230704_선형회귀.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "authorship_tag": "ABX9TyPYjM2iokQnZqDCcJZ/liKk",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/khwwang/coding-training/blob/master/230704_%EC%84%A0%ED%98%95%ED%9A%8C%EA%B7%80.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 선형회귀"
      ],
      "metadata": {
        "id": "Nc6yfgF8PpD9"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "단순선형회귀: 독립변수가 1개인 선형 회귀 -> 독립변수가 여러개면 다중선형회귀이다, x^2 + x^3처럼 제곱이 되어있는 것은 곡선회귀이므로 선형이 아니다\n",
        "\n",
        "> 데이터를 잘 나타내는 직선은 예측값과 실제값의 차이(error)->최소화\n",
        "잔차 제곱합 rss, mse등을 사용\n",
        "\n",
        "> 회귀함수는 이러한 cost function을 최소가 되는 파라미터를 찾는 알고리즘이다\n",
        "\n",
        "> 최소제곱법(sklearn의 linear regression에서는 normal equation을 이용)\n",
        "\n",
        "> 경사 하강법: 점진적 학습\n",
        "mse에 대한 그래프(error에 대한)에서 미분하여 가장 작은 것을 찾는 것\n",
        "기울기(gradient)가 작아지는 방향으로 진행\n",
        "\n",
        "> 학습률: 얼마만큼 학습해야 하는지 학습 양을 의미하며 한 번의 학습량으로 학습한 이후에 가중치 매개변수가 갱신된다\n",
        "- 가중치 업데이트: 기울기를 계산한 후에 기울기 값을 기존 가중치(기울기)에 더하거나 빼서 가중치를 업데이트하는데 이때 계산된 기울기에 미리 설정해둔 학습률을 곱해서 나온 값으로 가중치를 업데이트하도록 한다\n",
        "\n",
        "> convex funtion 볼록함수이고 매우 이상정인 cost function이라고 할 수 있다\n",
        "\n",
        "> gradient 종류\n",
        "- batch : 전체의 데이터를 이용\n",
        "- stochastic : 램덤으로 하나의 뽑아서 이용\n",
        "- mini batch : 몇개를 모아서 그룹을 설정한 데이터를 이용"
      ],
      "metadata": {
        "id": "g4oFEGN0PpGQ"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "다항회귀(polynomial regression) 직선이 아닐 경우 -> 곡선회귀와 같은말\n",
        "- 문제점: 고차 다항식으로 회귀를 할대 과적합이 일어나기 쉽다(모델이 너무 복잡)\n",
        "- 모델의 규제를 통해 차수를 줄인다\n",
        "\n",
        "> 규제: ridge, lasso등의 규제가 있다\n",
        "- 한번 자세히 살펴보자"
      ],
      "metadata": {
        "id": "dunVrHShcBHn"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 로지스틱 회귀"
      ],
      "metadata": {
        "id": "7V0YzlkomNR9"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "비선형이고 시그모이드 함수를 통해 cost function을 구한다\n",
        "확률을 이용한 cost function -> log likelihood\n",
        "\n",
        "- cross entropy를 구하여 두 확률 분포 사이의 오차를 측정![로지스틱 cost.png](data:image/png;base64,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)"
      ],
      "metadata": {
        "id": "lGcb1nO0mPqh"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 다중 분류(Categorical Cross Entropy)"
      ],
      "metadata": {
        "id": "be2wEBO8vrUI"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "soft max(다중 클래스 분류의 경우) 확률이 가장 높은 것을 정답으로 1로 한다\n",
        "- sigmoid와 다른 점은 그저 확률로만 나오는 것이 아니라 정답이 나온다(one hot encoding)\n",
        "- cross entropy를 cost로 사용한다![soft max cost.png](data:image/png;base64,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)"
      ],
      "metadata": {
        "id": "cX-evBrOsxmz"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Neural network"
      ],
      "metadata": {
        "id": "rlLixuHCsx0o"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### perceptron\n",
        "이진분류 -> 선형분류\n",
        "- activation function: step funtion\n",
        "\n",
        "> mlp(multilayered perceptron)\n",
        "학습할때 w값을 찾기가 어렵다\n",
        "\n",
        "- 오차 역전파(backpropagation)을 통하여 weight를 갱신시키는 방법\n",
        "- 순전파(forwardpropagation)을 통해 계산(미분값도)하고 반대로 오면서 chain rule을 이용하여 weight를 갱신하는 방법으로 역전파를 사용\n",
        "가지고있는 미분값(gradient 계산)을 통해 가중치 업데이트\n",
        "![image.png](data:image/png;base64,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)"
      ],
      "metadata": {
        "id": "C5T-1hYRsx7W"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "깊은 신경망의 경우 입력층으로 갈수록 역전파 과정에서 미분값이 사라지면서 학습이 안되는 현상"
      ],
      "metadata": {
        "id": "H7_ifxSCpE5F"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "optimizer에서 gd에서 발전한 것이 sgd이고 기본적으로 성능이 좋은 adam을 사용한다\n",
        "\n",
        "- cross validation: 교차검증은 검증을 데이터를 나눠서 진해앟고 평균을 내는 것인데 딥러닝에서는 시간이 오래 걸려서 많이 사용하지는 않는다\n",
        "\n",
        "- 데이터 증강: 데이터가 없으면 과소적합이나 과적합이 생기므로 기존의 데이터를 변형한다(축소, 확대, 자르기 등)\n",
        "\n",
        "- dropout: 모델의 다양성을 만들기 위한 방법으로 이해 (앙상블)\n",
        "\n",
        "- 가중치 초기값 설정, 초기화(Xaiver Glorot 초기화, He 초기화)\n",
        "\n",
        "- 배치 정규화 (Batch Normalization): 미니배치 단위 Output을 표준정규분포로 정규화하는 작업"
      ],
      "metadata": {
        "id": "TxjeJbsV2i-q"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import tensorflow as tf\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt"
      ],
      "metadata": {
        "id": "zqNAlSokTc3t"
      },
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "tf.__version__"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "id": "08ePjwp6UT53",
        "outputId": "2eda5b21-53b8-4982-c9db-f49b7526f692"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'2.12.0'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 2
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "X_train =np.array([1., 2., 3., 4., 5., 6.])\n",
        "y_train = np.array([9.,12.,15.,18.,21.,24.])"
      ],
      "metadata": {
        "id": "paHs7Bs_UWcz"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# 파라미터 초기화\n",
        "W = tf.Variable(0.0)\n",
        "b  =tf.Variable(0.0)\n",
        "\n",
        "# 가설\n",
        "y_hat = W * X_train + b\n",
        "\n",
        "cost = tf.reduce_mean(tf.square(y_train - y_hat))"
      ],
      "metadata": {
        "id": "pPXe-6dVU3eC"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# 학습율\n",
        "lr = 0.01\n",
        "\n",
        "with tf.GradientTape() as tape:\n",
        "  y_hat = W * X_train + b\n",
        "  cost = tf.reduce_mean(tf.square(y_train - y_hat))\n",
        "\n",
        "W_grad, b_grad = tape.gradient(cost, [W, b])\n",
        "# update\n",
        "W.assign_sub(lr*W_grad)\n",
        "b.assign_sub(lr*b_grad)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FI-jfAOnWCOJ",
        "outputId": "f7afdd1c-8aef-4244-e24c-36457e94a56a"
      },
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Variable 'UnreadVariable' shape=() dtype=float32, numpy=0.32999998>"
            ]
          },
          "metadata": {},
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 파라미터 초기화\n",
        "W = tf.Variable(0.0)\n",
        "b  =tf.Variable(0.0)\n",
        "\n",
        "# epoch돌리기\n",
        "epochs = 5000\n",
        "lr = 0.01\n",
        "\n",
        "for i in range(epochs):\n",
        "  with tf.GradientTape() as tape:\n",
        "    y_hat = W * X_train + b\n",
        "    cost = tf.reduce_mean(tf.square(y_train - y_hat))\n",
        "\n",
        "    W_grad, b_grad = tape.gradient(cost, [W, b])\n",
        "    # update\n",
        "    W.assign_sub(lr*W_grad)\n",
        "    b.assign_sub(lr*b_grad)\n",
        "\n",
        "    if i%200 == 0:\n",
        "      print('Epoch:', i , 'cost:', cost.numpy(), 'W:', W.numpy(), 'b:', b.numpy())"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KS07gg-YX3Lw",
        "outputId": "a0e452d4-57c0-4e86-9340-9c792629bcc2"
      },
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch: 0 cost: 298.5 W: 1.3299999 b: 0.32999998\n",
            "Epoch: 200 cost: 1.1257037 W: 3.5628922 b: 3.5901465\n",
            "Epoch: 400 cost: 0.26079097 W: 3.2709315 b: 4.840088\n",
            "Epoch: 600 cost: 0.06041649 W: 3.1304042 b: 5.441714\n",
            "Epoch: 800 cost: 0.013996857 W: 3.0627668 b: 5.7312837\n",
            "Epoch: 1000 cost: 0.0032424245 W: 3.03021 b: 5.870665\n",
            "Epoch: 1200 cost: 0.0007511492 W: 3.0145404 b: 5.9377494\n",
            "Epoch: 1400 cost: 0.00017403124 W: 3.006999 b: 5.970036\n",
            "Epoch: 1600 cost: 4.0324994e-05 W: 3.003369 b: 5.9855766\n",
            "Epoch: 1800 cost: 9.340783e-06 W: 3.0016217 b: 5.9930577\n",
            "Epoch: 2000 cost: 2.1644023e-06 W: 3.0007808 b: 5.9966583\n",
            "Epoch: 2200 cost: 5.027193e-07 W: 3.000376 b: 5.9983897\n",
            "Epoch: 2400 cost: 1.173236e-07 W: 3.0001814 b: 5.9992223\n",
            "Epoch: 2600 cost: 2.7551778e-08 W: 3.000088 b: 5.9996233\n",
            "Epoch: 2800 cost: 6.3532752e-09 W: 3.0000422 b: 5.9998183\n",
            "Epoch: 3000 cost: 1.4401849e-09 W: 3.0000205 b: 5.9999137\n",
            "Epoch: 3200 cost: 8.8220986e-10 W: 3.0000157 b: 5.999933\n",
            "Epoch: 3400 cost: 8.8220986e-10 W: 3.0000157 b: 5.999933\n",
            "Epoch: 3600 cost: 8.8220986e-10 W: 3.0000157 b: 5.999933\n",
            "Epoch: 3800 cost: 8.8220986e-10 W: 3.0000157 b: 5.999933\n",
            "Epoch: 4000 cost: 8.8220986e-10 W: 3.0000157 b: 5.999933\n",
            "Epoch: 4200 cost: 8.8220986e-10 W: 3.0000157 b: 5.999933\n",
            "Epoch: 4400 cost: 8.8220986e-10 W: 3.0000157 b: 5.999933\n",
            "Epoch: 4600 cost: 8.8220986e-10 W: 3.0000157 b: 5.999933\n",
            "Epoch: 4800 cost: 8.8220986e-10 W: 3.0000157 b: 5.999933\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Fully connected layer"
      ],
      "metadata": {
        "id": "JgGkcdDCYqC-"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import tensorflow as tf\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt"
      ],
      "metadata": {
        "id": "uN9HrweoasAf"
      },
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X_train =np.array([1., 2., 3., 4., 5., 6.])\n",
        "y_train = np.array([9.,12.,15.,18.,21.,24.])"
      ],
      "metadata": {
        "id": "DKMCSYVqbQaT"
      },
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow import keras\n",
        "from keras import layers\n",
        "\n",
        "model = keras.Sequential(\n",
        "    layers.Dense(units=1, input_shape=(1,)),\n",
        ")"
      ],
      "metadata": {
        "id": "YNNRY43-bRE4"
      },
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from keras.engine.training import optimizer\n",
        "# optimizer: sgd, adam\n",
        "# cost: mse\n",
        "# metrics: mse, mae\n",
        "model.compile(\n",
        "    optimizer='sgd',\n",
        "    loss='mse',\n",
        "    metrics=['mse','mae']\n",
        ")"
      ],
      "metadata": {
        "id": "IUx-LTLTdTsg"
      },
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model.fit(\n",
        "    X_train, y_train,\n",
        "    epochs=1000,\n",
        "    batch_size=3,\n",
        "    verbose=1\n",
        ")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hXGlJILjd_RY",
        "outputId": "8fa34cd8-24be-4010-bc8b-a5f9a3d8c4fe"
      },
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/1000\n",
            "2/2 [==============================] - 1s 11ms/step - loss: 326.4970 - mse: 326.4970 - mae: 17.1785\n",
            "Epoch 2/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 72.3549 - mse: 72.3549 - mae: 8.3215\n",
            "Epoch 3/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 19.0105 - mse: 19.0105 - mae: 4.2986\n",
            "Epoch 4/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 7.3225 - mse: 7.3225 - mae: 2.3397\n",
            "Epoch 5/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 4.9365 - mse: 4.9365 - mae: 1.8142\n",
            "Epoch 6/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 4.2451 - mse: 4.2451 - mae: 1.7321\n",
            "Epoch 7/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 4.0879 - mse: 4.0879 - mae: 1.7158\n",
            "Epoch 8/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 4.2458 - mse: 4.2458 - mae: 1.7860\n",
            "Epoch 9/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 4.0011 - mse: 4.0011 - mae: 1.6707\n",
            "Epoch 10/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 4.1744 - mse: 4.1744 - mae: 1.7839\n",
            "Epoch 11/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.8816 - mse: 3.8816 - mae: 1.6535\n",
            "Epoch 12/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 4.0076 - mse: 4.0076 - mae: 1.7294\n",
            "Epoch 13/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.7087 - mse: 3.7087 - mae: 1.6368\n",
            "Epoch 14/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.8789 - mse: 3.8789 - mae: 1.7096\n",
            "Epoch 15/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.9480 - mse: 3.9480 - mae: 1.6871\n",
            "Epoch 16/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 4.0055 - mse: 4.0055 - mae: 1.7206\n",
            "Epoch 17/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.5473 - mse: 3.5473 - mae: 1.5764\n",
            "Epoch 18/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.8518 - mse: 3.8518 - mae: 1.6719\n",
            "Epoch 19/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.5509 - mse: 3.5509 - mae: 1.6014\n",
            "Epoch 20/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.3441 - mse: 3.3441 - mae: 1.5524\n",
            "Epoch 21/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.5317 - mse: 3.5317 - mae: 1.6480\n",
            "Epoch 22/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.4811 - mse: 3.4811 - mae: 1.6411\n",
            "Epoch 23/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.4334 - mse: 3.4334 - mae: 1.6313\n",
            "Epoch 24/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.3846 - mse: 3.3846 - mae: 1.6201\n",
            "Epoch 25/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.2852 - mse: 3.2852 - mae: 1.5843\n",
            "Epoch 26/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.0809 - mse: 3.0809 - mae: 1.4840\n",
            "Epoch 27/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.3665 - mse: 3.3665 - mae: 1.5624\n",
            "Epoch 28/1000\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 3.3048 - mse: 3.3048 - mae: 1.5514\n",
            "Epoch 29/1000\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 2.9649 - mse: 2.9649 - mae: 1.4558\n",
            "Epoch 30/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.0328 - mse: 3.0328 - mae: 1.4894\n",
            "Epoch 31/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.8559 - mse: 2.8559 - mae: 1.4224\n",
            "Epoch 32/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.8152 - mse: 2.8152 - mae: 1.4112\n",
            "Epoch 33/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.9064 - mse: 2.9064 - mae: 1.4863\n",
            "Epoch 34/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.7318 - mse: 2.7318 - mae: 1.3889\n",
            "Epoch 35/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.8620 - mse: 2.8620 - mae: 1.4610\n",
            "Epoch 36/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.6824 - mse: 2.6824 - mae: 1.3749\n",
            "Epoch 37/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.6294 - mse: 2.6294 - mae: 1.3602\n",
            "Epoch 38/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.5487 - mse: 2.5487 - mae: 1.3549\n",
            "Epoch 39/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 2.5536 - mse: 2.5536 - mae: 1.3431\n",
            "Epoch 40/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.5189 - mse: 2.5189 - mae: 1.3371\n",
            "Epoch 41/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 2.5987 - mse: 2.5987 - mae: 1.4057\n",
            "Epoch 42/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 2.5387 - mse: 2.5387 - mae: 1.3901\n",
            "Epoch 43/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.4096 - mse: 2.4096 - mae: 1.3077\n",
            "Epoch 44/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.3262 - mse: 2.3262 - mae: 1.2935\n",
            "Epoch 45/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.3009 - mse: 2.3009 - mae: 1.2882\n",
            "Epoch 46/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.5520 - mse: 2.5520 - mae: 1.3683\n",
            "Epoch 47/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.3357 - mse: 2.3357 - mae: 1.3053\n",
            "Epoch 48/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.2357 - mse: 2.2357 - mae: 1.2485\n",
            "Epoch 49/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.1622 - mse: 2.1622 - mae: 1.2457\n",
            "Epoch 50/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 2.1359 - mse: 2.1359 - mae: 1.2380\n",
            "Epoch 51/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 2.2393 - mse: 2.2393 - mae: 1.2932\n",
            "Epoch 52/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.1882 - mse: 2.1882 - mae: 1.2905\n",
            "Epoch 53/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 2.2819 - mse: 2.2819 - mae: 1.2900\n",
            "Epoch 54/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.0091 - mse: 2.0091 - mae: 1.2043\n",
            "Epoch 55/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.0138 - mse: 2.0138 - mae: 1.1954\n",
            "Epoch 56/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.9512 - mse: 1.9512 - mae: 1.1881\n",
            "Epoch 57/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.9239 - mse: 1.9239 - mae: 1.1756\n",
            "Epoch 58/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.9249 - mse: 1.9249 - mae: 1.1590\n",
            "Epoch 59/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.9009 - mse: 1.9009 - mae: 1.1524\n",
            "Epoch 60/1000\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 1.8577 - mse: 1.8577 - mae: 1.1466\n",
            "Epoch 61/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.8161 - mse: 1.8161 - mae: 1.1440\n",
            "Epoch 62/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.7889 - mse: 1.7889 - mae: 1.1356\n",
            "Epoch 63/1000\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 1.8632 - mse: 1.8632 - mae: 1.1890\n",
            "Epoch 64/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.8576 - mse: 1.8576 - mae: 1.1801\n",
            "Epoch 65/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.7362 - mse: 1.7362 - mae: 1.1017\n",
            "Epoch 66/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.6999 - mse: 1.6999 - mae: 1.0966\n",
            "Epoch 67/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.6876 - mse: 1.6876 - mae: 1.0839\n",
            "Epoch 68/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 1.6371 - mse: 1.6371 - mae: 1.0823\n",
            "Epoch 69/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.7120 - mse: 1.7120 - mae: 1.1294\n",
            "Epoch 70/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.6946 - mse: 1.6946 - mae: 1.1271\n",
            "Epoch 71/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.5913 - mse: 1.5913 - mae: 1.0522\n",
            "Epoch 72/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.5679 - mse: 1.5679 - mae: 1.0444\n",
            "Epoch 73/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.6118 - mse: 1.6118 - mae: 1.1006\n",
            "Epoch 74/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.5210 - mse: 1.5210 - mae: 1.0286\n",
            "Epoch 75/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.4754 - mse: 1.4754 - mae: 1.0328\n",
            "Epoch 76/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.5228 - mse: 1.5228 - mae: 1.0688\n",
            "Epoch 77/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.6184 - mse: 1.6184 - mae: 1.0968\n",
            "Epoch 78/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.4159 - mse: 1.4159 - mae: 1.0025\n",
            "Epoch 79/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 1.3874 - mse: 1.3874 - mae: 0.9980\n",
            "Epoch 80/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 1.3629 - mse: 1.3629 - mae: 0.9903\n",
            "Epoch 81/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.3432 - mse: 1.3432 - mae: 0.9834\n",
            "Epoch 82/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.3286 - mse: 1.3286 - mae: 0.9779\n",
            "Epoch 83/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.3078 - mse: 1.3078 - mae: 0.9718\n",
            "Epoch 84/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.2857 - mse: 1.2857 - mae: 0.9623\n",
            "Epoch 85/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.4390 - mse: 1.4390 - mae: 1.0347\n",
            "Epoch 86/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.3276 - mse: 1.3276 - mae: 0.9909\n",
            "Epoch 87/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.3896 - mse: 1.3896 - mae: 1.0110\n",
            "Epoch 88/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.2410 - mse: 1.2410 - mae: 0.9249\n",
            "Epoch 89/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.2214 - mse: 1.2214 - mae: 0.9277\n",
            "Epoch 90/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.3195 - mse: 1.3195 - mae: 0.9783\n",
            "Epoch 91/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 1.2517 - mse: 1.2517 - mae: 0.9856\n",
            "Epoch 92/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.1709 - mse: 1.1709 - mae: 0.8969\n",
            "Epoch 93/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.1377 - mse: 1.1377 - mae: 0.8958\n",
            "Epoch 94/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.1256 - mse: 1.1256 - mae: 0.8872\n",
            "Epoch 95/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.0900 - mse: 1.0900 - mae: 0.8857\n",
            "Epoch 96/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.0782 - mse: 1.0782 - mae: 0.8819\n",
            "Epoch 97/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.0627 - mse: 1.0627 - mae: 0.8755\n",
            "Epoch 98/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.1232 - mse: 1.1232 - mae: 0.9151\n",
            "Epoch 99/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.0406 - mse: 1.0406 - mae: 0.8571\n",
            "Epoch 100/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 1.0129 - mse: 1.0129 - mae: 0.8534\n",
            "Epoch 101/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.0156 - mse: 1.0156 - mae: 0.8425\n",
            "Epoch 102/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.9857 - mse: 0.9857 - mae: 0.8434\n",
            "Epoch 103/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.9813 - mse: 0.9813 - mae: 0.8343\n",
            "Epoch 104/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.0280 - mse: 1.0280 - mae: 0.8755\n",
            "Epoch 105/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.0301 - mse: 1.0301 - mae: 0.8593\n",
            "Epoch 106/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.9379 - mse: 0.9379 - mae: 0.8151\n",
            "Epoch 107/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.9169 - mse: 0.9169 - mae: 0.8129\n",
            "Epoch 108/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.9033 - mse: 0.9033 - mae: 0.8070\n",
            "Epoch 109/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.9336 - mse: 0.9336 - mae: 0.8378\n",
            "Epoch 110/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.9291 - mse: 0.9291 - mae: 0.8369\n",
            "Epoch 111/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.8614 - mse: 0.8614 - mae: 0.7857\n",
            "Epoch 112/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.8584 - mse: 0.8584 - mae: 0.7796\n",
            "Epoch 113/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.9216 - mse: 0.9216 - mae: 0.8166\n",
            "Epoch 114/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.8796 - mse: 0.8796 - mae: 0.8111\n",
            "Epoch 115/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.8268 - mse: 0.8268 - mae: 0.7660\n",
            "Epoch 116/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.8478 - mse: 0.8478 - mae: 0.8016\n",
            "Epoch 117/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.8471 - mse: 0.8471 - mae: 0.8093\n",
            "Epoch 118/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.8133 - mse: 0.8133 - mae: 0.7724\n",
            "Epoch 119/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.7786 - mse: 0.7786 - mae: 0.7430\n",
            "Epoch 120/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.8096 - mse: 0.8096 - mae: 0.7896\n",
            "Epoch 121/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.7465 - mse: 0.7465 - mae: 0.7340\n",
            "Epoch 122/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.7425 - mse: 0.7425 - mae: 0.7208\n",
            "Epoch 123/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.7339 - mse: 0.7339 - mae: 0.7158\n",
            "Epoch 124/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.7217 - mse: 0.7217 - mae: 0.7136\n",
            "Epoch 125/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.7128 - mse: 0.7128 - mae: 0.7056\n",
            "Epoch 126/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.6912 - mse: 0.6912 - mae: 0.7059\n",
            "Epoch 127/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.6802 - mse: 0.6802 - mae: 0.7015\n",
            "Epoch 128/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.7042 - mse: 0.7042 - mae: 0.7282\n",
            "Epoch 129/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.6716 - mse: 0.6716 - mae: 0.6904\n",
            "Epoch 130/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.6513 - mse: 0.6513 - mae: 0.6853\n",
            "Epoch 131/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.6393 - mse: 0.6393 - mae: 0.6779\n",
            "Epoch 132/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.6315 - mse: 0.6315 - mae: 0.6746\n",
            "Epoch 133/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.6208 - mse: 0.6208 - mae: 0.6684\n",
            "Epoch 134/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.6118 - mse: 0.6118 - mae: 0.6635\n",
            "Epoch 135/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.6044 - mse: 0.6044 - mae: 0.6612\n",
            "Epoch 136/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.6311 - mse: 0.6311 - mae: 0.6913\n",
            "Epoch 137/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.6148 - mse: 0.6148 - mae: 0.6753\n",
            "Epoch 138/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.5833 - mse: 0.5833 - mae: 0.6426\n",
