-88bb49e4eece"
      },
      "source": [
        "model.predict([20])"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[38.160328]], dtype=float32)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9oEwdQTsEyUw",
        "colab_type": "text"
      },
      "source": [
        "답을 41로 예상을 하였는데, 생뚱 맞게 -6 정도가 나와 버렸네요.\n",
        "\n",
        "학습을 제대로 하지 못한 경우겠지요 !\n",
        "\n",
        "새로운 **epochs**과 **batch_size**로 학습을 다시 해보도록 하겠습니다. 여러분도 다양한 값들을 시도해보세요!"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3G21DPjaCjrg",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 386
        },
        "outputId": "f4605319-204d-4061-8eff-e98bb4ca84af"
      },
      "source": [
        "model.fit(data, outcome, epochs=10, batch_size=1)"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/10\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4622\n",
            "Epoch 2/10\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 0.4607\n",
            "Epoch 3/10\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4596\n",
            "Epoch 4/10\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4567\n",
            "Epoch 5/10\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4547\n",
            "Epoch 6/10\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4534\n",
            "Epoch 7/10\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4514\n",
            "Epoch 8/10\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4491\n",
            "Epoch 9/10\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4480\n",
            "Epoch 10/10\n",
            "10/10 [==============================] - 0s 1ms/step - loss: 0.4453\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tensorflow.python.keras.callbacks.History at 0x7fe205f0c5c0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ea51HWv5JBuq",
        "colab_type": "text"
      },
      "source": [
        "Epoch이 진행되면서 지속적으로 loss도 떨어져서 낮은 loss를 갖도록 학습을 하였습니다. \n",
        "\n",
        "그럼 얼마나 잘 예측하는지 새로운 데이터를 가지고 예측을 해볼까요 ?"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cC-2_dN5JPGC",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "2696ba68-3ec1-41b4-ac11-69abce68c308"
      },
      "source": [
        "model.predict([20])"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[38.208874]], dtype=float32)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jTX2e84eJatK",
        "colab_type": "text"
      },
      "source": [
        "아주 잘 예측하고 있음을 알 수 있습니다!"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "A74h3bahPVmb",
        "colab_type": "text"
      },
      "source": [
        "\n",
        "\n",
        "---\n",
        "\n",
        "\n",
        "마지막으로 **가중치 (w)**와 **바이어스(b)**가 어떻게 찾아졌는지 살펴보도록 하겠습니다.\n",
        "\n",
        "model.layers[n].get_weights()[0] : n 번째 layer의 가중차 값을 반환합니다.\n",
        "\n",
        "model.layers[n].get_weights()[1]: n번째 layer의 바이어스 값을 반환합니다."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "d0j0osfGP9bc",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        },
        "outputId": "860a948c-20c9-4167-e823-abfaae62659a"
      },
      "source": [
        "layer1_weight = model.layers[0].get_weights()[0]   # Weight 값 가져오기\n",
        "layer1_bias = model.layers[0].get_weights()[1]     # Bias 값 가져오기\n",
        "\n",
        "print(layer1_weight)\n",
        "print(layer1_bias)"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[1.7885612]]\n",
            "[2.4376488]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0cogLhNLQTiE",
        "colab_type": "text"
      },
      "source": [
        "우리의 예상은 layer1_weight는 2, layer1_bias는 1이기를 기대하였지만, 조금 다른 값이 나온 것을 알 수 있습니다.\n",
        "\n",
        "하지만, 꽤 잘 찾은 것 같지 않나요 ? ^^"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UEp4GP-6QKjt",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 10,
      "outputs": []
    }
  ]
}