            "Epoch 139/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.6335 - mse: 0.6335 - mae: 0.6791\n",
            "Epoch 140/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.5611 - mse: 0.5611 - mae: 0.6368\n",
            "Epoch 141/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.5585 - mse: 0.5585 - mae: 0.6294\n",
            "Epoch 142/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.5542 - mse: 0.5542 - mae: 0.6269\n",
            "Epoch 143/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.5451 - mse: 0.5451 - mae: 0.6201\n",
            "Epoch 144/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.5275 - mse: 0.5275 - mae: 0.6163\n",
            "Epoch 145/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.5788 - mse: 0.5788 - mae: 0.6489\n",
            "Epoch 146/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.5229 - mse: 0.5229 - mae: 0.6045\n",
            "Epoch 147/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.5431 - mse: 0.5431 - mae: 0.6460\n",
            "Epoch 148/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.5056 - mse: 0.5056 - mae: 0.5948\n",
            "Epoch 149/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.5419 - mse: 0.5419 - mae: 0.6269\n",
            "Epoch 150/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.5129 - mse: 0.5129 - mae: 0.6219\n",
            "Epoch 151/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.5104 - mse: 0.5104 - mae: 0.6181\n",
            "Epoch 152/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.4978 - mse: 0.4978 - mae: 0.6068\n",
            "Epoch 153/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.4964 - mse: 0.4964 - mae: 0.6171\n",
            "Epoch 154/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.4978 - mse: 0.4978 - mae: 0.5980\n",
            "Epoch 155/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.4560 - mse: 0.4560 - mae: 0.5687\n",
            "Epoch 156/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.4635 - mse: 0.4635 - mae: 0.5895\n",
            "Epoch 157/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.4430 - mse: 0.4430 - mae: 0.5563\n",
            "Epoch 158/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.4323 - mse: 0.4323 - mae: 0.5532\n",
            "Epoch 159/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.4542 - mse: 0.4542 - mae: 0.5894\n",
            "Epoch 160/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.4420 - mse: 0.4420 - mae: 0.5754\n",
            "Epoch 161/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.4167 - mse: 0.4167 - mae: 0.5425\n",
            "Epoch 162/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.4081 - mse: 0.4081 - mae: 0.5372\n",
            "Epoch 163/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.3980 - mse: 0.3980 - mae: 0.5362\n",
            "Epoch 164/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.3921 - mse: 0.3921 - mae: 0.5325\n",
            "Epoch 165/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.4154 - mse: 0.4154 - mae: 0.5640\n",
            "Epoch 166/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.3806 - mse: 0.3806 - mae: 0.5218\n",
            "Epoch 167/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.4027 - mse: 0.4027 - mae: 0.5482\n",
            "Epoch 168/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.3973 - mse: 0.3973 - mae: 0.5545\n",
            "Epoch 169/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.3665 - mse: 0.3665 - mae: 0.5097\n",
            "Epoch 170/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.3592 - mse: 0.3592 - mae: 0.5075\n",
            "Epoch 171/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.3538 - mse: 0.3538 - mae: 0.5052\n",
            "Epoch 172/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.3484 - mse: 0.3484 - mae: 0.5020\n",
            "Epoch 173/1000\n",
            "2/2 [==============================] - 0s 19ms/step - loss: 0.3650 - mse: 0.3650 - mae: 0.5214\n",
            "Epoch 174/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.3619 - mse: 0.3619 - mae: 0.5272\n",
            "Epoch 175/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.3334 - mse: 0.3334 - mae: 0.4910\n",
            "Epoch 176/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.3317 - mse: 0.3317 - mae: 0.4851\n",
            "Epoch 177/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.3234 - mse: 0.3234 - mae: 0.4825\n",
            "Epoch 178/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.3179 - mse: 0.3179 - mae: 0.4783\n",
            "Epoch 179/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.3144 - mse: 0.3144 - mae: 0.4762\n",
            "Epoch 180/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.3490 - mse: 0.3490 - mae: 0.5064\n",
            "Epoch 181/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.3192 - mse: 0.3192 - mae: 0.4827\n",
            "Epoch 182/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.3224 - mse: 0.3224 - mae: 0.4904\n",
            "Epoch 183/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.2971 - mse: 0.2971 - mae: 0.4599\n",
            "Epoch 184/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.2958 - mse: 0.2958 - mae: 0.4548\n",
            "Epoch 185/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.2924 - mse: 0.2924 - mae: 0.4519\n",
            "Epoch 186/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.2874 - mse: 0.2874 - mae: 0.4504\n",
            "Epoch 187/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.2788 - mse: 0.2788 - mae: 0.4490\n",
            "Epoch 188/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.3104 - mse: 0.3104 - mae: 0.4788\n",
            "Epoch 189/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.2834 - mse: 0.2834 - mae: 0.4552\n",
            "Epoch 190/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.2660 - mse: 0.2660 - mae: 0.4373\n",
            "Epoch 191/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.2783 - mse: 0.2783 - mae: 0.4593\n",
            "Epoch 192/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.2613 - mse: 0.2613 - mae: 0.4297\n",
            "Epoch 193/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.2550 - mse: 0.2550 - mae: 0.4284\n",
            "Epoch 194/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.2507 - mse: 0.2507 - mae: 0.4248\n",
            "Epoch 195/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.2500 - mse: 0.2500 - mae: 0.4206\n",
            "Epoch 196/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.2483 - mse: 0.2483 - mae: 0.4188\n",
            "Epoch 197/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.2431 - mse: 0.2431 - mae: 0.4153\n",
            "Epoch 198/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.2493 - mse: 0.2493 - mae: 0.4337\n",
            "Epoch 199/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.2376 - mse: 0.2376 - mae: 0.4096\n",
            "Epoch 200/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.2478 - mse: 0.2478 - mae: 0.4341\n",
            "Epoch 201/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.2408 - mse: 0.2408 - mae: 0.4257\n",
            "Epoch 202/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.2400 - mse: 0.2400 - mae: 0.4283\n",
            "Epoch 203/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.2356 - mse: 0.2356 - mae: 0.4201\n",
            "Epoch 204/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.2206 - mse: 0.2206 - mae: 0.3932\n",
            "Epoch 205/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.2272 - mse: 0.2272 - mae: 0.4153\n",
            "Epoch 206/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.2247 - mse: 0.2247 - mae: 0.4105\n",
            "Epoch 207/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.2073 - mse: 0.2073 - mae: 0.3867\n",
            "Epoch 208/1000\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.2263 - mse: 0.2263 - mae: 0.4057\n",
            "Epoch 209/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.2005 - mse: 0.2005 - mae: 0.3800\n",
            "Epoch 210/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.2115 - mse: 0.2115 - mae: 0.3982\n",
            "Epoch 211/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.2089 - mse: 0.2089 - mae: 0.3962\n",
            "Epoch 212/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.2043 - mse: 0.2043 - mae: 0.3916\n",
            "Epoch 213/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.2078 - mse: 0.2078 - mae: 0.3874\n",
            "Epoch 214/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.2003 - mse: 0.2003 - mae: 0.3915\n",
            "Epoch 215/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1863 - mse: 0.1863 - mae: 0.3611\n",
            "Epoch 216/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.1811 - mse: 0.1811 - mae: 0.3614\n",
            "Epoch 217/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1915 - mse: 0.1915 - mae: 0.3834\n",
            "Epoch 218/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1765 - mse: 0.1765 - mae: 0.3561\n",
            "Epoch 219/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.1735 - mse: 0.1735 - mae: 0.3535\n",
            "Epoch 220/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.1706 - mse: 0.1706 - mae: 0.3513\n",
            "Epoch 221/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1779 - mse: 0.1779 - mae: 0.3673\n",
            "Epoch 222/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1687 - mse: 0.1687 - mae: 0.3421\n",
            "Epoch 223/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.1635 - mse: 0.1635 - mae: 0.3417\n",
            "Epoch 224/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.1610 - mse: 0.1610 - mae: 0.3407\n",
            "Epoch 225/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.1701 - mse: 0.1701 - mae: 0.3615\n",
            "Epoch 226/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.1567 - mse: 0.1567 - mae: 0.3364\n",
            "Epoch 227/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.1651 - mse: 0.1651 - mae: 0.3566\n",
            "Epoch 228/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.1547 - mse: 0.1547 - mae: 0.3302\n",
            "Epoch 229/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.1506 - mse: 0.1506 - mae: 0.3262\n",
            "Epoch 230/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.1563 - mse: 0.1563 - mae: 0.3411\n",
            "Epoch 231/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.1448 - mse: 0.1448 - mae: 0.3234\n",
            "Epoch 232/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.1428 - mse: 0.1428 - mae: 0.3210\n",
            "Epoch 233/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.1408 - mse: 0.1408 - mae: 0.3182\n",
            "Epoch 234/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.1492 - mse: 0.1492 - mae: 0.3328\n",
            "Epoch 235/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.1370 - mse: 0.1370 - mae: 0.3126\n",
            "Epoch 236/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.1342 - mse: 0.1342 - mae: 0.3106\n",
            "Epoch 237/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.1349 - mse: 0.1349 - mae: 0.3093\n",
            "Epoch 238/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1308 - mse: 0.1308 - mae: 0.3076\n",
            "Epoch 239/1000\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.1367 - mse: 0.1367 - mae: 0.3187\n",
            "Epoch 240/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.1270 - mse: 0.1270 - mae: 0.3023\n",
            "Epoch 241/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1319 - mse: 0.1319 - mae: 0.3167\n",
            "Epoch 242/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.1229 - mse: 0.1229 - mae: 0.2976\n",
            "Epoch 243/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.1236 - mse: 0.1236 - mae: 0.2939\n",
            "Epoch 244/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.1273 - mse: 0.1273 - mae: 0.3109\n",
            "Epoch 245/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1190 - mse: 0.1190 - mae: 0.2901\n",
            "Epoch 246/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1162 - mse: 0.1162 - mae: 0.2894\n",
            "Epoch 247/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.1145 - mse: 0.1145 - mae: 0.2872\n",
            "Epoch 248/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.1128 - mse: 0.1128 - mae: 0.2851\n",
            "Epoch 249/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1110 - mse: 0.1110 - mae: 0.2832\n",
            "Epoch 250/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.1113 - mse: 0.1113 - mae: 0.2810\n",
            "Epoch 251/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.1079 - mse: 0.1079 - mae: 0.2783\n",
            "Epoch 252/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1201 - mse: 0.1201 - mae: 0.2979\n",
            "Epoch 253/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.1070 - mse: 0.1070 - mae: 0.2716\n",
            "Epoch 254/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.1053 - mse: 0.1053 - mae: 0.2724\n",
            "Epoch 255/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.1079 - mse: 0.1079 - mae: 0.2841\n",
            "Epoch 256/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.1109 - mse: 0.1109 - mae: 0.2840\n",
            "Epoch 257/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0983 - mse: 0.0983 - mae: 0.2660\n",
            "Epoch 258/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0971 - mse: 0.0971 - mae: 0.2646\n",
            "Epoch 259/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1030 - mse: 0.1030 - mae: 0.2765\n",
            "Epoch 260/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0942 - mse: 0.0942 - mae: 0.2596\n",
            "Epoch 261/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0943 - mse: 0.0943 - mae: 0.2567\n",
            "Epoch 262/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0915 - mse: 0.0915 - mae: 0.2572\n",
            "Epoch 263/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0910 - mse: 0.0910 - mae: 0.2538\n",
            "Epoch 264/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0898 - mse: 0.0898 - mae: 0.2523\n",
            "Epoch 265/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0885 - mse: 0.0885 - mae: 0.2503\n",
            "Epoch 266/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0870 - mse: 0.0870 - mae: 0.2482\n",
            "Epoch 267/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0848 - mse: 0.0848 - mae: 0.2470\n",
            "Epoch 268/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0946 - mse: 0.0946 - mae: 0.2641\n",
            "Epoch 269/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0828 - mse: 0.0828 - mae: 0.2427\n",
            "Epoch 270/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0913 - mse: 0.0913 - mae: 0.2573\n",
            "Epoch 271/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0850 - mse: 0.0850 - mae: 0.2549\n",
            "Epoch 272/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0805 - mse: 0.0805 - mae: 0.2372\n",
            "Epoch 273/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0775 - mse: 0.0775 - mae: 0.2360\n",
            "Epoch 274/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0766 - mse: 0.0766 - mae: 0.2349\n",
            "Epoch 275/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0812 - mse: 0.0812 - mae: 0.2491\n",
            "Epoch 276/1000\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.0758 - mse: 0.0758 - mae: 0.2314\n",
            "Epoch 277/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0775 - mse: 0.0775 - mae: 0.2430\n",
            "Epoch 278/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0766 - mse: 0.0766 - mae: 0.2385\n",
            "Epoch 279/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0709 - mse: 0.0709 - mae: 0.2259\n",
            "Epoch 280/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0753 - mse: 0.0753 - mae: 0.2403\n",
            "Epoch 281/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0704 - mse: 0.0704 - mae: 0.2229\n",
            "Epoch 282/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0682 - mse: 0.0682 - mae: 0.2207\n",
            "Epoch 283/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0714 - mse: 0.0714 - mae: 0.2331\n",
            "Epoch 284/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0659 - mse: 0.0659 - mae: 0.2177\n",
            "Epoch 285/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0736 - mse: 0.0736 - mae: 0.2335\n",
            "Epoch 286/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0679 - mse: 0.0679 - mae: 0.2277\n",
            "Epoch 287/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0633 - mse: 0.0633 - mae: 0.2122\n",
            "Epoch 288/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0621 - mse: 0.0621 - mae: 0.2110\n",
            "Epoch 289/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0658 - mse: 0.0658 - mae: 0.2215\n",
            "Epoch 290/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0612 - mse: 0.0612 - mae: 0.2074\n",
            "Epoch 291/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0603 - mse: 0.0603 - mae: 0.2055\n",
            "Epoch 292/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0584 - mse: 0.0584 - mae: 0.2051\n",
            "Epoch 293/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0578 - mse: 0.0578 - mae: 0.2042\n",
            "Epoch 294/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0607 - mse: 0.0607 - mae: 0.2151\n",
            "Epoch 295/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0559 - mse: 0.0559 - mae: 0.2007\n",
            "Epoch 296/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0558 - mse: 0.0558 - mae: 0.1987\n",
            "Epoch 297/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0587 - mse: 0.0587 - mae: 0.2085\n",
            "Epoch 298/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0538 - mse: 0.0538 - mae: 0.1972\n",
            "Epoch 299/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0567 - mse: 0.0567 - mae: 0.2058\n",
            "Epoch 300/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0521 - mse: 0.0521 - mae: 0.1928\n",
            "Epoch 301/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0550 - mse: 0.0550 - mae: 0.2028\n",
            "Epoch 302/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0565 - mse: 0.0565 - mae: 0.2027\n",
            "Epoch 303/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0509 - mse: 0.0509 - mae: 0.1883\n",
            "Epoch 304/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0519 - mse: 0.0519 - mae: 0.1987\n",
            "Epoch 305/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0510 - mse: 0.0510 - mae: 0.1931\n",
            "Epoch 306/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0500 - mse: 0.0500 - mae: 0.1943\n",
            "Epoch 307/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0501 - mse: 0.0501 - mae: 0.1938\n",
            "Epoch 308/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0491 - mse: 0.0491 - mae: 0.1920\n",
            "Epoch 309/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0455 - mse: 0.0455 - mae: 0.1806\n",
            "Epoch 310/1000\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.0452 - mse: 0.0452 - mae: 0.1789\n",
            "Epoch 311/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0442 - mse: 0.0442 - mae: 0.1786\n",
            "Epoch 312/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0442 - mse: 0.0442 - mae: 0.1767\n",
            "Epoch 313/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0429 - mse: 0.0429 - mae: 0.1760\n",
            "Epoch 314/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0445 - mse: 0.0445 - mae: 0.1834\n",
            "Epoch 315/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0440 - mse: 0.0440 - mae: 0.1830\n",
            "Epoch 316/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0418 - mse: 0.0418 - mae: 0.1705\n",
            "Epoch 317/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0403 - mse: 0.0403 - mae: 0.1704\n",
            "Epoch 318/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0428 - mse: 0.0428 - mae: 0.1809\n",
            "Epoch 319/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0415 - mse: 0.0415 - mae: 0.1778\n",
            "Epoch 320/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0387 - mse: 0.0387 - mae: 0.1661\n",
            "Epoch 321/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0380 - mse: 0.0380 - mae: 0.1652\n",
            "Epoch 322/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0398 - mse: 0.0398 - mae: 0.1721\n",
            "Epoch 323/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0369 - mse: 0.0369 - mae: 0.1629\n",
            "Epoch 324/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0371 - mse: 0.0371 - mae: 0.1621\n",
            "Epoch 325/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0359 - mse: 0.0359 - mae: 0.1606\n",
            "Epoch 326/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0359 - mse: 0.0359 - mae: 0.1584\n",
            "Epoch 327/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0368 - mse: 0.0368 - mae: 0.1672\n",
            "Epoch 328/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0343 - mse: 0.0343 - mae: 0.1575\n",
            "Epoch 329/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0342 - mse: 0.0342 - mae: 0.1555\n",
            "Epoch 330/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0354 - mse: 0.0354 - mae: 0.1633\n",
            "Epoch 331/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0333 - mse: 0.0333 - mae: 0.1527\n",
            "Epoch 332/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0326 - mse: 0.0326 - mae: 0.1520\n",
            "Epoch 333/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0343 - mse: 0.0343 - mae: 0.1618\n",
            "Epoch 334/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0332 - mse: 0.0332 - mae: 0.1564\n",
            "Epoch 335/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0350 - mse: 0.0350 - mae: 0.1607\n",
            "Epoch 336/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0327 - mse: 0.0327 - mae: 0.1566\n",
            "Epoch 337/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0323 - mse: 0.0323 - mae: 0.1584\n",
            "Epoch 338/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0314 - mse: 0.0314 - mae: 0.1548\n",
            "Epoch 339/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0294 - mse: 0.0294 - mae: 0.1441\n",
            "Epoch 340/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0304 - mse: 0.0304 - mae: 0.1520\n",
            "Epoch 341/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0299 - mse: 0.0299 - mae: 0.1478\n",
            "Epoch 342/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0296 - mse: 0.0296 - mae: 0.1485\n",
            "Epoch 343/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0311 - mse: 0.0311 - mae: 0.1523\n",
            "Epoch 344/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0283 - mse: 0.0283 - mae: 0.1441\n",
            "Epoch 345/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0271 - mse: 0.0271 - mae: 0.1374\n",
            "Epoch 346/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0267 - mse: 0.0267 - mae: 0.1375\n",
            "Epoch 347/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0275 - mse: 0.0275 - mae: 0.1438\n",
            "Epoch 348/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0274 - mse: 0.0274 - mae: 0.1446\n",
            "Epoch 349/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0255 - mse: 0.0255 - mae: 0.1339\n",
            "Epoch 350/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0265 - mse: 0.0265 - mae: 0.1426\n",
            "Epoch 351/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0266 - mse: 0.0266 - mae: 0.1383\n",
            "Epoch 352/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0240 - mse: 0.0240 - mae: 0.1314\n",
            "Epoch 353/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0236 - mse: 0.0236 - mae: 0.1307\n",
            "Epoch 354/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0248 - mse: 0.0248 - mae: 0.1377\n",
            "Epoch 355/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0229 - mse: 0.0229 - mae: 0.1284\n",
            "Epoch 356/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0256 - mse: 0.0256 - mae: 0.1383\n",
            "Epoch 357/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0223 - mse: 0.0223 - mae: 0.1262\n",
            "Epoch 358/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0230 - mse: 0.0230 - mae: 0.1309\n",
            "Epoch 359/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0219 - mse: 0.0219 - mae: 0.1238\n",
            "Epoch 360/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0217 - mse: 0.0217 - mae: 0.1240\n",
            "Epoch 361/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0209 - mse: 0.0209 - mae: 0.1226\n",
            "Epoch 362/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0207 - mse: 0.0207 - mae: 0.1222\n",
            "Epoch 363/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0203 - mse: 0.0203 - mae: 0.1209\n",
            "Epoch 364/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0200 - mse: 0.0200 - mae: 0.1200\n",
            "Epoch 365/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0200 - mse: 0.0200 - mae: 0.1190\n",
            "Epoch 366/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0195 - mse: 0.0195 - mae: 0.1186\n",
            "Epoch 367/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0192 - mse: 0.0192 - mae: 0.1173\n",
            "Epoch 368/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0210 - mse: 0.0210 - mae: 0.1234\n",
            "Epoch 369/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0190 - mse: 0.0190 - mae: 0.1160\n",
            "Epoch 370/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0184 - mse: 0.0184 - mae: 0.1153\n",
            "Epoch 371/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0184 - mse: 0.0184 - mae: 0.1133\n",
            "Epoch 372/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0180 - mse: 0.0180 - mae: 0.1129\n",
            "Epoch 373/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0175 - mse: 0.0175 - mae: 0.1123\n",
            "Epoch 374/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0173 - mse: 0.0173 - mae: 0.1115\n",
            "Epoch 375/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0170 - mse: 0.0170 - mae: 0.1107\n",
            "Epoch 376/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0190 - mse: 0.0190 - mae: 0.1189\n",
            "Epoch 377/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0169 - mse: 0.0169 - mae: 0.1081\n",
            "Epoch 378/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0164 - mse: 0.0164 - mae: 0.1085\n",
            "Epoch 379/1000\n",
            "2/2 [==============================] - 0s 25ms/step - loss: 0.0163 - mse: 0.0163 - mae: 0.1073\n",
            "Epoch 380/1000\n",
            "2/2 [==============================] - 0s 16ms/step - loss: 0.0159 - mse: 0.0159 - mae: 0.1068\n",
            "Epoch 381/1000\n",
            "2/2 [==============================] - 0s 25ms/step - loss: 0.0165 - mse: 0.0165 - mae: 0.1118\n",
            "Epoch 382/1000\n",
            "2/2 [==============================] - 0s 18ms/step - loss: 0.0163 - mse: 0.0163 - mae: 0.1114\n",
            "Epoch 383/1000\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 0.0163 - mse: 0.0163 - mae: 0.1121\n",
            "Epoch 384/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.0150 - mse: 0.0150 - mae: 0.1038\n",
            "Epoch 385/1000\n",
            "2/2 [==============================] - 0s 22ms/step - loss: 0.0149 - mse: 0.0149 - mae: 0.1026\n",
            "Epoch 386/1000\n",
            "2/2 [==============================] - 0s 17ms/step - loss: 0.0153 - mse: 0.0153 - mae: 0.1079\n",
            "Epoch 387/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0144 - mse: 0.0144 - mae: 0.1009\n",
            "Epoch 388/1000\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 0.0140 - mse: 0.0140 - mae: 0.1004\n",
            "Epoch 389/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0139 - mse: 0.0139 - mae: 0.1000\n",
            "Epoch 390/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.0139 - mse: 0.0139 - mae: 0.0986\n",
            "Epoch 391/1000\n",
            "2/2 [==============================] - 0s 43ms/step - loss: 0.0137 - mse: 0.0137 - mae: 0.0977\n",
            "Epoch 392/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0133 - mse: 0.0133 - mae: 0.0977\n",
            "Epoch 393/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0133 - mse: 0.0133 - mae: 0.0964\n",
            "Epoch 394/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0136 - mse: 0.0136 - mae: 0.1017\n",
            "Epoch 395/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0129 - mse: 0.0129 - mae: 0.0952\n",
            "Epoch 396/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0134 - mse: 0.0134 - mae: 0.1015\n",
            "Epoch 397/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0126 - mse: 0.0126 - mae: 0.0942\n",
            "Epoch 398/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0124 - mse: 0.0124 - mae: 0.0926\n",
            "Epoch 399/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0125 - mse: 0.0125 - mae: 0.0965\n",
            "Epoch 400/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0120 - mse: 0.0120 - mae: 0.0922\n",
            "Epoch 401/1000\n",
            "2/2 [==============================] - 0s 16ms/step - loss: 0.0125 - mse: 0.0125 - mae: 0.0964\n",
            "Epoch 402/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.0115 - mse: 0.0115 - mae: 0.0904\n",
            "Epoch 403/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.0113 - mse: 0.0113 - mae: 0.0900\n",
            "Epoch 404/1000\n",
            "2/2 [==============================] - 0s 17ms/step - loss: 0.0123 - mse: 0.0123 - mae: 0.0945\n",
            "Epoch 405/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0109 - mse: 0.0109 - mae: 0.0888\n",
            "Epoch 406/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.0109 - mse: 0.0109 - mae: 0.0877\n",
            "Epoch 407/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0114 - mse: 0.0114 - mae: 0.0921\n",
            "Epoch 408/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0106 - mse: 0.0106 - mae: 0.0862\n",
            "Epoch 409/1000\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 0.0110 - mse: 0.0110 - mae: 0.0915\n",
            "Epoch 410/1000\n",
            "2/2 [==============================] - 0s 21ms/step - loss: 0.0101 - mse: 0.0101 - mae: 0.0854\n",
            "Epoch 411/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0102 - mse: 0.0102 - mae: 0.0842\n",
            "Epoch 412/1000\n",
            "2/2 [==============================] - 0s 17ms/step - loss: 0.0112 - mse: 0.0112 - mae: 0.0915\n",
            "Epoch 413/1000\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 0.0099 - mse: 0.0099 - mae: 0.0830\n",
            "Epoch 414/1000\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 0.0097 - mse: 0.0097 - mae: 0.0828\n",
            "Epoch 415/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0096 - mse: 0.0096 - mae: 0.0817\n",
            "Epoch 416/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0094 - mse: 0.0094 - mae: 0.0811\n",
            "Epoch 417/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0093 - mse: 0.0093 - mae: 0.0813\n",
            "Epoch 418/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0091 - mse: 0.0091 - mae: 0.0803\n",
            "Epoch 419/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0100 - mse: 0.0100 - mae: 0.0853\n",
            "Epoch 420/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0089 - mse: 0.0089 - mae: 0.0789\n",
            "Epoch 421/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0088 - mse: 0.0088 - mae: 0.0788\n",
            "Epoch 422/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0095 - mse: 0.0095 - mae: 0.0831\n",
            "Epoch 423/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0088 - mse: 0.0088 - mae: 0.0797\n",
            "Epoch 424/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0087 - mse: 0.0087 - mae: 0.0806\n",
            "Epoch 425/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0081 - mse: 0.0081 - mae: 0.0765\n",
            "Epoch 426/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0081 - mse: 0.0081 - mae: 0.0753\n",
            "Epoch 427/1000\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.0080 - mse: 0.0080 - mae: 0.0747\n",
            "Epoch 428/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0088 - mse: 0.0088 - mae: 0.0813\n",
            "Epoch 429/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0086 - mse: 0.0086 - mae: 0.0788\n",
            "Epoch 430/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0079 - mse: 0.0079 - mae: 0.0754\n",
            "Epoch 431/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0079 - mse: 0.0079 - mae: 0.0772\n",
            "Epoch 432/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0081 - mse: 0.0081 - mae: 0.0769\n",
            "Epoch 433/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0072 - mse: 0.0072 - mae: 0.0721\n",
            "Epoch 434/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0076 - mse: 0.0076 - mae: 0.0751\n",
            "Epoch 435/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0071 - mse: 0.0071 - mae: 0.0705\n",
            "Epoch 436/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0072 - mse: 0.0072 - mae: 0.0738\n",
            "Epoch 437/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0073 - mse: 0.0073 - mae: 0.0746\n",
            "Epoch 438/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0068 - mse: 0.0068 - mae: 0.0692\n",
            "Epoch 439/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0070 - mse: 0.0070 - mae: 0.0723\n",
            "Epoch 440/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0065 - mse: 0.0065 - mae: 0.0684\n",
            "Epoch 441/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0069 - mse: 0.0069 - mae: 0.0724\n",
            "Epoch 442/1000\n",
            "2/2 [==============================] - 0s 17ms/step - loss: 0.0063 - mse: 0.0063 - mae: 0.0672\n",
            "Epoch 443/1000\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 0.0067 - mse: 0.0067 - mae: 0.0715\n",
            "Epoch 444/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.0068 - mse: 0.0068 - mae: 0.0705\n",
            "Epoch 445/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0062 - mse: 0.0062 - mae: 0.0651\n",
            "Epoch 446/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0061 - mse: 0.0061 - mae: 0.0647\n",
            "Epoch 447/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0060 - mse: 0.0060 - mae: 0.0642\n",
            "Epoch 448/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0058 - mse: 0.0058 - mae: 0.0646\n",
            "Epoch 449/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0064 - mse: 0.0064 - mae: 0.0680\n",
            "Epoch 450/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0058 - mse: 0.0058 - mae: 0.0652\n",
            "Epoch 451/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0059 - mse: 0.0059 - mae: 0.0674\n",
            "Epoch 452/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0055 - mse: 0.0055 - mae: 0.0621\n",
            "Epoch 453/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0054 - mse: 0.0054 - mae: 0.0616\n",
            "Epoch 454/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0053 - mse: 0.0053 - mae: 0.0614\n",
            "Epoch 455/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0052 - mse: 0.0052 - mae: 0.0610\n",
            "Epoch 456/1000\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 0.0054 - mse: 0.0054 - mae: 0.0635\n",
            "Epoch 457/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0051 - mse: 0.0051 - mae: 0.0603\n",
            "Epoch 458/1000\n",
            "2/2 [==============================] - 0s 17ms/step - loss: 0.0049 - mse: 0.0049 - mae: 0.0596\n",
            "Epoch 459/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0049 - mse: 0.0049 - mae: 0.0592\n",
            "Epoch 460/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.0051 - mse: 0.0051 - mae: 0.0622\n",
            "Epoch 461/1000\n",
            "2/2 [==============================] - 0s 17ms/step - loss: 0.0048 - mse: 0.0048 - mae: 0.0582\n",
            "Epoch 462/1000\n",
            "2/2 [==============================] - 0s 19ms/step - loss: 0.0051 - mse: 0.0051 - mae: 0.0608\n",
            "Epoch 463/1000\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 0.0046 - mse: 0.0046 - mae: 0.0577   \n",
            "Epoch 464/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0045 - mse: 0.0045 - mae: 0.0572\n",
            "Epoch 465/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0045 - mse: 0.0045 - mae: 0.0568\n",
            "Epoch 466/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0045 - mse: 0.0045 - mae: 0.0562\n",
            "Epoch 467/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.0047 - mse: 0.0047 - mae: 0.0597\n",
            "Epoch 468/1000\n",
            "2/2 [==============================] - 0s 16ms/step - loss: 0.0043 - mse: 0.0043 - mae: 0.0553\n",
            "Epoch 469/1000\n",
            "2/2 [==============================] - 0s 16ms/step - loss: 0.0042 - mse: 0.0042 - mae: 0.0548\n",
            "Epoch 470/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0042 - mse: 0.0042 - mae: 0.0546\n",
            "Epoch 471/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0044 - mse: 0.0044 - mae: 0.0572\n",
            "Epoch 472/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0040 - mse: 0.0040 - mae: 0.0537\n",
            "Epoch 473/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0043 - mse: 0.0043 - mae: 0.0572\n",
            "Epoch 474/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0039 - mse: 0.0039 - mae: 0.0532\n",
            "Epoch 475/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0039 - mse: 0.0039 - mae: 0.0526\n",
            "Epoch 476/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0041 - mse: 0.0041 - mae: 0.0560\n",
            "Epoch 477/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0040 - mse: 0.0040 - mae: 0.0558\n",
            "Epoch 478/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0037 - mse: 0.0037 - mae: 0.0514\n",
            "Epoch 479/1000\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.0037 - mse: 0.0037 - mae: 0.0508\n",
            "Epoch 480/1000\n",
            "2/2 [==============================] - 0s 20ms/step - loss: 0.0036 - mse: 0.0036 - mae: 0.0509   \n",
            "Epoch 481/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0039 - mse: 0.0039 - mae: 0.0532\n",
            "Epoch 482/1000\n",
            "2/2 [==============================] - 0s 16ms/step - loss: 0.0037 - mse: 0.0037 - mae: 0.0534\n",
            "Epoch 483/1000\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 0.0036 - mse: 0.0036 - mae: 0.0522\n",
            "Epoch 484/1000\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 0.0036 - mse: 0.0036 - mae: 0.0521\n",
            "Epoch 485/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0035 - mse: 0.0035 - mae: 0.0518\n",
            "Epoch 486/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.0033 - mse: 0.0033 - mae: 0.0486\n",
            "Epoch 487/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.0036 - mse: 0.0036 - mae: 0.0512\n",
            "Epoch 488/1000\n",
            "2/2 [==============================] - 0s 19ms/step - loss: 0.0032 - mse: 0.0032 - mae: 0.0475\n",
            "Epoch 489/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0032 - mse: 0.0032 - mae: 0.0470\n",
            "Epoch 490/1000\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 0.0033 - mse: 0.0033 - mae: 0.0497\n",
            "Epoch 491/1000\n",
            "2/2 [==============================] - 0s 17ms/step - loss: 0.0031 - mse: 0.0031 - mae: 0.0465\n",
            "Epoch 492/1000\n",
            "2/2 [==============================] - 0s 17ms/step - loss: 0.0032 - mse: 0.0032 - mae: 0.0485\n",
            "Epoch 493/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0029 - mse: 0.0029 - mae: 0.0461\n",
            "Epoch 494/1000\n",
            "2/2 [==============================] - 0s 18ms/step - loss: 0.0030 - mse: 0.0030 - mae: 0.0454\n",
            "Epoch 495/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0031 - mse: 0.0031 - mae: 0.0479\n",
            "Epoch 496/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0029 - mse: 0.0029 - mae: 0.0447\n",
            "Epoch 497/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0030 - mse: 0.0030 - mae: 0.0479\n",
            "Epoch 498/1000\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 0.0029 - mse: 0.0029 - mae: 0.0460\n",
            "Epoch 499/1000\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 0.0027 - mse: 0.0027 - mae: 0.0441\n",
            "Epoch 500/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0026 - mse: 0.0026 - mae: 0.0436\n",
            "Epoch 501/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.0028 - mse: 0.0028 - mae: 0.0461\n",
            "Epoch 502/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0027 - mse: 0.0027 - mae: 0.0455\n",
            "Epoch 503/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0029 - mse: 0.0029 - mae: 0.0466\n",
            "Epoch 504/1000\n",
            "2/2 [==============================] - 0s 16ms/step - loss: 0.0026 - mse: 0.0026 - mae: 0.0450\n",
            "Epoch 505/1000\n",
            "2/2 [==============================] - 0s 24ms/step - loss: 0.0026 - mse: 0.0026 - mae: 0.0447\n",
            "Epoch 506/1000\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 0.0025 - mse: 0.0025 - mae: 0.0431\n",
            "Epoch 507/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.0024 - mse: 0.0024 - mae: 0.0416   \n",
            "Epoch 508/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0026 - mse: 0.0026 - mae: 0.0434   \n",
            "Epoch 509/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0025 - mse: 0.0025 - mae: 0.0436\n",
            "Epoch 510/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0023 - mse: 0.0023 - mae: 0.0406\n",
            "Epoch 511/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0024 - mse: 0.0024 - mae: 0.0425\n",
            "Epoch 512/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0024 - mse: 0.0024 - mae: 0.0430\n",
            "Epoch 513/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0022 - mse: 0.0022 - mae: 0.0395\n",
            "Epoch 514/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0023 - mse: 0.0023 - mae: 0.0422\n",
            "Epoch 515/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0021 - mse: 0.0021 - mae: 0.0389\n",
            "Epoch 516/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0021 - mse: 0.0021 - mae: 0.0387\n",
            "Epoch 517/1000\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 0.0021 - mse: 0.0021 - mae: 0.0386\n",
            "Epoch 518/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0021 - mse: 0.0021 - mae: 0.0404\n",
            "Epoch 519/1000\n",
            "2/2 [==============================] - 0s 21ms/step - loss: 0.0020 - mse: 0.0020 - mae: 0.0378\n",
            "Epoch 520/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0020 - mse: 0.0020 - mae: 0.0376\n",
            "Epoch 521/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0020 - mse: 0.0020 - mae: 0.0371\n",
            "Epoch 522/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0020 - mse: 0.0020 - mae: 0.0393\n",
            "Epoch 523/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0019 - mse: 0.0019 - mae: 0.0368\n",
            "Epoch 524/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0020 - mse: 0.0020 - mae: 0.0392\n",
            "Epoch 525/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0018 - mse: 0.0018 - mae: 0.0363    \n",
            "Epoch 526/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0020 - mse: 0.0020 - mae: 0.0389\n",
            "Epoch 527/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0018 - mse: 0.0018 - mae: 0.0355    \n",
            "Epoch 528/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0018 - mse: 0.0018 - mae: 0.0354\n",
            "Epoch 529/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0017 - mse: 0.0017 - mae: 0.0353    \n",
            "Epoch 530/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0017 - mse: 0.0017 - mae: 0.0349\n",
            "Epoch 531/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0017 - mse: 0.0017 - mae: 0.0348    \n",
            "Epoch 532/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0017 - mse: 0.0017 - mae: 0.0344\n",
            "Epoch 533/1000\n",
            "2/2 [==============================] - 0s 17ms/step - loss: 0.0017 - mse: 0.0017 - mae: 0.0361\n",
            "Epoch 534/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.0016 - mse: 0.0016 - mae: 0.0338\n",
            "Epoch 535/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0016 - mse: 0.0016 - mae: 0.0337\n",
            "Epoch 536/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0016 - mse: 0.0016 - mae: 0.0352\n",
            "Epoch 537/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0016 - mse: 0.0016 - mae: 0.0351   \n",
            "Epoch 538/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0015 - mse: 0.0015 - mae: 0.0329\n",
            "Epoch 539/1000\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 0.0015 - mse: 0.0015 - mae: 0.0327   \n",
            "Epoch 540/1000\n",
            "2/2 [==============================] - 0s 19ms/step - loss: 0.0016 - mse: 0.0016 - mae: 0.0342\n",
            "Epoch 541/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0015 - mse: 0.0015 - mae: 0.0339\n",
            "Epoch 542/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0015 - mse: 0.0015 - mae: 0.0337\n",
            "Epoch 543/1000\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 0.0014 - mse: 0.0014 - mae: 0.0318   \n",
            "Epoch 544/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0015 - mse: 0.0015 - mae: 0.0332\n",
            "Epoch 545/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0014 - mse: 0.0014 - mae: 0.0313\n",
            "Epoch 546/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0014 - mse: 0.0014 - mae: 0.0311   \n",
            "Epoch 547/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0013 - mse: 0.0013 - mae: 0.0308\n",
            "Epoch 548/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0014 - mse: 0.0014 - mae: 0.0322    \n",
            "Epoch 549/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0013 - mse: 0.0013 - mae: 0.0301    \n",
            "Epoch 550/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0013 - mse: 0.0013 - mae: 0.0301\n",
            "Epoch 551/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0012 - mse: 0.0012 - mae: 0.0299\n",
            "Epoch 552/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0013 - mse: 0.0013 - mae: 0.0313\n",
            "Epoch 553/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0012 - mse: 0.0012 - mae: 0.0295    \n",
            "Epoch 554/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0012 - mse: 0.0012 - mae: 0.0293\n",
            "Epoch 555/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0013 - mse: 0.0013 - mae: 0.0306\n",
            "Epoch 556/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0013 - mse: 0.0013 - mae: 0.0306\n",
            "Epoch 557/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0012 - mse: 0.0012 - mae: 0.0304\n",
            "Epoch 558/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0012 - mse: 0.0012 - mae: 0.0306    \n",
            "Epoch 559/1000\n",
            "2/2 [==============================] - 0s 22ms/step - loss: 0.0012 - mse: 0.0012 - mae: 0.0293\n",
            "Epoch 560/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.0011 - mse: 0.0011 - mae: 0.0280   \n",
            "Epoch 561/1000\n",
            "2/2 [==============================] - 0s 17ms/step - loss: 0.0011 - mse: 0.0011 - mae: 0.0291   \n",
            "Epoch 562/1000\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 0.0011 - mse: 0.0011 - mae: 0.0291   \n",
            "Epoch 563/1000\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 0.0011 - mse: 0.0011 - mae: 0.0288\n",
            "Epoch 564/1000\n",
            "2/2 [==============================] - 0s 16ms/step - loss: 0.0010 - mse: 0.0010 - mae: 0.0271\n",
            "Epoch 565/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0011 - mse: 0.0011 - mae: 0.0282   \n",
            "Epoch 566/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 9.9285e-04 - mse: 9.9285e-04 - mae: 0.0268\n",
            "Epoch 567/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0010 - mse: 0.0010 - mae: 0.0283\n",
            "Epoch 568/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 9.6194e-04 - mse: 9.6194e-04 - mae: 0.0263\n",
            "Epoch 569/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 9.4972e-04 - mse: 9.4972e-04 - mae: 0.0262\n",
            "Epoch 570/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 9.3365e-04 - mse: 9.3365e-04 - mae: 0.0259\n",
            "Epoch 571/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 9.3118e-04 - mse: 9.3118e-04 - mae: 0.0257\n",
            "Epoch 572/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 9.0701e-04 - mse: 9.0701e-04 - mae: 0.0256\n",
            "Epoch 573/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 9.0503e-04 - mse: 9.0503e-04 - mae: 0.0253\n",
            "Epoch 574/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 8.9831e-04 - mse: 8.9831e-04 - mae: 0.0252\n",
            "Epoch 575/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 9.5659e-04 - mse: 9.5659e-04 - mae: 0.0263\n",
            "Epoch 576/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 8.7149e-04 - mse: 8.7149e-04 - mae: 0.0249\n",
            "Epoch 577/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 9.0650e-04 - mse: 9.0650e-04 - mae: 0.0264\n",
            "Epoch 578/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 8.3522e-04 - mse: 8.3522e-04 - mae: 0.0246\n",
            "Epoch 579/1000\n",
            "2/2 [==============================] - 0s 18ms/step - loss: 8.6804e-04 - mse: 8.6804e-04 - mae: 0.0253\n",
            "Epoch 580/1000\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 8.2157e-04 - mse: 8.2157e-04 - mae: 0.0240\n",
            "Epoch 581/1000\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 8.0346e-04 - mse: 8.0346e-04 - mae: 0.0238\n",
            "Epoch 582/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 8.6147e-04 - mse: 8.6147e-04 - mae: 0.0250\n",
            "Epoch 583/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 8.7285e-04 - mse: 8.7285e-04 - mae: 0.0254\n",
            "Epoch 584/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 7.9578e-04 - mse: 7.9578e-04 - mae: 0.0241\n",
            "Epoch 585/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 7.9533e-04 - mse: 7.9533e-04 - mae: 0.0243\n",
            "Epoch 586/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 7.3841e-04 - mse: 7.3841e-04 - mae: 0.0231\n",
            "Epoch 587/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 7.8212e-04 - mse: 7.8212e-04 - mae: 0.0244\n",
            "Epoch 588/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 7.6729e-04 - mse: 7.6729e-04 - mae: 0.0240\n",
            "Epoch 589/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 7.4744e-04 - mse: 7.4744e-04 - mae: 0.0239\n",
            "Epoch 590/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 7.3736e-04 - mse: 7.3736e-04 - mae: 0.0233\n",
            "Epoch 591/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 7.3495e-04 - mse: 7.3495e-04 - mae: 0.0234\n",
            "Epoch 592/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 7.1380e-04 - mse: 7.1380e-04 - mae: 0.0233\n",
            "Epoch 593/1000\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 6.6651e-04 - mse: 6.6651e-04 - mae: 0.0218\n",
            "Epoch 594/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 7.0264e-04 - mse: 7.0264e-04 - mae: 0.0233\n",
            "Epoch 595/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 6.4830e-04 - mse: 6.4830e-04 - mae: 0.0216\n",
            "Epoch 596/1000\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 6.4430e-04 - mse: 6.4430e-04 - mae: 0.0212\n",
            "Epoch 597/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 6.3594e-04 - mse: 6.3594e-04 - mae: 0.0212\n",
            "Epoch 598/1000\n",
            "2/2 [==============================] - 0s 21ms/step - loss: 6.6470e-04 - mse: 6.6470e-04 - mae: 0.0222\n",
            "Epoch 599/1000\n",
            "2/2 [==============================] - 0s 16ms/step - loss: 6.1730e-04 - mse: 6.1730e-04 - mae: 0.0208\n",
            "Epoch 600/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 5.9689e-04 - mse: 5.9689e-04 - mae: 0.0207\n",
            "Epoch 601/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 6.2537e-04 - mse: 6.2537e-04 - mae: 0.0217\n",
            "Epoch 602/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 6.3654e-04 - mse: 6.3654e-04 - mae: 0.0214\n",
            "Epoch 603/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 6.0643e-04 - mse: 6.0643e-04 - mae: 0.0214\n",
            "Epoch 604/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 5.6380e-04 - mse: 5.6380e-04 - mae: 0.0201\n",
            "Epoch 605/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 5.6421e-04 - mse: 5.6421e-04 - mae: 0.0199\n",
            "Epoch 606/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 5.5720e-04 - mse: 5.5720e-04 - mae: 0.0197\n",
            "Epoch 607/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 5.7403e-04 - mse: 5.7403e-04 - mae: 0.0209\n",
            "Epoch 608/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 6.0241e-04 - mse: 6.0241e-04 - mae: 0.0212\n",
            "Epoch 609/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 5.6279e-04 - mse: 5.6279e-04 - mae: 0.0209\n",
            "Epoch 610/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 5.2032e-04 - mse: 5.2032e-04 - mae: 0.0193\n",
            "Epoch 611/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 5.4419e-04 - mse: 5.4419e-04 - mae: 0.0202\n",
            "Epoch 612/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 5.3819e-04 - mse: 5.3819e-04 - mae: 0.0204\n",
            "Epoch 613/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 5.0509e-04 - mse: 5.0509e-04 - mae: 0.0188\n",
            "Epoch 614/1000\n",
            "2/2 [==============================] - 0s 19ms/step - loss: 4.8776e-04 - mse: 4.8776e-04 - mae: 0.0188\n",
            "Epoch 615/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 5.0625e-04 - mse: 5.0625e-04 - mae: 0.0196\n",
            "Epoch 616/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 4.7404e-04 - mse: 4.7404e-04 - mae: 0.0184\n",
            "Epoch 617/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 4.8682e-04 - mse: 4.8682e-04 - mae: 0.0191\n",
            "Epoch 618/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 4.6166e-04 - mse: 4.6166e-04 - mae: 0.0181\n",
            "Epoch 619/1000\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 5.0173e-04 - mse: 5.0173e-04 - mae: 0.0191\n",
            "Epoch 620/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 4.9541e-04 - mse: 4.9541e-04 - mae: 0.0190\n",
            "Epoch 621/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 4.3788e-04 - mse: 4.3788e-04 - mae: 0.0178\n",
            "Epoch 622/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 4.3854e-04 - mse: 4.3854e-04 - mae: 0.0175\n",
            "Epoch 623/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 4.2510e-04 - mse: 4.2510e-04 - mae: 0.0175\n",
            "Epoch 624/1000\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 4.1828e-04 - mse: 4.1828e-04 - mae: 0.0174\n",
            "Epoch 625/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 4.1200e-04 - mse: 4.1200e-04 - mae: 0.0173\n",
            "Epoch 626/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 4.3692e-04 - mse: 4.3692e-04 - mae: 0.0180\n",
            "Epoch 627/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 4.2279e-04 - mse: 4.2279e-04 - mae: 0.0176\n",
            "Epoch 628/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.9447e-04 - mse: 3.9447e-04 - mae: 0.0169\n",
            "Epoch 629/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 4.1129e-04 - mse: 4.1129e-04 - mae: 0.0177\n",
            "Epoch 630/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.8729e-04 - mse: 3.8729e-04 - mae: 0.0165\n",
            "Epoch 631/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 4.0479e-04 - mse: 4.0479e-04 - mae: 0.0177\n",
            "Epoch 632/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.7790e-04 - mse: 3.7790e-04 - mae: 0.0163\n",
            "Epoch 633/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.6489e-04 - mse: 3.6489e-04 - mae: 0.0162\n",
            "Epoch 634/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.8768e-04 - mse: 3.8768e-04 - mae: 0.0170\n",
            "Epoch 635/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.9761e-04 - mse: 3.9761e-04 - mae: 0.0170\n",
            "Epoch 636/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.5661e-04 - mse: 3.5661e-04 - mae: 0.0158\n",
            "Epoch 637/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.4491e-04 - mse: 3.4491e-04 - mae: 0.0157\n",
            "Epoch 638/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.5973e-04 - mse: 3.5973e-04 - mae: 0.0165\n",
            "Epoch 639/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.5412e-04 - mse: 3.5412e-04 - mae: 0.0164\n",
            "Epoch 640/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.3056e-04 - mse: 3.3056e-04 - mae: 0.0154\n",
            "Epoch 641/1000\n",
            "2/2 [==============================] - 0s 22ms/step - loss: 3.3064e-04 - mse: 3.3064e-04 - mae: 0.0152\n",
            "Epoch 642/1000\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 3.2463e-04 - mse: 3.2463e-04 - mae: 0.0151\n",
            "Epoch 643/1000\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 3.2019e-04 - mse: 3.2019e-04 - mae: 0.0150\n",
            "Epoch 644/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 3.1595e-04 - mse: 3.1595e-04 - mae: 0.0148\n",
            "Epoch 645/1000\n",
            "2/2 [==============================] - 0s 20ms/step - loss: 3.4747e-04 - mse: 3.4747e-04 - mae: 0.0161\n",
            "Epoch 646/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 3.1943e-04 - mse: 3.1943e-04 - mae: 0.0156\n",
            "Epoch 647/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 3.1960e-04 - mse: 3.1960e-04 - mae: 0.0157\n",
            "Epoch 648/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.9825e-04 - mse: 2.9825e-04 - mae: 0.0144\n",
            "Epoch 649/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 3.2418e-04 - mse: 3.2418e-04 - mae: 0.0154\n",
            "Epoch 650/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.8523e-04 - mse: 2.8523e-04 - mae: 0.0142\n",
            "Epoch 651/1000\n",
            "2/2 [==============================] - 0s 20ms/step - loss: 2.9617e-04 - mse: 2.9617e-04 - mae: 0.0148\n",
            "Epoch 652/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 3.1151e-04 - mse: 3.1151e-04 - mae: 0.0152\n",
            "Epoch 653/1000\n",
            "2/2 [==============================] - 0s 16ms/step - loss: 2.7447e-04 - mse: 2.7447e-04 - mae: 0.0139\n",
            "Epoch 654/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 2.6765e-04 - mse: 2.6765e-04 - mae: 0.0139\n",
            "Epoch 655/1000\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 2.8052e-04 - mse: 2.8052e-04 - mae: 0.0145\n",
            "Epoch 656/1000\n",
            "2/2 [==============================] - 0s 20ms/step - loss: 2.7554e-04 - mse: 2.7554e-04 - mae: 0.0144\n",
            "Epoch 657/1000\n",
            "2/2 [==============================] - 0s 17ms/step - loss: 2.6029e-04 - mse: 2.6029e-04 - mae: 0.0135\n",
            "Epoch 658/1000\n",
            "2/2 [==============================] - 0s 23ms/step - loss: 2.5555e-04 - mse: 2.5555e-04 - mae: 0.0134\n",
            "Epoch 659/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 2.5203e-04 - mse: 2.5203e-04 - mae: 0.0133\n",
            "Epoch 660/1000\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 2.4902e-04 - mse: 2.4902e-04 - mae: 0.0132\n",
            "Epoch 661/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 2.4138e-04 - mse: 2.4138e-04 - mae: 0.0132\n",
            "Epoch 662/1000\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 2.3783e-04 - mse: 2.3783e-04 - mae: 0.0131\n",
            "Epoch 663/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.3749e-04 - mse: 2.3749e-04 - mae: 0.0129\n",
            "Epoch 664/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.3472e-04 - mse: 2.3472e-04 - mae: 0.0128\n",
            "Epoch 665/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.2747e-04 - mse: 2.2747e-04 - mae: 0.0128\n",
            "Epoch 666/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.2777e-04 - mse: 2.2777e-04 - mae: 0.0127\n",
            "Epoch 667/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.2017e-04 - mse: 2.2017e-04 - mae: 0.0126\n",
            "Epoch 668/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.2122e-04 - mse: 2.2122e-04 - mae: 0.0124\n",
            "Epoch 669/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.1806e-04 - mse: 2.1806e-04 - mae: 0.0123\n",
            "Epoch 670/1000\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 2.3681e-04 - mse: 2.3681e-04 - mae: 0.0131\n",
            "Epoch 671/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 2.3218e-04 - mse: 2.3218e-04 - mae: 0.0130\n",
            "Epoch 672/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 2.0776e-04 - mse: 2.0776e-04 - mae: 0.0120\n",
            "Epoch 673/1000\n",
            "2/2 [==============================] - 0s 28ms/step - loss: 2.0216e-04 - mse: 2.0216e-04 - mae: 0.0120\n",
            "Epoch 674/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 2.1107e-04 - mse: 2.1107e-04 - mae: 0.0125\n",
            "Epoch 675/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 2.0774e-04 - mse: 2.0774e-04 - mae: 0.0125\n",
            "Epoch 676/1000\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 2.0413e-04 - mse: 2.0413e-04 - mae: 0.0125\n",
            "Epoch 677/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 1.9419e-04 - mse: 1.9419e-04 - mae: 0.0117\n",
            "Epoch 678/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.0059e-04 - mse: 2.0059e-04 - mae: 0.0123\n",
            "Epoch 679/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 1.8501e-04 - mse: 1.8501e-04 - mae: 0.0115\n",
            "Epoch 680/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 1.9252e-04 - mse: 1.9252e-04 - mae: 0.0119\n",
            "Epoch 681/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.8192e-04 - mse: 1.8192e-04 - mae: 0.0113\n",
            "Epoch 682/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.7633e-04 - mse: 1.7633e-04 - mae: 0.0113\n",
            "Epoch 683/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.7552e-04 - mse: 1.7552e-04 - mae: 0.0112\n",
            "Epoch 684/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 1.8385e-04 - mse: 1.8385e-04 - mae: 0.0117\n",
            "Epoch 685/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.7129e-04 - mse: 1.7129e-04 - mae: 0.0109\n",
            "Epoch 686/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.8724e-04 - mse: 1.8724e-04 - mae: 0.0117\n",
            "Epoch 687/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.7378e-04 - mse: 1.7378e-04 - mae: 0.0113\n",
            "Epoch 688/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.6408e-04 - mse: 1.6408e-04 - mae: 0.0107\n",
            "Epoch 689/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.7039e-04 - mse: 1.7039e-04 - mae: 0.0113\n",
            "Epoch 690/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 1.6388e-04 - mse: 1.6388e-04 - mae: 0.0110\n",
            "Epoch 691/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 1.5690e-04 - mse: 1.5690e-04 - mae: 0.0105\n",
            "Epoch 692/1000\n",
            "2/2 [==============================] - 0s 16ms/step - loss: 1.5354e-04 - mse: 1.5354e-04 - mae: 0.0104\n",
            "Epoch 693/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 1.4992e-04 - mse: 1.4992e-04 - mae: 0.0104\n",
            "Epoch 694/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.5833e-04 - mse: 1.5833e-04 - mae: 0.0109\n",
            "Epoch 695/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 1.4827e-04 - mse: 1.4827e-04 - mae: 0.0101\n",
            "Epoch 696/1000\n",
            "2/2 [==============================] - 0s 16ms/step - loss: 1.5384e-04 - mse: 1.5384e-04 - mae: 0.0109\n",
            "Epoch 697/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 1.4363e-04 - mse: 1.4363e-04 - mae: 0.0100\n",
            "Epoch 698/1000\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 1.4115e-04 - mse: 1.4115e-04 - mae: 0.0100\n",
            "Epoch 699/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 1.4443e-04 - mse: 1.4443e-04 - mae: 0.0105\n",
            "Epoch 700/1000\n",
            "2/2 [==============================] - 0s 17ms/step - loss: 1.5067e-04 - mse: 1.5067e-04 - mae: 0.0105\n",
            "Epoch 701/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.3521e-04 - mse: 1.3521e-04 - mae: 0.0098\n",
            "Epoch 702/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 1.3296e-04 - mse: 1.3296e-04 - mae: 0.0097\n",
            "Epoch 703/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 1.3132e-04 - mse: 1.3132e-04 - mae: 0.0096\n",
            "Epoch 704/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 1.2727e-04 - mse: 1.2727e-04 - mae: 0.0096\n",
            "Epoch 705/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 1.2743e-04 - mse: 1.2743e-04 - mae: 0.0095\n",
            "Epoch 706/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 1.2353e-04 - mse: 1.2353e-04 - mae: 0.0095\n",
            "Epoch 707/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 1.2945e-04 - mse: 1.2945e-04 - mae: 0.0099\n",
            "Epoch 708/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 1.2890e-04 - mse: 1.2890e-04 - mae: 0.0099\n",
            "Epoch 709/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 1.3230e-04 - mse: 1.3230e-04 - mae: 0.0098\n",
            "Epoch 710/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 1.1764e-04 - mse: 1.1764e-04 - mae: 0.0092\n",
            "Epoch 711/1000\n",
            "2/2 [==============================] - 0s 16ms/step - loss: 1.2299e-04 - mse: 1.2299e-04 - mae: 0.0098\n",
            "Epoch 712/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 1.1519e-04 - mse: 1.1519e-04 - mae: 0.0089\n",
            "Epoch 713/1000\n",
            "2/2 [==============================] - 0s 22ms/step - loss: 1.1148e-04 - mse: 1.1148e-04 - mae: 0.0089\n",
            "Epoch 714/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 1.0954e-04 - mse: 1.0954e-04 - mae: 0.0089\n",
            "Epoch 715/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 1.0789e-04 - mse: 1.0789e-04 - mae: 0.0088\n",
            "Epoch 716/1000\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 1.1325e-04 - mse: 1.1325e-04 - mae: 0.0092\n",
            "Epoch 717/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 1.0489e-04 - mse: 1.0489e-04 - mae: 0.0087\n",
            "Epoch 718/1000\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 1.0321e-04 - mse: 1.0321e-04 - mae: 0.0086\n",
            "Epoch 719/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 1.0146e-04 - mse: 1.0146e-04 - mae: 0.0085\n",
            "Epoch 720/1000\n",
            "2/2 [==============================] - 0s 18ms/step - loss: 1.0797e-04 - mse: 1.0797e-04 - mae: 0.0089\n",
            "Epoch 721/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 1.0060e-04 - mse: 1.0060e-04 - mae: 0.0084\n",
            "Epoch 722/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.0901e-04 - mse: 1.0901e-04 - mae: 0.0089\n",
            "Epoch 723/1000\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 1.0178e-04 - mse: 1.0178e-04 - mae: 0.0088\n",
            "Epoch 724/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 9.9596e-05 - mse: 9.9596e-05 - mae: 0.0085\n",
            "Epoch 725/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 9.7599e-05 - mse: 9.7599e-05 - mae: 0.0086\n",
            "Epoch 726/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 9.8504e-05 - mse: 9.8504e-05 - mae: 0.0087\n",
            "Epoch 727/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 9.1646e-05 - mse: 9.1646e-05 - mae: 0.0080\n",
            "Epoch 728/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 9.8363e-05 - mse: 9.8363e-05 - mae: 0.0085\n",
            "Epoch 729/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 9.2401e-05 - mse: 9.2401e-05 - mae: 0.0084\n",
            "Epoch 730/1000\n",
            "2/2 [==============================] - 0s 17ms/step - loss: 8.7875e-05 - mse: 8.7875e-05 - mae: 0.0078\n",
            "Epoch 731/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 8.6232e-05 - mse: 8.6232e-05 - mae: 0.0078\n",
            "Epoch 732/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 8.9301e-05 - mse: 8.9301e-05 - mae: 0.0082\n",
            "Epoch 733/1000\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 8.7592e-05 - mse: 8.7592e-05 - mae: 0.0081\n",
            "Epoch 734/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 8.7078e-05 - mse: 8.7078e-05 - mae: 0.0081\n",
            "Epoch 735/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 8.4782e-05 - mse: 8.4782e-05 - mae: 0.0080\n",
            "Epoch 736/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 8.2535e-05 - mse: 8.2535e-05 - mae: 0.0078\n",
            "Epoch 737/1000\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 7.7645e-05 - mse: 7.7645e-05 - mae: 0.0075\n",
            "Epoch 738/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 7.7209e-05 - mse: 7.7209e-05 - mae: 0.0074\n",
            "Epoch 739/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 7.6698e-05 - mse: 7.6698e-05 - mae: 0.0074\n",
            "Epoch 740/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 7.5705e-05 - mse: 7.5705e-05 - mae: 0.0073\n",
            "Epoch 741/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 8.2472e-05 - mse: 8.2472e-05 - mae: 0.0078\n",
            "Epoch 742/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 7.6523e-05 - mse: 7.6523e-05 - mae: 0.0075\n",
            "Epoch 743/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 7.1081e-05 - mse: 7.1081e-05 - mae: 0.0071\n",
            "Epoch 744/1000\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 7.5286e-05 - mse: 7.5286e-05 - mae: 0.0075\n",
            "Epoch 745/1000\n",
            "2/2 [==============================] - 0s 16ms/step - loss: 7.2852e-05 - mse: 7.2852e-05 - mae: 0.0073\n",
            "Epoch 746/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 6.8572e-05 - mse: 6.8572e-05 - mae: 0.0070\n",
            "Epoch 747/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 6.7958e-05 - mse: 6.7958e-05 - mae: 0.0069\n",
            "Epoch 748/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 6.7035e-05 - mse: 6.7035e-05 - mae: 0.0069\n",
            "Epoch 749/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 7.3781e-05 - mse: 7.3781e-05 - mae: 0.0074\n",
            "Epoch 750/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 6.4668e-05 - mse: 6.4668e-05 - mae: 0.0068\n",
            "Epoch 751/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 6.3077e-05 - mse: 6.3077e-05 - mae: 0.0068\n",
            "Epoch 752/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 6.1977e-05 - mse: 6.1977e-05 - mae: 0.0067\n",
            "Epoch 753/1000\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 6.1830e-05 - mse: 6.1830e-05 - mae: 0.0066\n",
            "Epoch 754/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 6.4786e-05 - mse: 6.4786e-05 - mae: 0.0071\n",
            "Epoch 755/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 6.0614e-05 - mse: 6.0614e-05 - mae: 0.0065\n",
            "Epoch 756/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 5.8597e-05 - mse: 5.8597e-05 - mae: 0.0065\n",
            "Epoch 757/1000\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 5.7765e-05 - mse: 5.7765e-05 - mae: 0.0065\n",
            "Epoch 758/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 5.7348e-05 - mse: 5.7348e-05 - mae: 0.0064\n",
            "Epoch 759/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 5.6588e-05 - mse: 5.6588e-05 - mae: 0.0063\n",
            "Epoch 760/1000\n",
            "2/2 [==============================] - 0s 17ms/step - loss: 5.6182e-05 - mse: 5.6182e-05 - mae: 0.0063\n",
            "Epoch 761/1000\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 5.4987e-05 - mse: 5.4987e-05 - mae: 0.0062\n",
            "Epoch 762/1000\n",
            "2/2 [==============================] - 0s 18ms/step - loss: 5.3692e-05 - mse: 5.3692e-05 - mae: 0.0062\n",
            "Epoch 763/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 5.3275e-05 - mse: 5.3275e-05 - mae: 0.0061\n",
            "Epoch 764/1000\n",
            "2/2 [==============================] - 0s 16ms/step - loss: 5.6010e-05 - mse: 5.6010e-05 - mae: 0.0065\n",
            "Epoch 765/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 5.2271e-05 - mse: 5.2271e-05 - mae: 0.0060\n",
            "Epoch 766/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 5.0673e-05 - mse: 5.0673e-05 - mae: 0.0060\n",
            "Epoch 767/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 5.2722e-05 - mse: 5.2722e-05 - mae: 0.0063\n",
            "Epoch 768/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 5.1916e-05 - mse: 5.1916e-05 - mae: 0.0063\n",
            "Epoch 769/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 4.8580e-05 - mse: 4.8580e-05 - mae: 0.0059\n",
            "Epoch 770/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 4.8356e-05 - mse: 4.8356e-05 - mae: 0.0058\n",
            "Epoch 771/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 5.0446e-05 - mse: 5.0446e-05 - mae: 0.0062\n",
            "Epoch 772/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 4.6347e-05 - mse: 4.6347e-05 - mae: 0.0058\n",
            "Epoch 773/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 5.0095e-05 - mse: 5.0095e-05 - mae: 0.0060\n",
            "Epoch 774/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 4.5627e-05 - mse: 4.5627e-05 - mae: 0.0057\n",
            "Epoch 775/1000\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 4.7049e-05 - mse: 4.7049e-05 - mae: 0.0060\n",
            "Epoch 776/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 4.3536e-05 - mse: 4.3536e-05 - mae: 0.0056\n",
            "Epoch 777/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 4.2894e-05 - mse: 4.2894e-05 - mae: 0.0056\n",
            "Epoch 778/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 4.3045e-05 - mse: 4.3045e-05 - mae: 0.0055\n",
            "Epoch 779/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 4.4704e-05 - mse: 4.4704e-05 - mae: 0.0058\n",
            "Epoch 780/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 4.3342e-05 - mse: 4.3342e-05 - mae: 0.0056\n",
            "Epoch 781/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 4.1034e-05 - mse: 4.1034e-05 - mae: 0.0054\n",
            "Epoch 782/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 4.2498e-05 - mse: 4.2498e-05 - mae: 0.0056\n",
            "Epoch 783/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 4.1685e-05 - mse: 4.1685e-05 - mae: 0.0056\n",
            "Epoch 784/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 4.1429e-05 - mse: 4.1429e-05 - mae: 0.0056\n",
            "Epoch 785/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.8634e-05 - mse: 3.8634e-05 - mae: 0.0052\n",
            "Epoch 786/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.7531e-05 - mse: 3.7531e-05 - mae: 0.0052\n",
            "Epoch 787/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.9305e-05 - mse: 3.9305e-05 - mae: 0.0054\n",
            "Epoch 788/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.6313e-05 - mse: 3.6313e-05 - mae: 0.0051\n",
            "Epoch 789/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.8222e-05 - mse: 3.8222e-05 - mae: 0.0053\n",
            "Epoch 790/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.7344e-05 - mse: 3.7344e-05 - mae: 0.0053\n",
            "Epoch 791/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.4886e-05 - mse: 3.4886e-05 - mae: 0.0050\n",
            "Epoch 792/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.4316e-05 - mse: 3.4316e-05 - mae: 0.0050\n",
            "Epoch 793/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.5901e-05 - mse: 3.5901e-05 - mae: 0.0052\n",
            "Epoch 794/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.3786e-05 - mse: 3.3786e-05 - mae: 0.0049\n",
            "Epoch 795/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.4778e-05 - mse: 3.4778e-05 - mae: 0.0051\n",
            "Epoch 796/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.2973e-05 - mse: 3.2973e-05 - mae: 0.0048\n",
            "Epoch 797/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.3699e-05 - mse: 3.3699e-05 - mae: 0.0050\n",
            "Epoch 798/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.1925e-05 - mse: 3.1925e-05 - mae: 0.0048\n",
            "Epoch 799/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.0934e-05 - mse: 3.0934e-05 - mae: 0.0047\n",
            "Epoch 800/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.3723e-05 - mse: 3.3723e-05 - mae: 0.0049\n",
            "Epoch 801/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.1685e-05 - mse: 3.1685e-05 - mae: 0.0049\n",
            "Epoch 802/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.1928e-05 - mse: 3.1928e-05 - mae: 0.0049\n",
            "Epoch 803/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.1209e-05 - mse: 3.1209e-05 - mae: 0.0048\n",
            "Epoch 804/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.8920e-05 - mse: 2.8920e-05 - mae: 0.0046\n",
            "Epoch 805/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.9623e-05 - mse: 2.9623e-05 - mae: 0.0047\n",
            "Epoch 806/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.8100e-05 - mse: 2.8100e-05 - mae: 0.0045\n",
            "Epoch 807/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.7711e-05 - mse: 2.7711e-05 - mae: 0.0044\n",
            "Epoch 808/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.7080e-05 - mse: 2.7080e-05 - mae: 0.0044\n",
            "Epoch 809/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.8539e-05 - mse: 2.8539e-05 - mae: 0.0047\n",
            "Epoch 810/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.8382e-05 - mse: 2.8382e-05 - mae: 0.0046\n",
            "Epoch 811/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.5912e-05 - mse: 2.5912e-05 - mae: 0.0043\n",
            "Epoch 812/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.5500e-05 - mse: 2.5500e-05 - mae: 0.0043\n",
            "Epoch 813/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.5040e-05 - mse: 2.5040e-05 - mae: 0.0042\n",
            "Epoch 814/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.4766e-05 - mse: 2.4766e-05 - mae: 0.0042\n",
            "Epoch 815/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 2.4793e-05 - mse: 2.4793e-05 - mae: 0.0042\n",
            "Epoch 816/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.6799e-05 - mse: 2.6799e-05 - mae: 0.0044\n",
            "Epoch 817/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.3685e-05 - mse: 2.3685e-05 - mae: 0.0041\n",
            "Epoch 818/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.3329e-05 - mse: 2.3329e-05 - mae: 0.0041\n",
            "Epoch 819/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.3337e-05 - mse: 2.3337e-05 - mae: 0.0041\n",
            "Epoch 820/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.2866e-05 - mse: 2.2866e-05 - mae: 0.0040\n",
            "Epoch 821/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.2694e-05 - mse: 2.2694e-05 - mae: 0.0040\n",
            "Epoch 822/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.2294e-05 - mse: 2.2294e-05 - mae: 0.0039\n",
            "Epoch 823/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.1851e-05 - mse: 2.1851e-05 - mae: 0.0039\n",
            "Epoch 824/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.1364e-05 - mse: 2.1364e-05 - mae: 0.0039\n",
            "Epoch 825/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.2505e-05 - mse: 2.2505e-05 - mae: 0.0042\n",
            "Epoch 826/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.0682e-05 - mse: 2.0682e-05 - mae: 0.0039\n",
            "Epoch 827/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.3176e-05 - mse: 2.3176e-05 - mae: 0.0042\n",
            "Epoch 828/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.0523e-05 - mse: 2.0523e-05 - mae: 0.0038\n",
            "Epoch 829/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.9877e-05 - mse: 1.9877e-05 - mae: 0.0038\n",
            "Epoch 830/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.1971e-05 - mse: 2.1971e-05 - mae: 0.0040\n",
            "Epoch 831/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.9367e-05 - mse: 1.9367e-05 - mae: 0.0037\n",
            "Epoch 832/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.9865e-05 - mse: 1.9865e-05 - mae: 0.0039\n",
            "Epoch 833/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.9936e-05 - mse: 1.9936e-05 - mae: 0.0039\n",
            "Epoch 834/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 1.8730e-05 - mse: 1.8730e-05 - mae: 0.0036\n",
            "Epoch 835/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.8424e-05 - mse: 1.8424e-05 - mae: 0.0036\n",
            "Epoch 836/1000\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 1.8969e-05 - mse: 1.8969e-05 - mae: 0.0038\n",
            "Epoch 837/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.7876e-05 - mse: 1.7876e-05 - mae: 0.0036\n",
            "Epoch 838/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.7277e-05 - mse: 1.7277e-05 - mae: 0.0035\n",
            "Epoch 839/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 1.8184e-05 - mse: 1.8184e-05 - mae: 0.0037\n",
            "Epoch 840/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.7760e-05 - mse: 1.7760e-05 - mae: 0.0037\n",
            "Epoch 841/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.7606e-05 - mse: 1.7606e-05 - mae: 0.0036\n",
            "Epoch 842/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.7490e-05 - mse: 1.7490e-05 - mae: 0.0036\n",
            "Epoch 843/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.6819e-05 - mse: 1.6819e-05 - mae: 0.0035\n",
            "Epoch 844/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.6078e-05 - mse: 1.6078e-05 - mae: 0.0034\n",
            "Epoch 845/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.5612e-05 - mse: 1.5612e-05 - mae: 0.0034\n",
            "Epoch 846/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.6288e-05 - mse: 1.6288e-05 - mae: 0.0035\n",
            "Epoch 847/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.5324e-05 - mse: 1.5324e-05 - mae: 0.0033\n",
            "Epoch 848/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.5638e-05 - mse: 1.5638e-05 - mae: 0.0034\n",
            "Epoch 849/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.4693e-05 - mse: 1.4693e-05 - mae: 0.0033\n",
            "Epoch 850/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.4477e-05 - mse: 1.4477e-05 - mae: 0.0032\n",
            "Epoch 851/1000\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 1.4413e-05 - mse: 1.4413e-05 - mae: 0.0032\n",
            "Epoch 852/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.4088e-05 - mse: 1.4088e-05 - mae: 0.0032\n",
            "Epoch 853/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.4101e-05 - mse: 1.4101e-05 - mae: 0.0031\n",
            "Epoch 854/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.5001e-05 - mse: 1.5001e-05 - mae: 0.0033\n",
            "Epoch 855/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.4365e-05 - mse: 1.4365e-05 - mae: 0.0033\n",
            "Epoch 856/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.3389e-05 - mse: 1.3389e-05 - mae: 0.0031\n",
            "Epoch 857/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 1.4025e-05 - mse: 1.4025e-05 - mae: 0.0032\n",
            "Epoch 858/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.2878e-05 - mse: 1.2878e-05 - mae: 0.0030\n",
            "Epoch 859/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.2901e-05 - mse: 1.2901e-05 - mae: 0.0030\n",
            "Epoch 860/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.3383e-05 - mse: 1.3383e-05 - mae: 0.0032\n",
            "Epoch 861/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.2497e-05 - mse: 1.2497e-05 - mae: 0.0030\n",
            "Epoch 862/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.2141e-05 - mse: 1.2141e-05 - mae: 0.0030\n",
            "Epoch 863/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.2718e-05 - mse: 1.2718e-05 - mae: 0.0031\n",
            "Epoch 864/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.1975e-05 - mse: 1.1975e-05 - mae: 0.0029\n",
            "Epoch 865/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.1768e-05 - mse: 1.1768e-05 - mae: 0.0029\n",
            "Epoch 866/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.1612e-05 - mse: 1.1612e-05 - mae: 0.0028\n",
            "Epoch 867/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.1373e-05 - mse: 1.1373e-05 - mae: 0.0028\n",
            "Epoch 868/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.1091e-05 - mse: 1.1091e-05 - mae: 0.0028\n",
            "Epoch 869/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.1755e-05 - mse: 1.1755e-05 - mae: 0.0030\n",
            "Epoch 870/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.0982e-05 - mse: 1.0982e-05 - mae: 0.0028\n",
            "Epoch 871/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.0778e-05 - mse: 1.0778e-05 - mae: 0.0028\n",
            "Epoch 872/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.1234e-05 - mse: 1.1234e-05 - mae: 0.0029\n",
            "Epoch 873/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.0331e-05 - mse: 1.0331e-05 - mae: 0.0027\n",
            "Epoch 874/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.0327e-05 - mse: 1.0327e-05 - mae: 0.0027\n",
            "Epoch 875/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.0156e-05 - mse: 1.0156e-05 - mae: 0.0027\n",
            "Epoch 876/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 9.8612e-06 - mse: 9.8612e-06 - mae: 0.0027\n",
            "Epoch 877/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 9.6827e-06 - mse: 9.6827e-06 - mae: 0.0026\n",
            "Epoch 878/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 9.5641e-06 - mse: 9.5641e-06 - mae: 0.0026\n",
            "Epoch 879/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 9.9396e-06 - mse: 9.9396e-06 - mae: 0.0027\n",
            "Epoch 880/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 9.2928e-06 - mse: 9.2928e-06 - mae: 0.0026\n",
            "Epoch 881/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 9.8694e-06 - mse: 9.8694e-06 - mae: 0.0027\n",
            "Epoch 882/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 9.0363e-06 - mse: 9.0363e-06 - mae: 0.0025\n",
            "Epoch 883/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 9.0414e-06 - mse: 9.0414e-06 - mae: 0.0025\n",
            "Epoch 884/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 8.7677e-06 - mse: 8.7677e-06 - mae: 0.0025\n",
            "Epoch 885/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 8.6364e-06 - mse: 8.6364e-06 - mae: 0.0025\n",
            "Epoch 886/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 9.6244e-06 - mse: 9.6244e-06 - mae: 0.0027\n",
            "Epoch 887/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 8.8762e-06 - mse: 8.8762e-06 - mae: 0.0026\n",
            "Epoch 888/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 8.4072e-06 - mse: 8.4072e-06 - mae: 0.0024\n",
            "Epoch 889/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 8.1371e-06 - mse: 8.1371e-06 - mae: 0.0024\n",
            "Epoch 890/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 8.5293e-06 - mse: 8.5293e-06 - mae: 0.0025\n",
            "Epoch 891/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 8.4486e-06 - mse: 8.4486e-06 - mae: 0.0025\n",
            "Epoch 892/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 8.7082e-06 - mse: 8.7082e-06 - mae: 0.0025\n",
            "Epoch 893/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 8.2145e-06 - mse: 8.2145e-06 - mae: 0.0025\n",
            "Epoch 894/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 7.5303e-06 - mse: 7.5303e-06 - mae: 0.0023\n",
            "Epoch 895/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 7.4081e-06 - mse: 7.4081e-06 - mae: 0.0023\n",
            "Epoch 896/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 7.8718e-06 - mse: 7.8718e-06 - mae: 0.0025\n",
            "Epoch 897/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 7.2289e-06 - mse: 7.2289e-06 - mae: 0.0023\n",
            "Epoch 898/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 7.1071e-06 - mse: 7.1071e-06 - mae: 0.0023\n",
            "Epoch 899/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 7.8983e-06 - mse: 7.8983e-06 - mae: 0.0024\n",
            "Epoch 900/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 7.0618e-06 - mse: 7.0618e-06 - mae: 0.0022\n",
            "Epoch 901/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 6.8019e-06 - mse: 6.8019e-06 - mae: 0.0022\n",
            "Epoch 902/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 7.0200e-06 - mse: 7.0200e-06 - mae: 0.0023\n",
            "Epoch 903/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 6.7030e-06 - mse: 6.7030e-06 - mae: 0.0022\n",
            "Epoch 904/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 6.4979e-06 - mse: 6.4979e-06 - mae: 0.0022\n",
            "Epoch 905/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 6.7957e-06 - mse: 6.7957e-06 - mae: 0.0023\n",
            "Epoch 906/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 6.3163e-06 - mse: 6.3163e-06 - mae: 0.0021\n",
            "Epoch 907/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 6.3228e-06 - mse: 6.3228e-06 - mae: 0.0021\n",
            "Epoch 908/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 6.1888e-06 - mse: 6.1888e-06 - mae: 0.0021\n",
            "Epoch 909/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 6.0497e-06 - mse: 6.0497e-06 - mae: 0.0021\n",
            "Epoch 910/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 6.2974e-06 - mse: 6.2974e-06 - mae: 0.0022\n",
            "Epoch 911/1000\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 6.1359e-06 - mse: 6.1359e-06 - mae: 0.0021\n",
            "Epoch 912/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 6.0722e-06 - mse: 6.0722e-06 - mae: 0.0021\n",
            "Epoch 913/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 6.3346e-06 - mse: 6.3346e-06 - mae: 0.0021\n",
            "Epoch 914/1000\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 5.5973e-06 - mse: 5.5973e-06 - mae: 0.0020\n",
            "Epoch 915/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 5.5142e-06 - mse: 5.5142e-06 - mae: 0.0020\n",
            "Epoch 916/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 5.4833e-06 - mse: 5.4833e-06 - mae: 0.0020\n",
            "Epoch 917/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 5.7696e-06 - mse: 5.7696e-06 - mae: 0.0021\n",
            "Epoch 918/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 5.2758e-06 - mse: 5.2758e-06 - mae: 0.0019\n",
            "Epoch 919/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 5.5372e-06 - mse: 5.5372e-06 - mae: 0.0021\n",
            "Epoch 920/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 5.1739e-06 - mse: 5.1739e-06 - mae: 0.0019\n",
            "Epoch 921/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 5.3620e-06 - mse: 5.3620e-06 - mae: 0.0020\n",
            "Epoch 922/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 5.0634e-06 - mse: 5.0634e-06 - mae: 0.0019\n",
            "Epoch 923/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 4.9084e-06 - mse: 4.9084e-06 - mae: 0.0019\n",
            "Epoch 924/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 5.1315e-06 - mse: 5.1315e-06 - mae: 0.0020\n",
            "Epoch 925/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 5.0474e-06 - mse: 5.0474e-06 - mae: 0.0020\n",
            "Epoch 926/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 4.7526e-06 - mse: 4.7526e-06 - mae: 0.0018\n",
            "Epoch 927/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 4.6179e-06 - mse: 4.6179e-06 - mae: 0.0018\n",
            "Epoch 928/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 4.7826e-06 - mse: 4.7826e-06 - mae: 0.0019\n",
            "Epoch 929/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 4.5583e-06 - mse: 4.5583e-06 - mae: 0.0018\n",
            "Epoch 930/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 4.7546e-06 - mse: 4.7546e-06 - mae: 0.0019\n",
            "Epoch 931/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 4.6695e-06 - mse: 4.6695e-06 - mae: 0.0019\n",
            "Epoch 932/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 4.5438e-06 - mse: 4.5438e-06 - mae: 0.0018\n",
            "Epoch 933/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 4.7610e-06 - mse: 4.7610e-06 - mae: 0.0019\n",
            "Epoch 934/1000\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 4.4096e-06 - mse: 4.4096e-06 - mae: 0.0018\n",
            "Epoch 935/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 4.3950e-06 - mse: 4.3950e-06 - mae: 0.0018\n",
            "Epoch 936/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 4.1089e-06 - mse: 4.1089e-06 - mae: 0.0017\n",
            "Epoch 937/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 4.0353e-06 - mse: 4.0353e-06 - mae: 0.0017\n",
            "Epoch 938/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 4.1728e-06 - mse: 4.1728e-06 - mae: 0.0018\n",
            "Epoch 939/1000\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 3.8479e-06 - mse: 3.8479e-06 - mae: 0.0017\n",
            "Epoch 940/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 4.3076e-06 - mse: 4.3076e-06 - mae: 0.0018\n",
            "Epoch 941/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 3.7817e-06 - mse: 3.7817e-06 - mae: 0.0016\n",
            "Epoch 942/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 4.1468e-06 - mse: 4.1468e-06 - mae: 0.0017\n",
            "Epoch 943/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 3.7231e-06 - mse: 3.7231e-06 - mae: 0.0016\n",
            "Epoch 944/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.5782e-06 - mse: 3.5782e-06 - mae: 0.0016\n",
            "Epoch 945/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.5258e-06 - mse: 3.5258e-06 - mae: 0.0016\n",
            "Epoch 946/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.7265e-06 - mse: 3.7265e-06 - mae: 0.0017\n",
            "Epoch 947/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 3.4372e-06 - mse: 3.4372e-06 - mae: 0.0016\n",
            "Epoch 948/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.5821e-06 - mse: 3.5821e-06 - mae: 0.0016\n",
            "Epoch 949/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.3526e-06 - mse: 3.3526e-06 - mae: 0.0015\n",
            "Epoch 950/1000\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 3.4861e-06 - mse: 3.4861e-06 - mae: 0.0016\n",
            "Epoch 951/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.6594e-06 - mse: 3.6594e-06 - mae: 0.0017\n",
            "Epoch 952/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.3230e-06 - mse: 3.3230e-06 - mae: 0.0016\n",
            "Epoch 953/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.1172e-06 - mse: 3.1172e-06 - mae: 0.0015\n",
            "Epoch 954/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.4818e-06 - mse: 3.4818e-06 - mae: 0.0016\n",
            "Epoch 955/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 3.3074e-06 - mse: 3.3074e-06 - mae: 0.0015\n",
            "Epoch 956/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.9874e-06 - mse: 2.9874e-06 - mae: 0.0015\n",
            "Epoch 957/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.9471e-06 - mse: 2.9471e-06 - mae: 0.0015\n",
            "Epoch 958/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.9017e-06 - mse: 2.9017e-06 - mae: 0.0014\n",
            "Epoch 959/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 3.0301e-06 - mse: 3.0301e-06 - mae: 0.0015\n",
            "Epoch 960/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.9821e-06 - mse: 2.9821e-06 - mae: 0.0015\n",
            "Epoch 961/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.9461e-06 - mse: 2.9461e-06 - mae: 0.0015\n",
            "Epoch 962/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.7367e-06 - mse: 2.7367e-06 - mae: 0.0014\n",
            "Epoch 963/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.9752e-06 - mse: 2.9752e-06 - mae: 0.0015\n",
            "Epoch 964/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.6530e-06 - mse: 2.6530e-06 - mae: 0.0014\n",
            "Epoch 965/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.8132e-06 - mse: 2.8132e-06 - mae: 0.0014\n",
            "Epoch 966/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.6242e-06 - mse: 2.6242e-06 - mae: 0.0013\n",
            "Epoch 967/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.5859e-06 - mse: 2.5859e-06 - mae: 0.0013\n",
            "Epoch 968/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.6557e-06 - mse: 2.6557e-06 - mae: 0.0014\n",
            "Epoch 969/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.4624e-06 - mse: 2.4624e-06 - mae: 0.0013\n",
            "Epoch 970/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.5544e-06 - mse: 2.5544e-06 - mae: 0.0014\n",
            "Epoch 971/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.4362e-06 - mse: 2.4362e-06 - mae: 0.0013\n",
            "Epoch 972/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.3499e-06 - mse: 2.3499e-06 - mae: 0.0013\n",
            "Epoch 973/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.6165e-06 - mse: 2.6165e-06 - mae: 0.0014\n",
            "Epoch 974/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.3052e-06 - mse: 2.3052e-06 - mae: 0.0013\n",
            "Epoch 975/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.2845e-06 - mse: 2.2845e-06 - mae: 0.0013\n",
            "Epoch 976/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.3352e-06 - mse: 2.3352e-06 - mae: 0.0013\n",
            "Epoch 977/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.2256e-06 - mse: 2.2256e-06 - mae: 0.0013\n",
            "Epoch 978/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.2960e-06 - mse: 2.2960e-06 - mae: 0.0013\n",
            "Epoch 979/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.2852e-06 - mse: 2.2852e-06 - mae: 0.0013\n",
            "Epoch 980/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.1098e-06 - mse: 2.1098e-06 - mae: 0.0012\n",
            "Epoch 981/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.0777e-06 - mse: 2.0777e-06 - mae: 0.0012\n",
            "Epoch 982/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 2.0245e-06 - mse: 2.0245e-06 - mae: 0.0012\n",
            "Epoch 983/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.9958e-06 - mse: 1.9958e-06 - mae: 0.0012\n",
            "Epoch 984/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 2.0046e-06 - mse: 2.0046e-06 - mae: 0.0012\n",
            "Epoch 985/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.9384e-06 - mse: 1.9384e-06 - mae: 0.0012\n",
            "Epoch 986/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.9160e-06 - mse: 1.9160e-06 - mae: 0.0012\n",
            "Epoch 987/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.9160e-06 - mse: 1.9160e-06 - mae: 0.0012\n",
            "Epoch 988/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.9723e-06 - mse: 1.9723e-06 - mae: 0.0012\n",
            "Epoch 989/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.8593e-06 - mse: 1.8593e-06 - mae: 0.0011\n",
            "Epoch 990/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.7992e-06 - mse: 1.7992e-06 - mae: 0.0011\n",
            "Epoch 991/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.7940e-06 - mse: 1.7940e-06 - mae: 0.0011\n",
            "Epoch 992/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.9493e-06 - mse: 1.9493e-06 - mae: 0.0012\n",
            "Epoch 993/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.8619e-06 - mse: 1.8619e-06 - mae: 0.0012\n",
            "Epoch 994/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.6966e-06 - mse: 1.6966e-06 - mae: 0.0011\n",
            "Epoch 995/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.6760e-06 - mse: 1.6760e-06 - mae: 0.0011\n",
            "Epoch 996/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.7585e-06 - mse: 1.7585e-06 - mae: 0.0012\n",
            "Epoch 997/1000\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 1.6545e-06 - mse: 1.6545e-06 - mae: 0.0011\n",
            "Epoch 998/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.5985e-06 - mse: 1.5985e-06 - mae: 0.0011\n",
            "Epoch 999/1000\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 1.6897e-06 - mse: 1.6897e-06 - mae: 0.0011\n",
            "Epoch 1000/1000\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 1.6485e-06 - mse: 1.6485e-06 - mae: 0.0011\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7fa01d7cd4b0>"
            ]
          },
          "metadata": {},
          "execution_count": 22
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.get_weights()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BrI_g3p2euZT",
        "outputId": "aeb14a14-8fd2-4dee-a62e-f10710216bae"
      },
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[array([[3.0006714]], dtype=float32), array([5.997183], dtype=float32)]"
            ]
          },
          "metadata": {},
          "execution_count": 23
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.evaluate(X_train, y_train)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3U_dVoHIe0Dn",
        "outputId": "a99ae105-be97-432c-9787-fa2726afe676"
      },
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 1s 608ms/step - loss: 1.5331e-06 - mse: 1.5331e-06 - mae: 0.0011\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[1.533086674498918e-06, 1.533086674498918e-06, 0.0010509490966796875]"
            ]
          },
          "metadata": {},
          "execution_count": 24
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# test\n",
        "X_test = [10]\n",
        "\n",
        "model.predict(X_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "50zfTgsyfGjh",
        "outputId": "eb62b8ff-e0c8-458b-e803-2c47fbaf39aa"
      },
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 0s 101ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[36.0039]], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 25
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 선형회귀(예측)"
      ],
      "metadata": {
        "id": "6FwOfeBPfW9X"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import tensorflow as tf\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt"
      ],
      "metadata": {
        "id": "lNPDLGYSj075"
      },
      "execution_count": 26,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "mpg = pd.read_csv('auto-mpg.csv', na_values=['?', '-', ';']) #데이터 읽을때 ?나 - 같은 표시 nan으로"
      ],
      "metadata": {
        "id": "gblpeWrAj1c9"
      },
      "execution_count": 64,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df = mpg.copy()\n",
        "df.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "uychHhxqkwCs",
        "outputId": "2a13aa9f-678f-4eaf-e4a6-7c8813559247"
      },
      "execution_count": 65,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "    mpg  cylinders  displacement  horsepower  weight  acceleration  \\\n",
              "0  18.0          8         307.0       130.0    3504          12.0   \n",
              "1  15.0          8         350.0       165.0    3693          11.5   \n",
              "2  18.0          8         318.0       150.0    3436          11.0   \n",
              "3  16.0          8         304.0       150.0    3433          12.0   \n",
              "4  17.0          8         302.0       140.0    3449          10.5   \n",
              "\n",
              "   model year  origin                   car name  \n",
              "0          70       1  chevrolet chevelle malibu  \n",
              "1          70       1          buick skylark 320  \n",
              "2          70       1         plymouth satellite  \n",
              "3          70       1              amc rebel sst  \n",
              "4          70       1                ford torino  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-b9dd567f-f56c-41c1-901e-308209d4ffff\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>mpg</th>\n",
              "      <th>cylinders</th>\n",
              "      <th>displacement</th>\n",
              "      <th>horsepower</th>\n",
              "      <th>weight</th>\n",
              "      <th>acceleration</th>\n",
              "      <th>model year</th>\n",
              "      <th>origin</th>\n",
              "      <th>car name</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>18.0</td>\n",
              "      <td>8</td>\n",
              "      <td>307.0</td>\n",
              "      <td>130.0</td>\n",
              "      <td>3504</td>\n",
              "      <td>12.0</td>\n",
              "      <td>70</td>\n",
              "      <td>1</td>\n",
              "      <td>chevrolet chevelle malibu</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>15.0</td>\n",
              "      <td>8</td>\n",
              "      <td>350.0</td>\n",
              "      <td>165.0</td>\n",
              "      <td>3693</td>\n",
              "      <td>11.5</td>\n",
              "      <td>70</td>\n",
              "      <td>1</td>\n",
              "      <td>buick skylark 320</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>18.0</td>\n",
              "      <td>8</td>\n",
              "      <td>318.0</td>\n",
              "      <td>150.0</td>\n",
              "      <td>3436</td>\n",
              "      <td>11.0</td>\n",
              "      <td>70</td>\n",
              "      <td>1</td>\n",
              "      <td>plymouth satellite</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>16.0</td>\n",
              "      <td>8</td>\n",
              "      <td>304.0</td>\n",
              "      <td>150.0</td>\n",
              "      <td>3433</td>\n",
              "      <td>12.0</td>\n",
              "      <td>70</td>\n",
              "      <td>1</td>\n",
              "      <td>amc rebel sst</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>17.0</td>\n",
              "      <td>8</td>\n",
              "      <td>302.0</td>\n",
              "      <td>140.0</td>\n",
              "      <td>3449</td>\n",
              "      <td>10.5</td>\n",
              "      <td>70</td>\n",
              "      <td>1</td>\n",
              "      <td>ford torino</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-b9dd567f-f56c-41c1-901e-308209d4ffff')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-b9dd567f-f56c-41c1-901e-308209d4ffff button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-b9dd567f-f56c-41c1-901e-308209d4ffff');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 65
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df.isna().sum()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "spcHyg1Rk2Am",
        "outputId": "986c7914-78fe-4b7f-bd30-09a8684e5586"
      },
      "execution_count": 67,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "mpg             0\n",
              "cylinders       0\n",
              "displacement    0\n",
              "horsepower      6\n",
              "weight          0\n",
              "acceleration    0\n",
              "model year      0\n",
              "origin          0\n",
              "car name        0\n",
              "dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 67
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df.info()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qMUu7DWek_PY",
        "outputId": "15873b67-d9bb-4b1e-fe23-e31305c4a1ad"
      },
      "execution_count": 68,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 398 entries, 0 to 397\n",
            "Data columns (total 9 columns):\n",
            " #   Column        Non-Null Count  Dtype  \n",
            "---  ------        --------------  -----  \n",
            " 0   mpg           398 non-null    float64\n",
            " 1   cylinders     398 non-null    int64  \n",
            " 2   displacement  398 non-null    float64\n",
            " 3   horsepower    392 non-null    float64\n",
            " 4   weight        398 non-null    int64  \n",
            " 5   acceleration  398 non-null    float64\n",
            " 6   model year    398 non-null    int64  \n",
            " 7   origin        398 non-null    int64  \n",
            " 8   car name      398 non-null    object \n",
            "dtypes: float64(4), int64(4), object(1)\n",
            "memory usage: 28.1+ KB\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df=df.dropna()"
      ],
      "metadata": {
        "id": "8-SJxNMHnRzU"
      },
      "execution_count": 69,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df.isna().sum()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-gzFFWEYnZTF",
        "outputId": "3c2054fc-affd-4767-dd15-f5c131a30631"
      },
      "execution_count": 70,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "mpg             0\n",
              "cylinders       0\n",
              "displacement    0\n",
              "horsepower      0\n",
              "weight          0\n",
              "acceleration    0\n",
              "model year      0\n",
              "origin          0\n",
              "car name        0\n",
              "dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 70
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "X = df.iloc[:,1:6] #cylinders  displacement    horsepower   weight   acceleration\n",
        "y = df['mpg']"
      ],
      "metadata": {
        "id": "vrSyZC3Ylae6"
      },
      "execution_count": 71,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)"
      ],
      "metadata": {
        "id": "-wPoh5N7lAYw"
      },
      "execution_count": 72,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.preprocessing import StandardScaler\n",
        "\n",
        "ss = StandardScaler()\n",
        "X_train_s = ss.fit_transform(X_train)\n",
        "X_test_s = ss.transform(X_test)"
      ],
      "metadata": {
        "id": "oVdHaXv5mZCL"
      },
      "execution_count": 73,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow import keras\n",
        "from keras import layers\n",
        "\n",
        "model = keras.Sequential([\n",
        "    layers.Dense(units=32, activation='relu', input_shape=(5,)),\n",
        "    layers.Dense(units=16, activation='relu'),\n",
        "    layers.Dense(units=1) #종속변수가 mpg로 하나이므로 출력 layer는 1\n",
        "])"
      ],
      "metadata": {
        "id": "Zw53UOQdmgb1"
      },
      "execution_count": 76,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UfOUNqiapz5N",
        "outputId": "1c59546d-070a-4570-ee85-511fa550ba13"
      },
      "execution_count": 77,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " dense_16 (Dense)            (None, 32)                192       \n",
            "                                                                 \n",
            " dense_17 (Dense)            (None, 16)                528       \n",
            "                                                                 \n",
            " dense_18 (Dense)            (None, 1)                 17        \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 737\n",
            "Trainable params: 737\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "keras.utils.plot_model(model)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 369
        },
        "id": "DMRmJ1jFqttp",
        "outputId": "3eb553a5-bc41-47ab-9161-179ba2f6de90"
      },
      "execution_count": 78,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 78
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# compile"
      ],
      "metadata": {
        "id": "tlmC9by1rGuq"
      }
    },
    {
      "cell_type": "code",
      "source": [
        " model.compile(\n",
        "    optimizer='adam',\n",
        "    loss='mse',\n",
        "    metrics=['mse', 'mae']\n",
        " )"
      ],
      "metadata": {
        "id": "iaSzni7vrW_d"
      },
      "execution_count": 80,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "epochs = 100\n",
        "batch_size = 32\n",
        "\n",
        "model.fit(\n",
        "    X_train_s, y_train,\n",
        "    epochs = epochs,\n",
        "    batch_size = batch_size,\n",
        "    verbose = 1\n",
        ")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gufsQobjrmNc",
        "outputId": "c19165ee-27c0-42f6-97f6-faa164269cc7"
      },
      "execution_count": 83,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "10/10 [==============================] - 1s 4ms/step - loss: 596.0766 - mse: 596.0766 - mae: 23.1586\n",
            "Epoch 2/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 549.4010 - mse: 549.4010 - mae: 22.1697\n",
            "Epoch 3/100\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 484.4275 - mse: 484.4275 - mae: 20.7332\n",
            "Epoch 4/100\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 402.6164 - mse: 402.6164 - mae: 18.8053\n",
            "Epoch 5/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 311.8309 - mse: 311.8309 - mae: 16.4104\n",
            "Epoch 6/100\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 227.5779 - mse: 227.5779 - mae: 13.7350\n",
            "Epoch 7/100\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 157.3080 - mse: 157.3080 - mae: 11.0674\n",
            "Epoch 8/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 109.1187 - mse: 109.1187 - mae: 8.8506\n",
            "Epoch 9/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 84.0368 - mse: 84.0368 - mae: 7.4240\n",
            "Epoch 10/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 74.6849 - mse: 74.6849 - mae: 6.8599\n",
            "Epoch 11/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 71.3038 - mse: 71.3038 - mae: 6.6415\n",
            "Epoch 12/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 68.9504 - mse: 68.9504 - mae: 6.5124\n",
            "Epoch 13/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 66.3058 - mse: 66.3058 - mae: 6.3764\n",
            "Epoch 14/100\n",
            "10/10 [==============================] - 0s 4ms/step - loss: 63.7899 - mse: 63.7899 - mae: 6.2555\n",
            "Epoch 15/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 61.5416 - mse: 61.5416 - mae: 6.1324\n",
            "Epoch 16/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 59.2416 - mse: 59.2416 - mae: 6.0181\n",
            "Epoch 17/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 57.1065 - mse: 57.1065 - mae: 5.9060\n",
            "Epoch 18/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 55.0972 - mse: 55.0972 - mae: 5.8055\n",
            "Epoch 19/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 53.0414 - mse: 53.0414 - mae: 5.6945\n",
            "Epoch 20/100\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 51.0607 - mse: 51.0607 - mae: 5.5936\n",
            "Epoch 21/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 49.1738 - mse: 49.1738 - mae: 5.4974\n",
            "Epoch 22/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 47.2389 - mse: 47.2389 - mae: 5.3928\n",
            "Epoch 23/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 45.3417 - mse: 45.3417 - mae: 5.2875\n",
            "Epoch 24/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 43.4454 - mse: 43.4454 - mae: 5.1788\n",
            "Epoch 25/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 41.6850 - mse: 41.6850 - mae: 5.0768\n",
            "Epoch 26/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 39.9392 - mse: 39.9392 - mae: 4.9732\n",
            "Epoch 27/100\n",
            "10/10 [==============================] - 0s 4ms/step - loss: 38.3186 - mse: 38.3186 - mae: 4.8697\n",
            "Epoch 28/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 36.7657 - mse: 36.7657 - mae: 4.7665\n",
            "Epoch 29/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 35.3040 - mse: 35.3040 - mae: 4.6744\n",
            "Epoch 30/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 33.8927 - mse: 33.8927 - mae: 4.5811\n",
            "Epoch 31/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 32.3629 - mse: 32.3629 - mae: 4.4719\n",
            "Epoch 32/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 31.0667 - mse: 31.0667 - mae: 4.3802\n",
            "Epoch 33/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 29.6765 - mse: 29.6765 - mae: 4.2783\n",
            "Epoch 34/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 28.6155 - mse: 28.6155 - mae: 4.1987\n",
            "Epoch 35/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 27.4760 - mse: 27.4760 - mae: 4.1026\n",
            "Epoch 36/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 26.3932 - mse: 26.3932 - mae: 4.0169\n",
            "Epoch 37/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 25.4262 - mse: 25.4262 - mae: 3.9300\n",
            "Epoch 38/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 24.5721 - mse: 24.5721 - mae: 3.8495\n",
            "Epoch 39/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 23.7905 - mse: 23.7905 - mae: 3.7766\n",
            "Epoch 40/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 23.1011 - mse: 23.1011 - mae: 3.7134\n",
            "Epoch 41/100\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 22.4132 - mse: 22.4132 - mae: 3.6438\n",
            "Epoch 42/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 21.8604 - mse: 21.8604 - mae: 3.5835\n",
            "Epoch 43/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 21.3242 - mse: 21.3242 - mae: 3.5283\n",
            "Epoch 44/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 20.7918 - mse: 20.7918 - mae: 3.4686\n",
            "Epoch 45/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 20.3515 - mse: 20.3515 - mae: 3.4204\n",
            "Epoch 46/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 19.8545 - mse: 19.8545 - mae: 3.3718\n",
            "Epoch 47/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 19.4833 - mse: 19.4833 - mae: 3.3297\n",
            "Epoch 48/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 19.0820 - mse: 19.0820 - mae: 3.2911\n",
            "Epoch 49/100\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 18.7341 - mse: 18.7341 - mae: 3.2569\n",
            "Epoch 50/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 18.4544 - mse: 18.4544 - mae: 3.2277\n",
            "Epoch 51/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 18.1396 - mse: 18.1396 - mae: 3.1942\n",
            "Epoch 52/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 17.8520 - mse: 17.8520 - mae: 3.1668\n",
            "Epoch 53/100\n",
            "10/10 [==============================] - 0s 4ms/step - loss: 17.6011 - mse: 17.6011 - mae: 3.1383\n",
            "Epoch 54/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 17.4018 - mse: 17.4018 - mae: 3.1210\n",
            "Epoch 55/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 17.1566 - mse: 17.1566 - mae: 3.0961\n",
            "Epoch 56/100\n",
            "10/10 [==============================] - 0s 4ms/step - loss: 16.9638 - mse: 16.9638 - mae: 3.0754\n",
            "Epoch 57/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 16.7995 - mse: 16.7995 - mae: 3.0584\n",
            "Epoch 58/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 16.6098 - mse: 16.6098 - mae: 3.0402\n",
            "Epoch 59/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 16.4712 - mse: 16.4712 - mae: 3.0246\n",
            "Epoch 60/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 16.3343 - mse: 16.3343 - mae: 3.0057\n",
            "Epoch 61/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 16.2082 - mse: 16.2082 - mae: 2.9955\n",
            "Epoch 62/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 16.1012 - mse: 16.1012 - mae: 2.9877\n",
            "Epoch 63/100\n",
            "10/10 [==============================] - 0s 4ms/step - loss: 15.9960 - mse: 15.9960 - mae: 2.9780\n",
            "Epoch 64/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 15.9049 - mse: 15.9049 - mae: 2.9675\n",
            "Epoch 65/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 15.8312 - mse: 15.8312 - mae: 2.9626\n",
            "Epoch 66/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 15.7231 - mse: 15.7231 - mae: 2.9503\n",
            "Epoch 67/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 15.6485 - mse: 15.6485 - mae: 2.9469\n",
            "Epoch 68/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 15.6212 - mse: 15.6212 - mae: 2.9420\n",
            "Epoch 69/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 15.5461 - mse: 15.5461 - mae: 2.9405\n",
            "Epoch 70/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 15.4462 - mse: 15.4462 - mae: 2.9319\n",
            "Epoch 71/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 15.5012 - mse: 15.5012 - mae: 2.9311\n",
            "Epoch 72/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 15.3446 - mse: 15.3446 - mae: 2.9156\n",
            "Epoch 73/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 15.3088 - mse: 15.3088 - mae: 2.9173\n",
            "Epoch 74/100\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 15.2556 - mse: 15.2556 - mae: 2.9136\n",
            "Epoch 75/100\n",
            "10/10 [==============================] - 0s 2ms/step - loss: 15.2398 - mse: 15.2398 - mae: 2.9170\n",
            "Epoch 76/100\n",
            "10/10 [==============================] - 0s 4ms/step - loss: 15.2214 - mse: 15.2214 - mae: 2.9139\n",
            "Epoch 77/100\n",
            "10/10 [==============================] - 0s 4ms/step - loss: 15.1348 - mse: 15.1348 - mae: 2.9071\n",
            "Epoch 78/100\n",
            "10/10 [==============================] - 0s 4ms/step - loss: 15.1095 - mse: 15.1095 - mae: 2.8977\n",
            "Epoch 79/100\n",
            "10/10 [==============================] - 0s 5ms/step - loss: 15.0912 - mse: 15.0912 - mae: 2.8929\n",
            "Epoch 80/100\n",
            "10/10 [==============================] - 0s 5ms/step - loss: 15.0406 - mse: 15.0406 - mae: 2.8935\n",
            "Epoch 81/100\n",
            "10/10 [==============================] - 0s 4ms/step - loss: 15.0509 - mse: 15.0509 - mae: 2.9003\n",
            "Epoch 82/100\n",
            "10/10 [==============================] - 0s 4ms/step - loss: 14.9902 - mse: 14.9902 - mae: 2.8943\n",
            "Epoch 83/100\n",
            "10/10 [==============================] - 0s 4ms/step - loss: 14.9978 - mse: 14.9978 - mae: 2.8874\n",
            "Epoch 84/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 14.9626 - mse: 14.9626 - mae: 2.8849\n",
            "Epoch 85/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 14.9196 - mse: 14.9196 - mae: 2.8835\n",
            "Epoch 86/100\n",
            "10/10 [==============================] - 0s 4ms/step - loss: 14.8923 - mse: 14.8923 - mae: 2.8825\n",
            "Epoch 87/100\n",
            "10/10 [==============================] - 0s 4ms/step - loss: 14.9009 - mse: 14.9009 - mae: 2.8807\n",
            "Epoch 88/100\n",
            "10/10 [==============================] - 0s 4ms/step - loss: 14.8554 - mse: 14.8554 - mae: 2.8793\n",
            "Epoch 89/100\n",
            "10/10 [==============================] - 0s 4ms/step - loss: 14.8433 - mse: 14.8433 - mae: 2.8796\n",
            "Epoch 90/100\n",
            "10/10 [==============================] - 0s 4ms/step - loss: 14.8223 - mse: 14.8223 - mae: 2.8751\n",
            "Epoch 91/100\n",
            "10/10 [==============================] - 0s 4ms/step - loss: 14.8162 - mse: 14.8162 - mae: 2.8699\n",
            "Epoch 92/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 14.7918 - mse: 14.7918 - mae: 2.8689\n",
            "Epoch 93/100\n",
            "10/10 [==============================] - 0s 4ms/step - loss: 14.7617 - mse: 14.7617 - mae: 2.8674\n",
            "Epoch 94/100\n",
            "10/10 [==============================] - 0s 4ms/step - loss: 14.7513 - mse: 14.7513 - mae: 2.8649\n",
            "Epoch 95/100\n",
            "10/10 [==============================] - 0s 4ms/step - loss: 14.7587 - mse: 14.7587 - mae: 2.8654\n",
            "Epoch 96/100\n",
            "10/10 [==============================] - 0s 4ms/step - loss: 14.7532 - mse: 14.7532 - mae: 2.8642\n",
            "Epoch 97/100\n",
            "10/10 [==============================] - 0s 4ms/step - loss: 14.7196 - mse: 14.7196 - mae: 2.8603\n",
            "Epoch 98/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 14.7022 - mse: 14.7022 - mae: 2.8622\n",
            "Epoch 99/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 14.7537 - mse: 14.7537 - mae: 2.8618\n",
            "Epoch 100/100\n",
            "10/10 [==============================] - 0s 3ms/step - loss: 14.7771 - mse: 14.7771 - mae: 2.8657\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7fa0065efc40>"
            ]
          },
          "metadata": {},
          "execution_count": 83
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.evaluate(X_train_s, y_train)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pGvIGyg3sAPf",
        "outputId": "c5aa1a9b-f7a8-4a54-b4ea-e3fbc1560ce1"
      },
      "execution_count": 85,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "10/10 [==============================] - 0s 6ms/step - loss: 14.6621 - mse: 14.6621 - mae: 2.8584\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[14.662102699279785, 14.662102699279785, 2.858421802520752]"
            ]
          },
          "metadata": {},
          "execution_count": 85
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "y_pred = model.predict(X_test_s)\n",
        "y_pred = y_pred.reshape(-1,)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-2UTeGnqsvXE",
        "outputId": "43a6d816-785b-4f47-bac8-f8f0e0e52d88"
      },
      "execution_count": 92,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "3/3 [==============================] - 0s 12ms/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "y_train"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "K6Kx9Zplx9HH",
        "outputId": "a96f7ab9-80ec-40fa-a187-f5ae5ec53fb0"
      },
      "execution_count": 100,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "260    18.6\n",
              "184    25.0\n",
              "174    18.0\n",
              "64     15.0\n",
              "344    39.0\n",
              "       ... \n",
              "72     15.0\n",
              "107    18.0\n",
              "272    23.8\n",
              "352    29.9\n",
              "103    11.0\n",
              "Name: mpg, Length: 313, dtype: float64"
            ]
          },
          "metadata": {},
          "execution_count": 100
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "Md2vhwcS2N1A"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model = keras.Sequential([\n",
        "    layers.Dense(units=32, activation='relu', input_shape=(5,)),\n",
        "    layers.Dense(units=16, activation='relu'),\n",
        "    layers.Dense(units=1) #종속변수가 mpg로 하나이므로 출력 layer는 1\n",
        "])"
      ],
      "metadata": {
        "id": "C8blVT0S2NHL"
      },
      "execution_count": 104,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model.compile(\n",
        "    optimizer='adam',\n",
        "    loss='mse',\n",
        "    metrics=['mse', 'mae']\n",
        " )"
      ],
      "metadata": {
        "id": "ZbptkpRI2NJx"
      },
      "execution_count": 105,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "epochs = 100\n",
        "batch_size = 32\n",
        "\n",
        "history = model.fit(\n",
        "    X_train_s, y_train,\n",
        "    epochs = epochs,\n",
        "    batch_size = batch_size,\n",
        "    validation_split=0.2,\n",
        "    verbose = 1\n",
        ")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "00ZsZ25WyEBu",
        "outputId": "2f1ad197-9e7a-45ae-87dc-426d6195aeee"
      },
      "execution_count": 106,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "8/8 [==============================] - 1s 46ms/step - loss: 584.4835 - mse: 584.4835 - mae: 22.7338 - val_loss: 663.1355 - val_mse: 663.1355 - val_mae: 24.4579\n",
            "Epoch 2/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 575.5333 - mse: 575.5333 - mae: 22.5117 - val_loss: 653.5670 - val_mse: 653.5670 - val_mae: 24.2439\n",
            "Epoch 3/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 565.7418 - mse: 565.7418 - mae: 22.2571 - val_loss: 642.9825 - val_mse: 642.9825 - val_mae: 23.9989\n",
            "Epoch 4/100\n",
            "8/8 [==============================] - 0s 10ms/step - loss: 554.5839 - mse: 554.5839 - mae: 21.9648 - val_loss: 631.0683 - val_mse: 631.0683 - val_mae: 23.7095\n",
            "Epoch 5/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 542.2780 - mse: 542.2780 - mae: 21.6244 - val_loss: 617.3069 - val_mse: 617.3069 - val_mae: 23.3732\n",
            "Epoch 6/100\n",
            "8/8 [==============================] - 0s 12ms/step - loss: 528.1149 - mse: 528.1149 - mae: 21.2226 - val_loss: 601.1207 - val_mse: 601.1207 - val_mae: 22.9807\n",
            "Epoch 7/100\n",
            "8/8 [==============================] - 0s 11ms/step - loss: 511.8915 - mse: 511.8915 - mae: 20.7744 - val_loss: 582.6690 - val_mse: 582.6690 - val_mae: 22.5242\n",
            "Epoch 8/100\n",
            "8/8 [==============================] - 0s 12ms/step - loss: 493.7071 - mse: 493.7071 - mae: 20.2451 - val_loss: 561.2695 - val_mse: 561.2695 - val_mae: 21.9932\n",
            "Epoch 9/100\n",
            "8/8 [==============================] - 0s 12ms/step - loss: 472.5782 - mse: 472.5782 - mae: 19.6291 - val_loss: 534.7214 - val_mse: 534.7214 - val_mae: 21.3460\n",
            "Epoch 10/100\n",
            "8/8 [==============================] - 0s 13ms/step - loss: 447.4916 - mse: 447.4916 - mae: 18.9377 - val_loss: 505.2693 - val_mse: 505.2693 - val_mae: 20.6079\n",
            "Epoch 11/100\n",
            "8/8 [==============================] - 0s 10ms/step - loss: 421.1340 - mse: 421.1340 - mae: 18.1448 - val_loss: 472.7627 - val_mse: 472.7627 - val_mae: 19.8068\n",
            "Epoch 12/100\n",
            "8/8 [==============================] - 0s 12ms/step - loss: 392.0179 - mse: 392.0179 - mae: 17.3450 - val_loss: 437.8725 - val_mse: 437.8725 - val_mae: 18.9424\n",
            "Epoch 13/100\n",
            "8/8 [==============================] - 0s 10ms/step - loss: 360.9401 - mse: 360.9401 - mae: 16.5309 - val_loss: 400.7137 - val_mse: 400.7137 - val_mae: 18.0454\n",
            "Epoch 14/100\n",
            "8/8 [==============================] - 0s 13ms/step - loss: 329.1327 - mse: 329.1327 - mae: 15.7032 - val_loss: 360.2362 - val_mse: 360.2362 - val_mae: 17.0784\n",
            "Epoch 15/100\n",
            "8/8 [==============================] - 0s 12ms/step - loss: 294.6123 - mse: 294.6123 - mae: 14.8439 - val_loss: 317.1249 - val_mse: 317.1249 - val_mae: 16.0005\n",
            "Epoch 16/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 260.0009 - mse: 260.0009 - mae: 13.9972 - val_loss: 272.5577 - val_mse: 272.5577 - val_mae: 14.8134\n",
            "Epoch 17/100\n",
            "8/8 [==============================] - 0s 12ms/step - loss: 223.3109 - mse: 223.3109 - mae: 13.0163 - val_loss: 229.6011 - val_mse: 229.6011 - val_mae: 13.5680\n",
            "Epoch 18/100\n",
            "8/8 [==============================] - 0s 10ms/step - loss: 188.4320 - mse: 188.4320 - mae: 11.9898 - val_loss: 188.7824 - val_mse: 188.7824 - val_mae: 12.2388\n",
            "Epoch 19/100\n",
            "8/8 [==============================] - 0s 8ms/step - loss: 154.7146 - mse: 154.7146 - mae: 10.8441 - val_loss: 151.9485 - val_mse: 151.9485 - val_mae: 10.9098\n",
            "Epoch 20/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 125.4139 - mse: 125.4139 - mae: 9.7132 - val_loss: 119.6707 - val_mse: 119.6707 - val_mae: 9.5666\n",
            "Epoch 21/100\n",
            "8/8 [==============================] - 0s 6ms/step - loss: 99.6513 - mse: 99.6513 - mae: 8.5803 - val_loss: 93.6578 - val_mse: 93.6578 - val_mae: 8.2703\n",
            "Epoch 22/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 79.0640 - mse: 79.0640 - mae: 7.5217 - val_loss: 74.5896 - val_mse: 74.5896 - val_mae: 7.1147\n",
            "Epoch 23/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 64.8235 - mse: 64.8235 - mae: 6.6451 - val_loss: 61.9020 - val_mse: 61.9020 - val_mae: 6.1873\n",
            "Epoch 24/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 54.3483 - mse: 54.3483 - mae: 5.9763 - val_loss: 55.1408 - val_mse: 55.1408 - val_mae: 5.6047\n",
            "Epoch 25/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 48.9019 - mse: 48.9019 - mae: 5.5460 - val_loss: 51.5736 - val_mse: 51.5736 - val_mae: 5.3009\n",
            "Epoch 26/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 45.6573 - mse: 45.6573 - mae: 5.2996 - val_loss: 49.7937 - val_mse: 49.7937 - val_mae: 5.1779\n",
            "Epoch 27/100\n",
            "8/8 [==============================] - 0s 10ms/step - loss: 43.9148 - mse: 43.9148 - mae: 5.1670 - val_loss: 48.4880 - val_mse: 48.4880 - val_mae: 5.1345\n",
            "Epoch 28/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 42.6828 - mse: 42.6828 - mae: 5.0962 - val_loss: 47.0809 - val_mse: 47.0809 - val_mae: 5.0660\n",
            "Epoch 29/100\n",
            "8/8 [==============================] - 0s 8ms/step - loss: 41.4695 - mse: 41.4695 - mae: 5.0123 - val_loss: 45.5038 - val_mse: 45.5038 - val_mae: 4.9753\n",
            "Epoch 30/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 40.3978 - mse: 40.3978 - mae: 4.9492 - val_loss: 43.9405 - val_mse: 43.9405 - val_mae: 4.8784\n",
            "Epoch 31/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 39.3006 - mse: 39.3006 - mae: 4.8836 - val_loss: 42.4195 - val_mse: 42.4195 - val_mae: 4.7849\n",
            "Epoch 32/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 38.2836 - mse: 38.2836 - mae: 4.8213 - val_loss: 41.1320 - val_mse: 41.1320 - val_mae: 4.7067\n",
            "Epoch 33/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 37.3451 - mse: 37.3451 - mae: 4.7640 - val_loss: 39.6709 - val_mse: 39.6709 - val_mae: 4.6170\n",
            "Epoch 34/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 36.3247 - mse: 36.3247 - mae: 4.7064 - val_loss: 38.3800 - val_mse: 38.3800 - val_mae: 4.5363\n",
            "Epoch 35/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 35.4729 - mse: 35.4729 - mae: 4.6581 - val_loss: 37.2347 - val_mse: 37.2347 - val_mae: 4.4638\n",
            "Epoch 36/100\n",
            "8/8 [==============================] - 0s 10ms/step - loss: 34.6056 - mse: 34.6056 - mae: 4.5975 - val_loss: 36.0509 - val_mse: 36.0509 - val_mae: 4.3849\n",
            "Epoch 37/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 33.7915 - mse: 33.7915 - mae: 4.5424 - val_loss: 35.0604 - val_mse: 35.0604 - val_mae: 4.3165\n",
            "Epoch 38/100\n",
            "8/8 [==============================] - 0s 8ms/step - loss: 33.0481 - mse: 33.0481 - mae: 4.4926 - val_loss: 33.9192 - val_mse: 33.9192 - val_mae: 4.2485\n",
            "Epoch 39/100\n",
            "8/8 [==============================] - 0s 10ms/step - loss: 32.2152 - mse: 32.2152 - mae: 4.4373 - val_loss: 32.9755 - val_mse: 32.9755 - val_mae: 4.1882\n",
            "Epoch 40/100\n",
            "8/8 [==============================] - 0s 8ms/step - loss: 31.5122 - mse: 31.5122 - mae: 4.3929 - val_loss: 32.0858 - val_mse: 32.0858 - val_mae: 4.1297\n",
            "Epoch 41/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 30.8256 - mse: 30.8256 - mae: 4.3461 - val_loss: 31.2395 - val_mse: 31.2395 - val_mae: 4.0744\n",
            "Epoch 42/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 30.1512 - mse: 30.1512 - mae: 4.3012 - val_loss: 30.3954 - val_mse: 30.3954 - val_mae: 4.0191\n",
            "Epoch 43/100\n",
            "8/8 [==============================] - 0s 8ms/step - loss: 29.5194 - mse: 29.5194 - mae: 4.2528 - val_loss: 29.6119 - val_mse: 29.6119 - val_mae: 3.9697\n",
            "Epoch 44/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 28.9432 - mse: 28.9432 - mae: 4.2082 - val_loss: 28.8544 - val_mse: 28.8544 - val_mae: 3.9241\n",
            "Epoch 45/100\n",
            "8/8 [==============================] - 0s 8ms/step - loss: 28.3962 - mse: 28.3962 - mae: 4.1606 - val_loss: 28.1824 - val_mse: 28.1824 - val_mae: 3.8761\n",
            "Epoch 46/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 27.8610 - mse: 27.8610 - mae: 4.1180 - val_loss: 27.4926 - val_mse: 27.4926 - val_mae: 3.8285\n",
            "Epoch 47/100\n",
            "8/8 [==============================] - 0s 10ms/step - loss: 27.3911 - mse: 27.3911 - mae: 4.0848 - val_loss: 26.7428 - val_mse: 26.7428 - val_mae: 3.7775\n",
            "Epoch 48/100\n",
            "8/8 [==============================] - 0s 8ms/step - loss: 26.8844 - mse: 26.8844 - mae: 4.0488 - val_loss: 26.1424 - val_mse: 26.1424 - val_mae: 3.7306\n",
            "Epoch 49/100\n",
            "8/8 [==============================] - 0s 10ms/step - loss: 26.4034 - mse: 26.4034 - mae: 4.0027 - val_loss: 25.5972 - val_mse: 25.5972 - val_mae: 3.6899\n",
            "Epoch 50/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 25.9714 - mse: 25.9714 - mae: 3.9623 - val_loss: 25.0623 - val_mse: 25.0623 - val_mae: 3.6461\n",
            "Epoch 51/100\n",
            "8/8 [==============================] - 0s 8ms/step - loss: 25.6083 - mse: 25.6083 - mae: 3.9230 - val_loss: 24.5709 - val_mse: 24.5709 - val_mae: 3.5993\n",
            "Epoch 52/100\n",
            "8/8 [==============================] - 0s 10ms/step - loss: 25.1878 - mse: 25.1878 - mae: 3.8860 - val_loss: 23.9469 - val_mse: 23.9469 - val_mae: 3.5582\n",
            "Epoch 53/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 24.7750 - mse: 24.7750 - mae: 3.8512 - val_loss: 23.4858 - val_mse: 23.4858 - val_mae: 3.5173\n",
            "Epoch 54/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 24.4203 - mse: 24.4203 - mae: 3.8190 - val_loss: 23.0592 - val_mse: 23.0592 - val_mae: 3.4744\n",
            "Epoch 55/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 24.0711 - mse: 24.0711 - mae: 3.7885 - val_loss: 22.5252 - val_mse: 22.5252 - val_mae: 3.4341\n",
            "Epoch 56/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 23.6976 - mse: 23.6976 - mae: 3.7537 - val_loss: 22.1868 - val_mse: 22.1868 - val_mae: 3.3976\n",
            "Epoch 57/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 23.3930 - mse: 23.3930 - mae: 3.7177 - val_loss: 21.8024 - val_mse: 21.8024 - val_mae: 3.3626\n",
            "Epoch 58/100\n",
            "8/8 [==============================] - 0s 8ms/step - loss: 23.0653 - mse: 23.0653 - mae: 3.6856 - val_loss: 21.3572 - val_mse: 21.3572 - val_mae: 3.3243\n",
            "Epoch 59/100\n",
            "8/8 [==============================] - 0s 8ms/step - loss: 22.7742 - mse: 22.7742 - mae: 3.6561 - val_loss: 20.8984 - val_mse: 20.8984 - val_mae: 3.2879\n",
            "Epoch 60/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 22.4962 - mse: 22.4962 - mae: 3.6305 - val_loss: 20.5049 - val_mse: 20.5049 - val_mae: 3.2457\n",
            "Epoch 61/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 22.2145 - mse: 22.2145 - mae: 3.5979 - val_loss: 20.2614 - val_mse: 20.2614 - val_mae: 3.2103\n",
            "Epoch 62/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 21.8821 - mse: 21.8821 - mae: 3.5646 - val_loss: 19.8859 - val_mse: 19.8859 - val_mae: 3.1793\n",
            "Epoch 63/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 21.6342 - mse: 21.6342 - mae: 3.5400 - val_loss: 19.5165 - val_mse: 19.5165 - val_mae: 3.1483\n",
            "Epoch 64/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 21.3649 - mse: 21.3649 - mae: 3.5059 - val_loss: 19.2544 - val_mse: 19.2544 - val_mae: 3.1229\n",
            "Epoch 65/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 21.1721 - mse: 21.1721 - mae: 3.4801 - val_loss: 18.9919 - val_mse: 18.9919 - val_mae: 3.0958\n",
            "Epoch 66/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 20.8708 - mse: 20.8708 - mae: 3.4566 - val_loss: 18.7049 - val_mse: 18.7049 - val_mae: 3.0711\n",
            "Epoch 67/100\n",
            "8/8 [==============================] - 0s 10ms/step - loss: 20.6837 - mse: 20.6837 - mae: 3.4400 - val_loss: 18.4828 - val_mse: 18.4828 - val_mae: 3.0498\n",
            "Epoch 68/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 20.4750 - mse: 20.4750 - mae: 3.4132 - val_loss: 18.2302 - val_mse: 18.2302 - val_mae: 3.0278\n",
            "Epoch 69/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 20.2609 - mse: 20.2609 - mae: 3.3959 - val_loss: 17.9634 - val_mse: 17.9634 - val_mae: 3.0063\n",
            "Epoch 70/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 20.0339 - mse: 20.0339 - mae: 3.3711 - val_loss: 17.7591 - val_mse: 17.7591 - val_mae: 2.9883\n",
            "Epoch 71/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 19.8530 - mse: 19.8530 - mae: 3.3545 - val_loss: 17.5643 - val_mse: 17.5643 - val_mae: 2.9718\n",
            "Epoch 72/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 19.6724 - mse: 19.6724 - mae: 3.3402 - val_loss: 17.4050 - val_mse: 17.4050 - val_mae: 2.9577\n",
            "Epoch 73/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 19.4669 - mse: 19.4669 - mae: 3.3224 - val_loss: 17.2215 - val_mse: 17.2215 - val_mae: 2.9428\n",
            "Epoch 74/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 19.3009 - mse: 19.3009 - mae: 3.3061 - val_loss: 17.0288 - val_mse: 17.0288 - val_mae: 2.9286\n",
            "Epoch 75/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 19.1533 - mse: 19.1533 - mae: 3.2883 - val_loss: 16.8841 - val_mse: 16.8841 - val_mae: 2.9151\n",
            "Epoch 76/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 18.9774 - mse: 18.9774 - mae: 3.2707 - val_loss: 16.7049 - val_mse: 16.7049 - val_mae: 2.9010\n",
            "Epoch 77/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 18.8054 - mse: 18.8054 - mae: 3.2570 - val_loss: 16.5306 - val_mse: 16.5306 - val_mae: 2.8856\n",
            "Epoch 78/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 18.6614 - mse: 18.6614 - mae: 3.2473 - val_loss: 16.4122 - val_mse: 16.4122 - val_mae: 2.8755\n",
            "Epoch 79/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 18.5523 - mse: 18.5523 - mae: 3.2371 - val_loss: 16.3865 - val_mse: 16.3865 - val_mae: 2.8686\n",
            "Epoch 80/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 18.3763 - mse: 18.3763 - mae: 3.2184 - val_loss: 16.1772 - val_mse: 16.1772 - val_mae: 2.8573\n",
            "Epoch 81/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 18.2553 - mse: 18.2553 - mae: 3.2050 - val_loss: 15.9733 - val_mse: 15.9733 - val_mae: 2.8462\n",
            "Epoch 82/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 18.1178 - mse: 18.1178 - mae: 3.1887 - val_loss: 15.8415 - val_mse: 15.8415 - val_mae: 2.8374\n",
            "Epoch 83/100\n",
            "8/8 [==============================] - 0s 8ms/step - loss: 17.9889 - mse: 17.9889 - mae: 3.1745 - val_loss: 15.7389 - val_mse: 15.7389 - val_mae: 2.8282\n",
            "Epoch 84/100\n",
            "8/8 [==============================] - 0s 11ms/step - loss: 17.8720 - mse: 17.8720 - mae: 3.1618 - val_loss: 15.6004 - val_mse: 15.6004 - val_mae: 2.8212\n",
            "Epoch 85/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 17.7229 - mse: 17.7229 - mae: 3.1486 - val_loss: 15.5590 - val_mse: 15.5590 - val_mae: 2.8172\n",
            "Epoch 86/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 17.6500 - mse: 17.6500 - mae: 3.1422 - val_loss: 15.4827 - val_mse: 15.4827 - val_mae: 2.8152\n",
            "Epoch 87/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 17.4835 - mse: 17.4835 - mae: 3.1269 - val_loss: 15.4386 - val_mse: 15.4386 - val_mae: 2.8158\n",
            "Epoch 88/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 17.3826 - mse: 17.3826 - mae: 3.1184 - val_loss: 15.3769 - val_mse: 15.3769 - val_mae: 2.8155\n",
            "Epoch 89/100\n",
            "8/8 [==============================] - 0s 10ms/step - loss: 17.2447 - mse: 17.2447 - mae: 3.1074 - val_loss: 15.2531 - val_mse: 15.2531 - val_mae: 2.8088\n",
            "Epoch 90/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 17.1538 - mse: 17.1538 - mae: 3.0939 - val_loss: 15.1179 - val_mse: 15.1179 - val_mae: 2.8002\n",
            "Epoch 91/100\n",
            "8/8 [==============================] - 0s 8ms/step - loss: 17.0641 - mse: 17.0641 - mae: 3.0859 - val_loss: 15.1049 - val_mse: 15.1049 - val_mae: 2.8053\n",
            "Epoch 92/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 16.9643 - mse: 16.9643 - mae: 3.0736 - val_loss: 14.9963 - val_mse: 14.9963 - val_mae: 2.7988\n",
            "Epoch 93/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 16.9226 - mse: 16.9226 - mae: 3.0733 - val_loss: 14.9487 - val_mse: 14.9487 - val_mae: 2.8006\n",
            "Epoch 94/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 16.7511 - mse: 16.7511 - mae: 3.0575 - val_loss: 14.9314 - val_mse: 14.9314 - val_mae: 2.8035\n",
            "Epoch 95/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 16.6966 - mse: 16.6966 - mae: 3.0495 - val_loss: 14.9348 - val_mse: 14.9348 - val_mae: 2.8078\n",
            "Epoch 96/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 16.5872 - mse: 16.5872 - mae: 3.0441 - val_loss: 14.8562 - val_mse: 14.8562 - val_mae: 2.8054\n",
            "Epoch 97/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 16.5112 - mse: 16.5112 - mae: 3.0364 - val_loss: 14.7704 - val_mse: 14.7704 - val_mae: 2.8015\n",
            "Epoch 98/100\n",
            "8/8 [==============================] - 0s 7ms/step - loss: 16.4744 - mse: 16.4744 - mae: 3.0321 - val_loss: 14.6962 - val_mse: 14.6962 - val_mae: 2.7978\n",
            "Epoch 99/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 16.3604 - mse: 16.3604 - mae: 3.0198 - val_loss: 14.6152 - val_mse: 14.6152 - val_mae: 2.7951\n",
            "Epoch 100/100\n",
            "8/8 [==============================] - 0s 9ms/step - loss: 16.2944 - mse: 16.2944 - mae: 3.0159 - val_loss: 14.6423 - val_mse: 14.6423 - val_mae: 2.8019\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "history.history.keys()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pBHPwyXlzANP",
        "outputId": "e7dc398c-1ec1-4cd5-fb0d-523e1aa697e0"
      },
      "execution_count": 107,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "dict_keys(['loss', 'mse', 'mae', 'val_loss', 'val_mse', 'val_mae'])"
            ]
          },
          "metadata": {},
          "execution_count": 107
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "hist = pd.DataFrame(history.history)\n",
        "hist['epoch'] = history.epoch\n",
        "hist"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 423
        },
        "id": "IXyifvg_zLvd",
        "outputId": "2a6898b9-ad48-4322-f863-e841b87e0e69"
      },
      "execution_count": 108,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "          loss         mse        mae    val_loss     val_mse    val_mae  \\\n",
              "0   584.483521  584.483521  22.733809  663.135498  663.135498  24.457890   \n",
              "1   575.533325  575.533325  22.511660  653.566956  653.566956  24.243902   \n",
              "2   565.741821  565.741821  22.257097  642.982544  642.982544  23.998941   \n",
              "3   554.583862  554.583862  21.964846  631.068298  631.068298  23.709518   \n",
              "4   542.278015  542.278015  21.624388  617.306885  617.306885  23.373186   \n",
              "..         ...         ...        ...         ...         ...        ...   \n",
              "95   16.587162   16.587162   3.044141   14.856238   14.856238   2.805383   \n",
              "96   16.511187   16.511187   3.036433   14.770385   14.770385   2.801529   \n",
              "97   16.474401   16.474401   3.032139   14.696231   14.696231   2.797776   \n",
              "98   16.360373   16.360373   3.019833   14.615176   14.615176   2.795108   \n",
              "99   16.294367   16.294367   3.015908   14.642282   14.642282   2.801860   \n",
              "\n",
              "    epoch  \n",
              "0       0  \n",
              "1       1  \n",
              "2       2  \n",
              "3       3  \n",
              "4       4  \n",
              "..    ...  \n",
              "95     95  \n",
              "96     96  \n",
              "97     97  \n",
              "98     98  \n",
              "99     99  \n",
              "\n",
              "[100 rows x 7 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-61076325-7ebd-40e6-9a7b-c3967df57655\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>loss</th>\n",
              "      <th>mse</th>\n",
              "      <th>mae</th>\n",
              "      <th>val_loss</th>\n",
              "      <th>val_mse</th>\n",
              "      <th>val_mae</th>\n",
              "      <th>epoch</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>584.483521</td>\n",
              "      <td>584.483521</td>\n",
              "      <td>22.733809</td>\n",
              "      <td>663.135498</td>\n",
              "      <td>663.135498</td>\n",
              "      <td>24.457890</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>575.533325</td>\n",
              "      <td>575.533325</td>\n",
              "      <td>22.511660</td>\n",
              "      <td>653.566956</td>\n",
              "      <td>653.566956</td>\n",
              "      <td>24.243902</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>565.741821</td>\n",
              "      <td>565.741821</td>\n",
              "      <td>22.257097</td>\n",
              "      <td>642.982544</td>\n",
              "      <td>642.982544</td>\n",
              "      <td>23.998941</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>554.583862</td>\n",
              "      <td>554.583862</td>\n",
              "      <td>21.964846</td>\n",
              "      <td>631.068298</td>\n",
              "      <td>631.068298</td>\n",
              "      <td>23.709518</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>542.278015</td>\n",
              "      <td>542.278015</td>\n",
              "      <td>21.624388</td>\n",
              "      <td>617.306885</td>\n",
              "      <td>617.306885</td>\n",
              "      <td>23.373186</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>95</th>\n",
              "      <td>16.587162</td>\n",
              "      <td>16.587162</td>\n",
              "      <td>3.044141</td>\n",
              "      <td>14.856238</td>\n",
              "      <td>14.856238</td>\n",
              "      <td>2.805383</td>\n",
              "      <td>95</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>96</th>\n",
              "      <td>16.511187</td>\n",
              "      <td>16.511187</td>\n",
              "      <td>3.036433</td>\n",
              "      <td>14.770385</td>\n",
              "      <td>14.770385</td>\n",
              "      <td>2.801529</td>\n",
              "      <td>96</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>97</th>\n",
              "      <td>16.474401</td>\n",
              "      <td>16.474401</td>\n",
              "      <td>3.032139</td>\n",
              "      <td>14.696231</td>\n",
              "      <td>14.696231</td>\n",
              "      <td>2.797776</td>\n",
              "      <td>97</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>98</th>\n",
              "      <td>16.360373</td>\n",
              "      <td>16.360373</td>\n",
              "      <td>3.019833</td>\n",
              "      <td>14.615176</td>\n",
              "      <td>14.615176</td>\n",
              "      <td>2.795108</td>\n",
              "      <td>98</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>99</th>\n",
              "      <td>16.294367</td>\n",
              "      <td>16.294367</td>\n",
              "      <td>3.015908</td>\n",
              "      <td>14.642282</td>\n",
              "      <td>14.642282</td>\n",
              "      <td>2.801860</td>\n",
              "      <td>99</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>100 rows × 7 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-61076325-7ebd-40e6-9a7b-c3967df57655')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-61076325-7ebd-40e6-9a7b-c3967df57655 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-61076325-7ebd-40e6-9a7b-c3967df57655');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 108
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.plot(hist['epoch'], hist['mse'], label='train_mse')\n",
        "plt.plot(hist['epoch'], hist['val_mse'], label='val_mse')\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 430
        },
        "id": "90dEcPwjzYPt",
        "outputId": "f73a35a4-9376-43f2-cfef-d1e9f4cc3478"
      },
      "execution_count": 109,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAigAAAGdCAYAAAA44ojeAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAABRCklEQVR4nO3deXwV9b3/8decNes5IYEkBAiEncgqKATUWqUiolVBrRZ361ZQkWotrUXrhj9rtdrrUm0LtmK9tVdaRRERFBfCKjuy70sSICQn+3LO/P44ySGRLYEkk5O8n4/HPM5k5jsznzNaz7sz3/mOYZqmiYiIiEgzYrO6ABEREZHvU0ARERGRZkcBRURERJodBRQRERFpdhRQREREpNlRQBEREZFmRwFFREREmh0FFBEREWl2HFYXcDoCgQD79+8nNjYWwzCsLkdERETqwDRNCgoKSElJwWY7+TWSsAwo+/fvp1OnTlaXISIiIqdhz549dOzY8aRtwjKgxMbGAsEv6PF4LK5GRERE6sLn89GpU6fQ7/jJhGVAqb6t4/F4FFBERETCTF26Z6iTrIiIiDQ7CigiIiLS7CigiIiISLMTln1QRESk5fD7/VRUVFhdhjQAu92Ow+FokCFAFFBERMQyhYWF7N27F9M0rS5FGkhUVBTt27fH5XKd0X4UUERExBJ+v5+9e/cSFRVFu3btNPBmmDNNk/Lycg4ePMiOHTvo0aPHKQdjOxkFFBERsURFRQWmadKuXTsiIyOtLkcaQGRkJE6nk127dlFeXk5ERMRp70udZEVExFK6ctKynMlVk1r7aZC9iIiIiDQgBRQRERFpdhRQRERELNKlSxf++Mc/Wl1Gs6ROsiIiIvVw4YUXMnDgwAYJFsuWLSM6OvrMi2qBdAWlppIj8I+xsG+F1ZWIiEiYMk2TysrKOrVt164dUVFRjVxReFJAqWnBU7BtPrx9DRzcbHU1IiKtimmaFJdXWjLVdaC4W2+9lYULF/LSSy9hGAaGYTBjxgwMw2DOnDkMHjwYt9vN119/zbZt27jyyitJSkoiJiaGc845h88++6zW/r5/i8cwDP7yl79w9dVXExUVRY8ePfjggw/qVNsXX3yBYRjMnTuXQYMGERkZyUUXXUROTg5z5syhT58+eDwefvrTn1JcXBza7t///jf9+vUjMjKShIQERo4cSVFRUWj9X/7yF/r06UNERAS9e/fm1VdfrVM9Z0q3eGoa+TvY9y3s/xb+cTXcMRe8Ha2uSkSkVSip8JM+da4lx97wxCiiXKf+SXzppZfYvHkzffv25YknngBg/fr1APzqV7/i+eefp2vXrrRp04Y9e/Zw2WWX8fTTT+N2u/n73//OFVdcwaZNm0hNTT3hMX73u9/x3HPP8fvf/54//elPjB8/nl27dhEfH1+n7/L444/zP//zP0RFRXHddddx3XXX4Xa7eeeddygsLOTqq6/mT3/6E4888ggHDhzghhtu4LnnnuPqq6+moKCAr776KhTYZs6cydSpU/mf//kfBg0axMqVK7nzzjuJjo7mlltuqVM9p0sBpSZ3DIz/N0y/FA5tDoaU2z6B6ASrKxMRkWbA6/XicrmIiooiOTkZgI0bNwLwxBNP8KMf/SjUNj4+ngEDBoT+fvLJJ5k1axYffPABEydOPOExbr31Vm644QYAnnnmGV5++WWWLl3KpZdeWqcan3rqKUaMGAHAHXfcwZQpU9i2bRtdu3YF4JprruHzzz8PBZTKykrGjh1L586dAejXr19oX4899hh/+MMfGDt2LABpaWls2LCBP//5zwooTS46AW6aBX8dFQwp71wLN38QDC8iItJoIp12NjwxyrJjn6khQ4bU+ruwsJDHH3+cjz76KBQESkpK2L1790n3079//9B8dHQ0Ho+HnJycOtdRc/ukpCSioqJC4aR62dKlSwEYMGAAF198Mf369WPUqFFccsklXHPNNbRp04aioiK2bdvGHXfcwZ133hnavrKyEq/XW+d6TpcCyvF4OwZDyt9GBTvM/u+N8NN/gePMXnwkIiInZhhGnW6zNFfffxrnoYceYt68eTz//PN0796dyMhIrrnmGsrLy0+6H6fTWetvwzAIBAJ1rqPm9oZhnHR/drudefPmsWjRIj799FP+9Kc/8Zvf/IYlS5aEOu+++eabDB06tNY+7PYzD3Snok6yJ9KuZ/B2jzMatn8OH/8C9LZNEZFWz+Vy4ff7T9num2++4dZbb+Xqq6+mX79+JCcns3PnzsYvsJ4Mw2DEiBH87ne/Y+XKlbhcLmbNmkVSUhIpKSls376d7t2715rS0tIava7wjapNoeNguO4teOc6+PbvkJgOw+61uioREbFQly5dWLJkCTt37iQmJuaEVzd69OjB+++/zxVXXIFhGPz2t7+t15WQprBkyRLmz5/PJZdcQmJiIkuWLOHgwYP06dMHCHbYvf/++/F6vVx66aWUlZWxfPlyjhw5wuTJkxu1Nl1BOZUeP4IfPRmcn/tr2PLZyduLiEiL9tBDD2G320lPT6ddu3Yn7FPywgsv0KZNG4YPH84VV1zBqFGjOPvss5u42pPzeDx8+eWXXHbZZfTs2ZNHH32UP/zhD4wePRqAn/3sZ/zlL39h+vTp9OvXjx/84AfMmDGjSa6gGGZdH/5uRnw+H16vl/z8fDweT+Mf0DThg4mw8m1we+Bnn0G7Xo1/XBGRFqy0tJQdO3aQlpZGRESE1eVIAznZP9f6/H7rCkpdGAaMeQFSM6DMB+/8BIpzra5KRESkxVJAqSuHG37yNnhT4cgO+Pft0MzuJYqISMt1zz33EBMTc9zpnnvusbq8BqdOsvUR3RZ++i78ZWTwyZ6vX4ALHrK6KhERaQWeeOIJHnro+L85TdLdoYkpoNRX0lkw+rlgn5TPn4Eu50HqMKurEhGRFi4xMZHExESry2gyusVzOgbdCP2uBdMP/75D/VFEREQamALK6TAMuPxFiO8Kvr3wwX0axE1ERKQBKaCcLncsXDMd7C7YOBuWvmF1RSIiIi2GAsqZSBl4dBC3Tx+FrHWWliMiItJSKKCcqaF3Q6/LwF8e7DgbOPX7GUREROTkFFDOVHV/FLcX9q+EJa9bXZGIiDRjXbp04Y9//KPVZTR7CigNITYZLnkiOL/gKTiyy9p6REREwpwCSkMZdDN0Pg8qimH2g3qqR0RE5AwooDQUmw2ueAnsbtg2H9a+Z3VFIiLhxTShvMiaqY7/p/KNN94gJSWFwPdedXLllVdy++23s23bNq688kqSkpKIiYnhnHPO4bPPPjvtU2IYBn/+85+5/PLLiYqKok+fPmRmZrJ161YuvPBCoqOjGT58ONu2bQtts3r1an74wx8SGxuLx+Nh8ODBLF++PLT+66+/5vzzzycyMpJOnTpx//33U1RUdNo1Nha9zbihffk8LHgSohJgwjKITrC6IhGRZumYt96WF8EzKdYU8+v94Io+ZbMjR46QnJzMxx9/zMUXXwxAbm4u7du35+OPP6Zt27YsXryYESNG4Ha7+fvf/87zzz/Ppk2bSE1NBYJ9UCZNmsSkSZNOeTzDMOjQoQMvvPACAwcO5JFHHmHVqlV07dqVX/7yl6SmpnL77bcTFxfHnDlzAOjbty+DBg3iN7/5DXa7nVWrVtGzZ08GDBjAtm3bGDBgAE899RRjxozh4MGDTJw4kQEDBjB9+vTTP3816G3GzdWIByDxLCg+DHN/bXU1IiLSgNq0acPo0aN55513Qsv+/e9/07ZtW374wx8yYMAA7r77bvr27UuPHj148skn6datGx988MFpH/O2227juuuuo2fPnjzyyCPs3LmT8ePHM2rUKPr06cMDDzzAF198EWq/e/duRo4cSe/evenRowfXXnstAwYMAGDatGmMHz+eSZMm0aNHD4YPH87LL7/M3//+d0pLS0+7xsagd/E0NLsTfvwn+MvFsObd4GPIHc62uioRkebPGRW8kmHVseto/Pjx3Hnnnbz66qu43W5mzpzJ9ddfj81mo7CwkMcff5yPPvqIAwcOUFlZSUlJCbt37z7t0vr37x+aT0pKAqBfv361lpWWluLz+fB4PEyePJmf/exn/OMf/2DkyJFce+21dOvWDQje/lmzZg0zZ84MbW+aJoFAgB07dtCnT5/TrrOh6QpKY+g4GPr/JDj/2WPqMCsiUheGEbzNYsVkGHUu84orrsA0TT766CP27NnDV199xfjx4wF46KGHmDVrFs888wxfffUVq1atol+/fpSXl5/2aXE6nTVOkXHCZdX9Yh5//HHWr1/PmDFjWLBgAenp6cyaNQuAwsJC7r77blatWhWaVq9ezZYtW0IhprnQFZTGctFvYP37sONL2Dofeoy0uiIREWkAERERjB07lpkzZ7J161Z69erF2WcHr5R/88033HrrrVx99dVAMBDs3LmzyWvs2bMnPXv25MEHH+SGG25g+vTpXH311Zx99tls2LCB7t27N3lN9aUrKI0lLhXOvSs4/9ljGmFWRKQFGT9+PB999BF/+9vfQldPAHr06MH7778fujLx05/+9JgnfhpTSUkJEydO5IsvvmDXrl188803LFu2LHTr5pFHHmHRokVMnDiRVatWsWXLFv773/8yceLEJquxrhRQGtP5v4AIL2SvgzX/sroaERFpIBdddBHx8fFs2rSJn/70p6HlL7zwAm3atGH48OFcccUVjBo1KnR1pSnY7XYOHz7MzTffTM+ePbnuuusYPXo0v/vd74Bgf5aFCxeyefNmzj//fAYNGsTUqVNJSbHo6amTqPdjxvv27eORRx5hzpw5FBcX0717d6ZPn86QIUOAYGebxx57jDfffJO8vDxGjBjBa6+9Ro8ePUL7yM3N5b777uPDDz/EZrMxbtw4XnrpJWJiYupUQ7N+zPj7vv5j8AqKtxNMXA7OiFNuIiLSGpzscVQJX5Y8ZnzkyBFGjBiB0+lkzpw5bNiwgT/84Q+0adMm1Oa5557j5Zdf5vXXX2fJkiVER0czatSoWo8vjR8/nvXr1zNv3jxmz57Nl19+yV133VWfUsLH0LvB0wHy98DSN6yuRkREJCzU6wrKr371K7755hu++uqr4643TZOUlBR+8Ytf8NBDDwGQn59PUlISM2bM4Prrr+e7774jPT2dZcuWha66fPLJJ1x22WXs3bu3TpeZwuoKCsDKmfDfn0NEHDywCiLbnGoLEZEWr7VfQZk5cyZ33333cdd17tyZ9evXN3FFDcOSKygffPABQ4YM4dprryUxMZFBgwbx5ptvhtbv2LGDrKwsRo48+sSK1+tl6NChZGZmApCZmUlcXFwonACMHDkSm83GkiVLjnvcsrIyfD5frSmsDLg+OHhbaR4s+h+rqxERkWbgxz/+ca3HfWtOH3/8sdXlWa5eAWX79u2h/iRz587l3nvv5f777+ett94CICsrCzg6kEy1pKSk0LqsrCwSExNrrXc4HMTHx4fafN+0adPwer2hqVOnTvUp23o2O/xwSnB+6ZtQmm9tPSIiYrnY2Fi6d+9+3Klz585Wl2e5egWUQCDA2WefzTPPPMOgQYO46667uPPOO3n99dcbqz4ApkyZQn5+fmjas2dPox6vUfQaA+16Q1k+LPur1dWIiDQbYfhKODmJhvrnWa+A0r59e9LT02st69OnT2gI3+TkZACys7NrtcnOzg6tS05OJicnp9b6yspKcnNzQ22+z+124/F4ak1hx2aD8yYH5zNfgfJia+sREbGY3W4HOKNRVqX5KS4O/r7VHO32dNRrJNkRI0awadOmWss2b94cuhSVlpZGcnIy8+fPZ+DAgUCwQ8ySJUu49957AcjIyCAvL48VK1YwePBgABYsWEAgEGDo0KFn9GWavb7j4POnIG83rHwbhrbQJ5dEROrA4XAQFRXFwYMHcTqd2GwamiucmaZJcXExOTk5xMXFhQLo6apXQHnwwQcZPnw4zzzzDNdddx1Lly7ljTfe4I03go/PGobBpEmTeOqpp+jRowdpaWn89re/JSUlhauuugoIXnG59NJLQ7eGKioqmDhxItdff32zHCimQdkdMGISfDQZvnkJBt8KDpfVVYmIWMIwDNq3b8+OHTvYtWuX1eVIA4mLizvhHZH6qPdAbbNnz2bKlCls2bKFtLQ0Jk+ezJ133hlaXz1Q2xtvvEFeXh7nnXcer776Kj179gy1yc3NZeLEibUGanv55Zdb5kBt31dRCi/1h8JsuPJVGDT+1NuIiLRggUBAt3laCKfTedIrJ/X5/a53QGkOGjOg5JdU4I08s/tmp/TNyzDvt5DQAyYsCT7lIyIi0sI12jgoLd36/fn88Pkv+OfS3Y17oCG3BQdtO7wFvvuwcY8lIiIShhRQapi7LovconKmvL+Wfy1vxEeZ3bEw9J7g/Fd/gPC7iCUiItKoFFBqePBHPbl1eBcAHvm/Nbz/7d7GO9jQu8ERCVlrYM/xR9AVERFprRRQajAMg8euSOfGYamYJjz03mr+u2pf4xwsKh76XROc18BtIiIitSigfI9hGDzx477ccG4nAiY8+L+r+GjNgcY52Dl3BD83/AeKDjXOMURERMKQAspx2GwGT1/Vj2sHdyRgwv3vruTzjTmn3rC+UgZBytngL4eV/2j4/YuIiIQpBZQTsNkMnh3Xn6sGpuAPmPx85res2pPX8AeqvoqyfDoEAg2/fxERkTCkgHISdpvB768dwAU921FS4ef2GcvYfrCwYQ9y1liI8ELeLtg2v2H3LSIiEqYUUE7Babfx2viz6d/RS25ROTf/bSk5BaUNdwBXFAysGk122V8abr8iIiJhTAGlDqLdDv526zl0Tohi75ESbv3bMgpKKxruAENuD35unht8kaCIiEgrp4BSR21j3Pz99nNpG+NiwwEfP5/5Lf5AAw2w1rYHpF0AmLBiRsPsU0REJIwpoNRD54Ropt96LhFOG19tOcQfPt3UcDs/52fBz2//DpV6aZaIiLRuCij11K+jl/83rj8Ar36xjU/WZTXMjntdBjHJUHQQNur9PCIi0ropoJyGKwd24PYRaUBwtNmtOQ3wZI/dCWffHJxf/b9nvj8REZEwpoBymqZc1ptz0+IpLKvknrdXUFhWeeY77Tsu+LltAZTknfn+REREwpQCymly2m288tOzSfK42ZpTyMPvrcY807cSJ/aGdr0hUAGb5jRMoSIiImFIAeUMtIt18+r4wTjtBnPWZTFj0c4z32n6VcHPDf85832JiIiEKQWUMzS4cxseHZMOwLQ5G9mY5TuzHaZfGfzctgBK88+wOhERkfCkgNIAbs7ozEW9EymvDHD/P1dSWuE//Z0l9oG2PYMvENz0ScMVKSIiEkYUUBqAYRg8d01/2sa42ZxdyLSPvzuTnek2j4iItHoKKA2kbYyb568Njo/yVuYuPt+Yc/o7O+uq4OfW+VB6hreMREREwpACSgO6sFcit43oAsDD/17NwYKy09tRYjok9AB/WfD9PCIiIq2MAkoDe+TS3vROjuVQYTmP/N+a03v02DCOdpbVbR4REWmFFFAaWITTzss3DMJlt7FgYw4frT1wejuqvs2zZR6UFTRYfSIiIuFAAaUR9EyK5ec/7AbA7z7cgK+0ov47SeoL8d10m0dERFolBZRGcu+F3ejaNpqDBWU8P/c03npsGEevoqyf1aC1iYiINHcKKI3E7bDz1NV9AfjH4l2s2pNX/51UP2689TMoa4AXEoqIiIQJBZRGNLxbW8YO6oBpwq/fX0ulP1C/HST3gzZdoLIUtn/RGCWKiIg0SwoojezXY/rgjXSy4YCv/u/qMQzoeWlwfrNeHigiIq2HAkojaxvj5lejewPwwrzN7M8rqd8OQgHlUwjU8wqMiIhImFJAaQI/GdKJIZ3bUFzu5/f17TDbeQS4YqEoBw6sbJwCRUREmhkFlCZgsxk8dsVZAPxn1T7W76/HW4odLuh+UXBeLw8UEZFWQgGlifTr6OWKASmYJvy/T+p5FSV0m0cBRUREWgcFlCb08CW9cNoNvtx8kG+2Hqr7ht1/BBiQtQby9zVafSIiIs2FAkoTSk2IYvzQzgBMm/MdgUAd39MT0w46nhOc36JRZUVEpOVTQGli913UnRi3g3X7fHy4Zn/dN+w5KvipYe9FRKQVUEBpYgkxbu6+oCsAz3+6ifLKOj463Gt08HP7F1Be3DjFiYiINBMKKBa44/w02sW62ZNbwswlu+q2UWI6eDsFR5Xd8WXjFigiImIxBRQLRLkcTBrZA4A/LdhKUVnlqTcyjBq3eTSqrIiItGwKKBb5yZBOdEmIIreonH8u3V23jXpW3ebZPBfMOnawFRERCUMKKBZx2G3ce2E3AP785XZKK/yn3qjLeeCMhoIDwUeORUREWqh6BZTHH38cwzBqTb179w6tLy0tZcKECSQkJBATE8O4cePIzs6utY/du3czZswYoqKiSExM5OGHH6aysg63OFqgqwd1JMUbwcGCMt5bvufUGzgjoNsPg/MaVVZERFqwel9BOeusszhw4EBo+vrrr0PrHnzwQT788EPee+89Fi5cyP79+xk7dmxovd/vZ8yYMZSXl7No0SLeeustZsyYwdSpUxvm24QZl8PGPVVXUV5fuJ0Kfx2e6OlxSfBz67xGrExERMRa9Q4oDoeD5OTk0NS2bVsA8vPz+etf/8oLL7zARRddxODBg5k+fTqLFi1i8eLFAHz66ads2LCBt99+m4EDBzJ69GiefPJJXnnlFcrLyxv2m4WJ64Z0ol2sm315JcxaWYdRYrtVvZdn3wooyWvU2kRERKxS74CyZcsWUlJS6Nq1K+PHj2f37mAHzxUrVlBRUcHIkSNDbXv37k1qaiqZmZkAZGZm0q9fP5KSkkJtRo0ahc/nY/369Sc8ZllZGT6fr9bUUkQ47dx5fhoAr36+Ff+pRpeN6wQJPcAMwM6vmqBCERGRplevgDJ06FBmzJjBJ598wmuvvcaOHTs4//zzKSgoICsrC5fLRVxcXK1tkpKSyMrKAiArK6tWOKleX73uRKZNm4bX6w1NnTp1qk/Zzd74oZ2Ji3Ky83Axs+syumx1P5RtnzduYSIiIhapV0AZPXo01157Lf3792fUqFF8/PHH5OXl8a9//aux6gNgypQp5Ofnh6Y9e+rQoTSMRLsd3DEieBXllc+3nvodPV2rAsp2BRQREWmZzugx47i4OHr27MnWrVtJTk6mvLycvLy8Wm2ys7NJTk4GIDk5+Zineqr/rm5zPG63G4/HU2tqaW4e3oVYt4PN2YV8uiH75I27nAeGHXK3w5GdTVKfiIhIUzqjgFJYWMi2bdto3749gwcPxul0Mn/+/ND6TZs2sXv3bjIyMgDIyMhg7dq15OTkhNrMmzcPj8dDenr6mZQS9ryRTm7KCL7p+G/f7Dh54wgPdDo3OK/bPCIi0gLVK6A89NBDLFy4kJ07d7Jo0SKuvvpq7HY7N9xwA16vlzvuuIPJkyfz+eefs2LFCm677TYyMjIYNmwYAJdccgnp6encdNNNrF69mrlz5/Loo48yYcIE3G53o3zBcHJTRmfsNoOlO3LZlFVw8sa6zSMiIi1YvQLK3r17ueGGG+jVqxfXXXcdCQkJLF68mHbt2gHw4osvcvnllzNu3DguuOACkpOTef/990Pb2+12Zs+ejd1uJyMjgxtvvJGbb76ZJ554omG/VZhq743kR32CnYb/sXjnyRtXd5TdvhACdRiFVkREJIwYphl+L3Xx+Xx4vV7y8/NbXH+URVsP8dO/LCHaZWfxry8mNsJ5/Ib+SniuK5Tlw50LoMPgpi1URESknurz+6138TQzGd0S6NYumqJy/8kHbrM7IO384Py2BU1TnIiISBNRQGlmDMPgpmHBzrJ/z9zFSS9whcZD+aLxCxMREWlCCijN0NjBHYly2dmaU8ji7bknbljdUXbPEigrbJriREREmoACSjPkiXBy1aAOwCk6y8Z3hbhUCFTArkVNU5yIiEgTUEBppqpv88xdn022r/T4jQxDjxuLiEiLpIDSTPVp7+GcLm3wB0zeWbL7xA1D/VDUUVZERFoOBZRm7Maqqyj/XLqbCn/g+I3SfgAYcHAj+OrwokEREZEwoIDSjI3u256EaBc5BWV8veXQ8RtFxUPKwOD8jq+arDYREZHGpIDSjLkcNi7v3x6AD1af5OpIl6rxUHYqoIiISMuggNLM/Xhg8GmeueuzKCk/wZD2CigiItLCKKA0c2enxtGxTSTF5X4++y77+I1Sh4FhhyM7IW9Pk9YnIiLSGBRQmjnDMPjxgBTgJLd5IjxH+6Hs/LppChMREWlECihh4Mqq2zxfbMohv7ji+I1Ct3kUUEREJPwpoISBXsmx9E6OpcJvMmfdgeM3CgWUL5uuMBERkUaigBImrjjVbZ7qfih5u+HIriasTEREpOEpoISJ6n4omdsPH3/oe3cMdDg7OK/bPCIiEuYUUMJEp/goBndug2nChye6iqJ+KCIi0kIooISRKwcGr6KcOKCcF/zc+RWYZhNVJSIi0vAUUMLIZf3aY7cZrN6bz45DRcc2SB0GNgfk74E89UMREZHwpYASRtrGuBnRvS0AH6w6zlUUVzR0GByc13t5REQkjCmghJnqzrKz16gfioiItFwKKGHmR+lJOO0GW3IK2ZxdcGwD9UMREZEWQAElzHgjnVzQox0AH605zqBtnYaCzQm+fXBkRxNXJyIi0jAUUMLQZf3aA/Dx2uMEFFcUdDwnOK9+KCIiEqYUUMLQyPrc5hEREQlDCihh6JS3eaoDyq5F6ociIiJhSQElTJ30Nk/HIcHxUHz7gmOiiIiIhBkFlDB10ts8rmhoPyA4vyuz6YsTERE5QwooYeqUt3lSM4KfuxVQREQk/CighLHq2zwfHe82T+fhwU8FFBERCUMKKGGs+jbP1uPd5uk0LPh5cCMU5zZ9cSIiImdAASWMnfQ2T3QCtO0VnN+9uIkrExEROTMKKGHupLd5Uquuouxe1IQViYiInDkFlDB30ts8oX4ouoIiIiLhRQElzHkjnYzo3haAeRuya6+svoKyfyWUFzdxZSIiIqdPAaUFuLhPEgDzv/teQInrDLEpEKiEfSssqExEROT0KKC0ACP7JAKwck8ehwrLjq4wjBr9UPS4sYiIhA8FlBagvTeSs1I8mCYs2JhTe2V1P5Rd6igrIiLhQwGlhTjhbZ7qEWX3LgN/ZRNXJSIicnoUUFqIH1UFlK+2HKK0wn90RWIfcHuhvBCy11pUnYiISP0ooLQQfTt4SPK4KS73s3j74aMrbHZIHRqc1+PGIiISJs4ooDz77LMYhsGkSZNCy0pLS5kwYQIJCQnExMQwbtw4srNr33bYvXs3Y8aMISoqisTERB5++GEqK3X74UwYhsFFvatv83yvH0p1R1n1QxERkTBx2gFl2bJl/PnPf6Z///61lj/44IN8+OGHvPfeeyxcuJD9+/czduzY0Hq/38+YMWMoLy9n0aJFvPXWW8yYMYOpU6ee/rcQ4OjTPPO/y8Y0zaMrUmsM2FZzuYiISDN1WgGlsLCQ8ePH8+abb9KmTZvQ8vz8fP7617/ywgsvcNFFFzF48GCmT5/OokWLWLw4eHvh008/ZcOGDbz99tsMHDiQ0aNH8+STT/LKK69QXl7eMN+qlRrRvS0RThv780vZcMB3dEXKILC7oCgHcrdbV6CIiEgdnVZAmTBhAmPGjGHkyJG1lq9YsYKKiopay3v37k1qaiqZmcFxODIzM+nXrx9JSUmhNqNGjcLn87F+/frjHq+srAyfz1drkmNFOO2c1z348sBat3mcEdBhcHBe46GIiEgYqHdAeffdd/n222+ZNm3aMeuysrJwuVzExcXVWp6UlERWVlaoTc1wUr2+et3xTJs2Da/XG5o6depU37JbjZq3eWrpeE7wc+/yJq5IRESk/uoVUPbs2cMDDzzAzJkziYiIaKyajjFlyhTy8/ND0549e5rs2OHmoqqAsnpvPtm+0qMrOg4Jfu5TQBERkeavXgFlxYoV5OTkcPbZZ+NwOHA4HCxcuJCXX34Zh8NBUlIS5eXl5OXl1douOzub5ORkAJKTk495qqf67+o23+d2u/F4PLUmOb7E2AgGdIoDvjeqbPUVlOz1UF7U9IWJiIjUQ70CysUXX8zatWtZtWpVaBoyZAjjx48PzTudTubPnx/aZtOmTezevZuMjOCIphkZGaxdu5acnKM/nvPmzcPj8ZCent5AX6t1G9n7OLd5PCnBFweaAdi/yprCRERE6shRn8axsbH07du31rLo6GgSEhJCy++44w4mT55MfHw8Ho+H++67j4yMDIYNC47Fcckll5Cens5NN93Ec889R1ZWFo8++igTJkzA7XY30Ndq3S7qk8gf5m3mm62HKav043bYgys6DoHvPggOe99lhLVFioiInESDjyT74osvcvnllzNu3DguuOACkpOTef/990Pr7XY7s2fPxm63k5GRwY033sjNN9/ME0880dCltFrp7T20i3VTUuFn2Y4jR1dU90PZu8yawkREROrIMM3wG7nL5/Ph9XrJz89Xf5QTeOi91fx7xV7uPD+N34ypunW2axFMHw0xyfCLjWAY1hYpIiKtSn1+v/Uunhbqwl7B8VAWbj54dGH7gWDYoTALfPusKUxERKQOFFBaqPO6t8VmwObsQvbnlQQXuqIg6azgvMZDERGRZkwBpYWKi3IxKDX4GoJaV1FCA7apH4qIiDRfCigt2A96Bm/zfLGp5ngo1R1ldQVFRESaLwWUFqy6H8o3Ww9T4Q8EF1ZfQTmwCvwV1hQmIiJyCgooLVjfFC/x0S4KyypZsavqceP4bhARB5WlkL3O0vpERERORAGlBbPZDC7o0Rao0Q/FZjv6ZmPd5hERkWZKAaWFu7BXcNj7hZuO11FWAUVERJonBZQW7vwebTEM2HDAR0712401oqyIiDRzCigtXEKMm/4dvECN2zzVt3hyt0FxrkWViYiInJgCSisQety4OqBExUNC9+D8vhUWVSUiInJiCiitwA+q+qF8veUQldWPG3fQeCgiItJ8KaC0AgM7xeGNdJJfUsHqvXnBheqHIiIizZgCSitgtxmcF3rc+FBwYXVA2bccwu+F1iIi0sIpoLQSP+gR7IfyZXU/lMSzwO6C0nw4ssPCykRERI6lgNJKnN8zeAVlzd488orLweGCpL7BlftXWliZiIjIsRRQWon23kh6JsUQMIPv5gGgw9nBTwUUERFpZhRQWpHzv3+bJ2VQ8HP/KmsKEhEROQEFlFbkgqrxUL7cchDTNGsHlEDAusJERES+RwGlFRmaFo/bYeNAfinbDhZC217giITyAji81eryREREQhRQWpEIp51z0+KBqseN7Q5o3z+4Uv1QRESkGVFAaWUuOKYfijrKiohI86OA0spU90NZsuMwpRX+Gv1QvrWwKhERkdoUUFqZnkkxJHsiKK0IsHznkaMB5cAa8FdaW5yIiEgVBZRWxjAMzq8a9v7LLQeDbzV2xUJlCRzaZHF1IiIiQQoordD5PWv0Q7HZIGVgcIX6oYiISDOhgNIKnd+9LYYBG7MKyPaVKqCIiEizo4DSCrWJdtG/gxeAr7YcOtoPZZ86yoqISPOggNJKXVDzNk91QMleB5XlFlYlIiISpIDSSlW/l+frrYcIeLtARBz4yyFng6V1iYiIgAJKqzWwUxyRTju5ReVsPlhYYzwU9UMRERHrKaC0Ui6HjSFd2gCQue2wAoqIiDQrCiit2PBuwfFQagcUdZQVERHrKaC0YhndEgBYvP0w/vYDgwtzvoOKEuuKEhERQQGlVeub4iHG7cBXWsl3RR6IbgeBSsheb3VpIiLSyimgtGIOu41z0+IByNyeq/FQRESk2VBAaeUyugZv82RuPwzVt3kOrLauIBERERRQWr3qfihLd+TiT+ofXHhglXUFiYiIoIDS6vVp78Eb6aSwrJLvjK7BhTnfQUWptYWJiEirpoDSytltBkOr+qF8me2CqAQw/eooKyIilqpXQHnttdfo378/Ho8Hj8dDRkYGc+bMCa0vLS1lwoQJJCQkEBMTw7hx48jOzq61j927dzNmzBiioqJITEzk4YcfprKysmG+jZyW6ts8mdtza/RDWWVZPSIiIvUKKB07duTZZ59lxYoVLF++nIsuuogrr7yS9euD/2/7wQcf5MMPP+S9995j4cKF7N+/n7Fjx4a29/v9jBkzhvLychYtWsRbb73FjBkzmDp1asN+K6mX6oCyfOcR9UMREZFmwTBN0zyTHcTHx/P73/+ea665hnbt2vHOO+9wzTXXALBx40b69OlDZmYmw4YNY86cOVx++eXs37+fpKQkAF5//XUeeeQRDh48iMvlqtMxfT4fXq+X/Px8PB7PmZQvQCBgMuTpz8gtKuezS/Po/sXPof0AuPtLq0sTEZEWpD6/36fdB8Xv9/Puu+9SVFRERkYGK1asoKKigpEjR4ba9O7dm9TUVDIzMwHIzMykX79+oXACMGrUKHw+X+gqjDQ9m81gWNdgP5SvizoEF2ZvgMoyC6sSEZHWrN4BZe3atcTExOB2u7nnnnuYNWsW6enpZGVl4XK5iIuLq9U+KSmJrKwsALKysmqFk+r11etOpKysDJ/PV2uShpVR9V6euXvdEBEHgYrg0zwiIiIWqHdA6dWrF6tWrWLJkiXce++93HLLLWzYsKExaguZNm0aXq83NHXq1KlRj9caVQ/YtmJPHv7kAcGF6ociIiIWqXdAcblcdO/encGDBzNt2jQGDBjASy+9RHJyMuXl5eTl5dVqn52dTXJyMgDJycnHPNVT/Xd1m+OZMmUK+fn5oWnPnj31LVtOoVu7aNrFuimvDJAV3Su4UCPKioiIRc54HJRAIEBZWRmDBw/G6XQyf/780LpNmzaxe/duMjIyAMjIyGDt2rXk5OSE2sybNw+Px0N6evoJj+F2u0OPNldP0rAMwwhdRVlZ0Tm4cP8q6woSEZFWzVGfxlOmTGH06NGkpqZSUFDAO++8wxdffMHcuXPxer3ccccdTJ48mfj4eDweD/fddx8ZGRkMGzYMgEsuuYT09HRuuukmnnvuObKysnj00UeZMGECbre7Ub6g1F1GtwQ+WL2fubnJXA7Bwdr8FWB3Wl2aiIi0MvUKKDk5Odx8880cOHAAr9dL//79mTt3Lj/60Y8AePHFF7HZbIwbN46ysjJGjRrFq6++Gtrebrcze/Zs7r33XjIyMoiOjuaWW27hiSeeaNhvJael+grK3AMRmDGxGGUFcHAjJPezuDIREWltzngcFCtoHJTGYZomw59dwIH8UlZ1fom47CVw5Ssw6EarSxMRkRagScZBkZanZj+ULbZuwYXqhyIiIhZQQJFahlUNe/9V9YBtepJHREQsoIAitVRfQZlzqGpAvay14NfLHEVEpGkpoEgtneKj6Ngmkq2BZCod0VBZAoe3WF2WiIi0MgoocoyMrgmY2DgQ0T24QP1QRESkiSmgyDEyqvqhrKzsElygfigiItLEFFDkGNUBZWFB++ACvZNHRESamAKKHKO9N5IuCVGsCXQNLjiwBgJ+a4sSEZFWRQFFjiujWwLbzBTKbRFQUQSHt1pdkoiItCIKKHJcw7omEMDGZlvVVZT9K60tSEREWhUFFDmu6vFQlpRWvdl437cWViMiIq2NAoocV6Ingm7tolkTSAsu0BUUERFpQgoockIZ3RJYY1a9kydrDfgrrC1IRERaDQUUOaGMrm3ZaSZRRBRUlsLBjVaXJCIirYQCipzQsK7xmNhY7e8SXKDbPCIi0kQUUOSEEmLcdE+MOXqbRx1lRUSkiSigyEmd0yVeHWVFRKTJKaDISZ2b1uboFZTs9VBZZm1BIiLSKiigyEmdm5bAXrMtuWYMBCoge53VJYmISCuggCIn1SEukg5xUawNaERZERFpOgoockrndGnDarMqoOxTQBERkcangCKndG5agq6giIhIk1JAkVM6N60Na6oCinnwOygvtrgiERFp6RRQ5JS6tYuhIjqZbDMOwwwEh70XERFpRAoockqGYXBOl6NXUXSbR0REGpsCitRJcMA2BRQREWkaCihSJ0PTElhb9SSPqSHvRUSkkSmgSJ30aR/LVkcPAIzDW6DUZ3FFIiLSkimgSJ047DbSOndmr9k2uODAamsLEhGRFk0BRepsaFrNfii6zSMiIo1HAUXqrGZHWXPfCourERGRlkwBRepsQKc41hnBfij+3cssrkZERFoyBRSpswinHSNlEH7TwFG4H3wHrC5JRERaKAUUqZd+XTuw2ewY/GPfcmuLERGRFksBReplWNcEVgW6A2DuVUAREZHGoYAi9TKkSxvWEgwoZTuXWlyNiIi0VAooUi9RLgcliYMAsGetgoDf2oJERKRFUkCReuvYcxCFZgROfzEc3Gh1OSIi0gIpoEi9De2WeHQ8FPVDERGRRqCAIvU2uHMb1lT1QyncttjiakREpCVSQJF6i3TZKUgYCIB/jwZsExGRhlevgDJt2jTOOeccYmNjSUxM5KqrrmLTpk212pSWljJhwgQSEhKIiYlh3LhxZGdn12qze/duxowZQ1RUFImJiTz88MNUVlae+beRJuPpPiz4WbAVygosrkZERFqaegWUhQsXMmHCBBYvXsy8efOoqKjgkksuoaioKNTmwQcf5MMPP+S9995j4cKF7N+/n7Fjx4bW+/1+xowZQ3l5OYsWLeKtt95ixowZTJ06teG+lTS6fn16sc9MwIaJuU8vDhQRkYZlmKZpnu7GBw8eJDExkYULF3LBBReQn59Pu3bteOedd7jmmmsA2LhxI3369CEzM5Nhw4YxZ84cLr/8cvbv309SUhIAr7/+Oo888ggHDx7E5XKd8rg+nw+v10t+fj4ej+d0y5czUFrh5/MnRzPatoTDw6aQcOmvrC5JRESaufr8fp9RH5T8/HwA4uPjAVixYgUVFRWMHDky1KZ3796kpqaSmZkJQGZmJv369QuFE4BRo0bh8/lYv379cY9TVlaGz+erNYm1Ipx2Dnn7A1C8Y4nF1YiISEtz2gElEAgwadIkRowYQd++fQHIysrC5XIRFxdXq21SUhJZWVmhNjXDSfX66nXHM23aNLxeb2jq1KnT6ZYtDcjV5VwAPIdWw+lfiBMRETnGaQeUCRMmsG7dOt59992GrOe4pkyZQn5+fmjas2dPox9TTi2t33AqTRte/2HM/L1WlyMiIi3IaQWUiRMnMnv2bD7//HM6duwYWp6cnEx5eTl5eXm12mdnZ5OcnBxq8/2neqr/rm7zfW63G4/HU2sS6/VPa88mUgHI+u4bi6sREZGWpF4BxTRNJk6cyKxZs1iwYAFpaWm11g8ePBin08n8+fNDyzZt2sTu3bvJyMgAICMjg7Vr15KTkxNqM2/ePDweD+np6WfyXaSJRTjt7I8+C4DcTYssrkZERFoSR30aT5gwgXfeeYf//ve/xMbGhvqMeL1eIiMj8Xq93HHHHUyePJn4+Hg8Hg/33XcfGRkZDBsWHDfjkksuIT09nZtuuonnnnuOrKwsHn30USZMmIDb7W74byiNq+MQ2PwR7mw9aiwiIg2nXgHltddeA+DCCy+stXz69OnceuutALz44ovYbDbGjRtHWVkZo0aN4tVXXw21tdvtzJ49m3vvvZeMjAyio6O55ZZbeOKJJ87sm4glkvqcB5uhY8kmzMpyDMepHxMXERE5lTMaB8UqGgel+SirqKD0qc54jSJ2jv2ILv3Ps7okERFppppsHBQRt9PJ9sjgY+bZ6xZYXI2IiLQUCihyxkpThgLg3KsB20REpGEooMgZa5v+AwA6F62hstJvcTUiItISKKDIGes64HxKcZJg+Njy3UqryxERkRZAAUXOmN3pZldEcAybrLXqhyIiImdOAUUaRGn74Ht5nHsXW1yJiIi0BAoo0iDanvVDALoUraa0Qv1QRETkzCigSINI6XcBldjoaBxi3XfrrS5HRETCnAKKNAjDHcv+iB4AZK353OJqREQk3CmgSIMpaR8cD8WxT/1QRETkzCigSINJqBoPpWvRGgpKKyyuRkREwpkCijSYtukXAtDTtpdvN26zthgREQlrCijScKLbkuPuDMCBtV9YWoqIiIQ3BRRpUCVV46E49mZaXImIiIQzBRRpUPFVt3m6l6zlUGGZtcWIiEjYUkCRBhXb8wIAzjJ2snTTHourERGRcKWAIg0rLpV8ZxJOw8++dV9aXY2IiIQpBRRpcKUpwX4oxq5MTNO0uBoREQlHCijS4OLTg+/l6Ve5hvX7fRZXIyIi4UgBRRqcs0cwoJxtbGHh2u0WVyMiIuFIAUUaXnxXCqM64jT8HF6v9/KIiEj9KaBIo7D3uBiATkcWk+0rtbgaEREJNwoo0igie40E4ALbGhZszLG4GhERCTcKKNI40i4ggI1utgOsXLvG6mpERCTMKKBI44iMozRpEADOXQsprfBbXJCIiIQTBRRpNJG9g7d5Msw1fLP1kMXViIhIOFFAkUZjdLsIgBG2dSz47oDF1YiISDhRQJHG02Ewlc4Y2hiF7NuwRKPKiohInSmgSOOxOzHSfgBAeslyjSorIiJ1poAijcrePTiq7Pm2dXz2XbbF1YiISLhQQJHGVdUPZbBtE99s2GVxMSIiEi4UUKRxxXfF703FZfiJyVpCVr5GlRURkVNTQJHGZRjYuweHvT/ftpZP1ulpHhEROTUFFGl83ar7oaxl9hoFFBEROTUFFGl8aRdgGjZ62Paxb9dW9ueVWF2RiIg0cwoo0vgi22B0PBeAUfZlfKSrKCIicgoKKNI0zroagMvti5m9Zr/FxYiISHOngCJNI/1KTAyG2DaTs3c7uw4XWV2RiIg0Ywoo0jQ87TE6DwfgMvtidZYVEZGTUkCRphO6zbNEAUVERE5KAUWaTp8fYxo2Btm2UpC1ja05hVZXJCIizVS9A8qXX37JFVdcQUpKCoZh8J///KfWetM0mTp1Ku3btycyMpKRI0eyZcuWWm1yc3MZP348Ho+HuLg47rjjDgoL9WPV4sUmYXQeAcBlNnWWFRGRE6t3QCkqKmLAgAG88sorx13/3HPP8fLLL/P666+zZMkSoqOjGTVqFKWlR4c4Hz9+POvXr2fevHnMnj2bL7/8krvuuuv0v4WEj75jgeDTPB+u3o9pmhYXJCIizZFhnsEvhGEYzJo1i6uuugoIXj1JSUnhF7/4BQ899BAA+fn5JCUlMWPGDK6//nq+++470tPTWbZsGUOGDAHgk08+4bLLLmPv3r2kpKSc8rg+nw+v10t+fj4ej+d0yxcrFB3CfL4nhunngrIX+fP919Cnvf4Zioi0BvX5/W7QPig7duwgKyuLkSNHhpZ5vV6GDh1KZmYmAJmZmcTFxYXCCcDIkSOx2WwsWbLkuPstKyvD5/PVmiRMRbfFSLsAgMt1m0dERE6gQQNKVlYWAElJSbWWJyUlhdZlZWWRmJhYa73D4SA+Pj7U5vumTZuG1+sNTZ06dWrIsqWpVd3mGWNfzPvf7qPSH7C4IBERaW7C4imeKVOmkJ+fH5r27NljdUlyJnpfjmlzcJZtFxG+HXy6IdvqikREpJlp0ICSnJwMQHZ27R+c7Ozs0Lrk5GRycnJqra+srCQ3NzfU5vvcbjcej6fWJGEsKh6ja/ANx2Nsi5mxaKe19YiISLPToAElLS2N5ORk5s+fH1rm8/lYsmQJGRkZAGRkZJCXl8eKFStCbRYsWEAgEGDo0KENWY40Z1WDtl3r+JLlOw6xYb/6FYmIyFH1DiiFhYWsWrWKVatWAcGOsatWrWL37t0YhsGkSZN46qmn+OCDD1i7di0333wzKSkpoSd9+vTpw6WXXsqdd97J0qVL+eabb5g4cSLXX399nZ7gkRbirKsgsg2djWxG2Zbxlq6iiIhIDfUOKMuXL2fQoEEMGjQIgMmTJzNo0CCmTp0KwC9/+Uvuu+8+7rrrLs455xwKCwv55JNPiIiICO1j5syZ9O7dm4svvpjLLruM8847jzfeeKOBvpKEBVc0nBsc++Zux4f8Z9VejhSVW1yUiIg0F2c0DopVNA5KC1F0CPPFszAqS7m+/FF+cMlY7r2wm9VViYhII7FsHBSReoluizHoRgDutn/I24t36ZFjEREBFFDEahkTMA0bP7SvJjZ/E599l3PqbUREpMVTQBFrxXfFSL8SgLscs5mxaIfFBYmISHOggCLWG34/AD+2LWLP9k2s25dvcUEiImI1BRSxXoezIe0CHEaA2x2f8PgH6wkEwq7vtoiINCAFFGkeRjwAwA32BWzZtYf3Vuh1BiIirZkCijQP3S6GpH5EGWX81vk20+Zs5HBhmdVViYiIRRRQpHkwDLjs95gYXGP/kqGli5g2Z6PVVYmIiEUUUKT56JyBUXWr5xnnX1i4Yh2Ltx+2uCgREbGCAoo0Lz/8NST1I8Eo4Fnnmzw6ay3llRq8TUSktVFAkebF4Yaxb2DaXVxsX8mQ3A9586vtVlclIiJNTAFFmp+kdIyLgy+f/K3jH/zfvK9YtO2QxUWJiEhTUkCR5mnYBMzOI4g2ynjV8Qee+sccth8stLoqERFpIgoo0jzZbBhXv44Z1Zbetj380/wlf/3rq+QXV1hdmYiINAEFFGm+4lIx7l5IRfvBeI1ini59mi9enUBFRbnVlYmISCNTQJHmzdsR5x2fkNv3dgCuLPxf9vzxEsx9Ky0uTEREGpMCijR/Dhfx17zI2mEvUmhG0LVoJcabF1Lx2g/g239AeZHVFYqISAMzTNMMu7ey+Xw+vF4v+fn5eDweq8uRJvT+p5/j+Oo5RtmW4jYqATDdHoxel0GHwZAyCJL7gjPS4kpFROT76vP7rYAiYWfdvnym/d/XpGfPZrx9Pl1s2bUbGHZI7ANte0JC9xpTN4iMs6RmERFRQJFWwB8wmblkF89/8h1nVaxlmO07hkbspr+xnaiK3BNvGJMUDC7tekHbXpCUDonpEBXfdMWLiLRSCijSauT4Svl/n2ziwzX7q4bEN0kxcrmmQy5DPUfoasuibdkenEe2QWHWiXfk6QBJZwWnxKrPtj3A7myy7yIi0tIpoEirk19SwcdrD/D+t3tZtvPIMes7xEUyMNHG2VEH6eU4QGpgD21LdhB5ZBNG3u7j79TmDF5pSUwP3jJKTA9ecfF2Cr59WURE6kUBRVq1XYeLmLs+i7X7fKzbl8+OQyd+ysduM+gVF2BYTA6DXHvpZu4iuXQ7Xt9m7BUnGLnWFQPxXYN9Wqr7t7TtEbxl5I5ppG8lIhL+FFBEavCVVrB+n49tBwvZdbiInYeLQ58nflOySQcO0d+5jyFRB0i37yMtsIt2Zbuwm5UnPpg3FRJ7B/u5tO0B8VUhJjZZV11EpNVTQBGpg0DAJLuglF1VgWXX4WL2HClh35Fi9h4pIaeg7JhtHFTS2cgmzciii5FFN1sWvZzZpLGPNoFjby2FOKMhPg3iOkObLtCm+rNLcJkzorG+pohIs6GAItIAyir9HMgrZV9eCXurQktwKmZ/XilZvlL8gaP/84mjgB7GPnra9tLd2EcXI4s0I4uOxkEcxomu1AT5Y9pji0/DaNM52GHX2wE8HcGTAnGpEKF/z0Uk/CmgiDQBf8Akp6CU/XklZOWXcSC/hGxfKVm+Mg7klXAgPxhibIEKOhk5pBo5dDJy6GQcJLXG37FGySmPVeaKozymE2ZcZxzxnYlI6ITNmxIMM7Htg49P2x1N8K1FRE5ffX6/9V80kdNktxm090bS3nviUWtrhphsXxk5vlJyCsrYXFBGTkEZOfklVBQcIrZkL6lGNh2Mw7SvmlKqPuONQtzlebhz8yB3LWw/9jgBDIqd8RS721ERlYQZnYTNk4wzLpnINh2Iiu+AzZMUDDIOd+OdFBGRBqKAItKI6hJiACr8AQ4WlHGwoIxDhcHPjVUhpqggD4dvFxGFe/GW7qNNRTZJRi7JxhGSjCMkcQSn4Sem4jAxFYehcONJj1VgxFLgiKfYlUCZuy2VUW0xo9thi0nE4UnC5U0iqk0yMfHJxMbEYqhzr4hYQAFFpBlw2m2kxEWSEneiIHNeaK7SH+BwUTkHC8rYUlDGN75iSvJyqMzfj+nLwlGUhbs0h6jyw3gqDtHGzKOdkUc78nAblcSaBcRWFEDFLigCTjLwbqEZQb4Ri8/mpdjuocQRR6kzjoqINgQi4jEjEzBi2mKPbofdk4g7NoGYSBcxbgexEQ5iI5xEOe3YbAo5IlI/CigiYcZht5HkiSDJU/PJny4nbF9eGeBIcTk7isooOHKIkiP7KM87QMCXhVF0EHvJIdylh4msOExMZR7eQB5tzHxcRiUxRikxlNIhcBACQAVQAviOfyy/aZBLLLmmhz1Ek2/GkE80JfYYyhweyh2xVDo9+N0ezAgPuL0YEV4cUXHYozxEu11EOu1EuuxEuexEOu1Eux1Eu+1EuRxEuxxEue047XoRu0hLp4Ai0sK5HDUCTXsv0O3UG5kmpYV5FOUeoDg/hzLfISoKDuIvPATFR7CVHsZRegRX+REiK44QXZlHTKAAu2HSDh/tjOMkmMqqqRQoOP5hC8xICoik0IykgCgKzChyiCbPjCGfGPLMaAqIotQWRaUjmoAzBr8zBtMVDa4YDHcMLpebyKpwE+kMBp0I19H5SJeDqKoQFOG04XbYiXAG5yOqtolw2rHrqo+IpRRQRORYhkFEbBsiYtuQQHrdtvFXQPFhKDoIRYcwS/KoLMqlrOAwFUW5BEryMUvyoNSHUebDUZ6Ps6IQV2UBDrMcgFijhFhKoC7ZwATKq6YagwWXmU4KiKTAjMRHdFXoiaKYCEpMFz7clOCi1HRTTHAqMd0UEUExERSaERQRSbktigpnNKYjCndVaHE7bKFPl8OG21EdcIKf1eurw07NwFPd3lVr26PbuRw2nPbqyVDfH2n1FFBEpGHYncERc2OTgWDGcFZNp1RRCmU+KCuA0vzgZ/V8aR4U50LJEfzFuQRKfJilPsyyAigvxFZeiL2yCFugAgC3UYGbCtoe7yrOaQhUGBRWBENLkRlBIRGUmBGhcFNsRlCOg3Kcoc8i08XBGuGnGDclNeaLcVNmuijDSVnVdia1b1s57QZOe1XQcdhwVwUjt8OGoyrE1Aw0TrsNV9XfobDjMI5Z5nLYcFW1d9iD8w6bDUeN/TnsBk5bcPua+3XWaFvdRv2LpLEooIiI9ZwRwSkm8aTN7FXTcVWWQ3lhcCorgFJfVdip+qwohoqSGlMRlBcHl5cXQXkRZnlRKPgY5YUYZgCbYeKhBE9dr+ycpnLTThmuUGApM53BvyuclFa4KCs+GmbKcVJuOqnAXhWMnJThoNx0hkJPIU7KzOp1TipwUIGDctMRmq/AHvw0HVRipwI7/qrPyqrpVF/aZhAKLXabgcNmBANU1Wd1kAmGmurlweATbG/DYQtuWz3ZDAO7Dey2o+Grej/277V11Po8ui+HPbif7+/76Ha22ttXtbcZVH0eXe6oqqN6O5sN7FXrdaWr8SigiEjL4HCBIx6i4k97FwY1fo5NMxhkqgNP9VRRXBWEqoJNRVEwHPnLjn5WlFStK64KQUVVy4qPBqPK2gP0uQw/LqpucUGjhqH6qMBOpRkMK+VVwaYSeyjoVAcZPzYqA3Yq/XYqK+yhdUfDjoMK82jwqawKQ34M/NgIYMM0DUwMTMDEwI+d8qoQVVwVwEKhqkbI8pt2Kqv2UYmdADb8VVMAI3SsStNWdczg8upjgUEAI7SdWVVTzVpO9A+kOqBVhx1bjZBlMzg6XxVqgvNGcN5Wu41hBI9SPW+3VYWz6rBX1Y5QiKoZpqgR7o7uz169X8MIfoOqbew16rDbqKrl6D4Nw6BPcizDu7dtmn/RjkMBRUTkeAwDXFHB6RRXdk6LaQb77VSWQmVZ8NNfXnu+ouToZ2VZMNT4K4Lz/vKj7b+/XWXZ0b+rPwMVwW395cEgFaia91cvP/bdUwBO/DgN/3HOT8OfkuYuYFYHGQOzKuQEqoKUHxuV2PAH7FQG7PjN6pBkD4UhgOqh2wOhAHU0TFW3CZi2GsGIWvPfZ1bt2x+qp3qfR2s8GshsBMyj4evouqO3F43QHmF31xEM7/5gI53NU1NAERGxgmFUXfVxWV3JUQE/BCqDgSVQAf7Ko8EmUFk70IQCTjkEAsG/Q9v6q5Z/fx8VR7c3/VXHqzqm6Q+GNgDM4Lzpr7oqVX40RPlrBC1/We19BCqD89X7rnkMs0Yb8+TvxjoRm2FiC0WM44S2mlpAgNtKnKXHV0AREZEgmz04tfTXIQQCHA1B1fOBo6HIDARDDVSFpur1gdp/B/xHl1eHo0DF90JTRe3gVb1P0x+so2aYqnWsGm1DjvPqvOrvULPumrWaNf8O1DgWx35XCAbnqnTVvcPZDXO+T5MCioiItC42DfQXDvRPSURERJodSwPKK6+8QpcuXYiIiGDo0KEsXbrUynJERESkmbAsoPzv//4vkydP5rHHHuPbb79lwIABjBo1ipycHKtKEhERkWbCsoDywgsvcOedd3LbbbeRnp7O66+/TlRUFH/729+sKklERESaCUsCSnl5OStWrGDkyJFHC7HZGDlyJJmZmce0Lysrw+fz1ZpERESk5bIkoBw6dAi/309SUlKt5UlJSWRlZR3Tftq0aXi93tDUqVOnpipVRERELBAWT/FMmTKF/Pz80LRnzx6rSxIREZFGZMk4KG3btsVut5OdnV1reXZ2NsnJyce0d7vduN0tfOAgERERCbHkCorL5WLw4MHMnz8/tCwQCDB//nwyMjKsKElERESaEctGkp08eTK33HILQ4YM4dxzz+WPf/wjRUVF3HbbbVaVJCIiIs2EZQHlJz/5CQcPHmTq1KlkZWUxcOBAPvnkk2M6zoqIiEjrY5imeZy3DzVvPp8Pr9dLfn4+Ho/H6nJERESkDurz+x0WT/GIiIhI6xKWbzOuvuijAdtERETCR/Xvdl1u3oRlQCkoKADQgG0iIiJhqKCgAK/Xe9I2YdkHJRAIsH//fmJjYzEMo0H37fP56NSpE3v27FH/lkamc910dK6bjs5109G5bjoNda5N06SgoICUlBRstpP3MgnLKyg2m42OHTs26jE8Ho/+hW8iOtdNR+e66ehcNx2d66bTEOf6VFdOqqmTrIiIiDQ7CigiIiLS7CigfI/b7eaxxx7Tu3+agM5109G5bjo6101H57rpWHGuw7KTrIiIiLRsuoIiIiIizY4CioiIiDQ7CigiIiLS7CigiIiISLOjgFLDK6+8QpcuXYiIiGDo0KEsXbrU6pLC3rRp0zjnnHOIjY0lMTGRq666ik2bNtVqU1payoQJE0hISCAmJoZx48aRnZ1tUcUtx7PPPothGEyaNCm0TOe64ezbt48bb7yRhIQEIiMj6devH8uXLw+tN02TqVOn0r59eyIjIxk5ciRbtmyxsOLw5Pf7+e1vf0taWhqRkZF069aNJ598sta7XHSuT8+XX37JFVdcQUpKCoZh8J///KfW+rqc19zcXMaPH4/H4yEuLo477riDwsLChinQFNM0TfPdd981XS6X+be//c1cv369eeedd5pxcXFmdna21aWFtVGjRpnTp083161bZ65atcq87LLLzNTUVLOwsDDU5p577jE7depkzp8/31y+fLk5bNgwc/jw4RZWHf6WLl1qdunSxezfv7/5wAMPhJbrXDeM3Nxcs3Pnzuatt95qLlmyxNy+fbs5d+5cc+vWraE2zz77rOn1es3//Oc/5urVq80f//jHZlpamllSUmJh5eHn6aefNhMSEszZs2ebO3bsMN977z0zJibGfOmll0JtdK5Pz8cff2z+5je/Md9//30TMGfNmlVrfV3O66WXXmoOGDDAXLx4sfnVV1+Z3bt3N2+44YYGqU8Bpcq5555rTpgwIfS33+83U1JSzGnTpllYVcuTk5NjAubChQtN0zTNvLw80+l0mu+9916ozXfffWcCZmZmplVlhrWCggKzR48e5rx588wf/OAHoYCic91wHnnkEfO888474fpAIGAmJyebv//970PL8vLyTLfbbf7zn/9sihJbjDFjxpi33357rWVjx441x48fb5qmznVD+X5Aqct53bBhgwmYy5YtC7WZM2eOaRiGuW/fvjOuSbd4gPLyclasWMHIkSNDy2w2GyNHjiQzM9PCylqe/Px8AOLj4wFYsWIFFRUVtc597969SU1N1bk/TRMmTGDMmDG1zinoXDekDz74gCFDhnDttdeSmJjIoEGDePPNN0Prd+zYQVZWVq1z7fV6GTp0qM51PQ0fPpz58+ezefNmAFavXs3XX3/N6NGjAZ3rxlKX85qZmUlcXBxDhgwJtRk5ciQ2m40lS5accQ1h+bLAhnbo0CH8fj9JSUm1liclJbFx40aLqmp5AoEAkyZNYsSIEfTt2xeArKwsXC4XcXFxtdomJSWRlZVlQZXh7d133+Xbb79l2bJlx6zTuW4427dv57XXXmPy5Mn8+te/ZtmyZdx///24XC5uueWW0Pk83n9TdK7r51e/+hU+n4/evXtjt9vx+/08/fTTjB8/HkDnupHU5bxmZWWRmJhYa73D4SA+Pr5Bzr0CijSZCRMmsG7dOr7++murS2mR9uzZwwMPPMC8efOIiIiwupwWLRAIMGTIEJ555hkABg0axLp163j99de55ZZbLK6uZfnXv/7FzJkzeeeddzjrrLNYtWoVkyZNIiUlRee6hdMtHqBt27bY7fZjnmbIzs4mOTnZoqpalokTJzJ79mw+//xzOnbsGFqenJxMeXk5eXl5tdrr3NffihUryMnJ4eyzz8bhcOBwOFi4cCEvv/wyDoeDpKQknesG0r59e9LT02st69OnD7t37wYInU/9N+XMPfzww/zqV7/i+uuvp1+/ftx00008+OCDTJs2DdC5bix1Oa/Jycnk5OTUWl9ZWUlubm6DnHsFFMDlcjF48GDmz58fWhYIBJg/fz4ZGRkWVhb+TNNk4sSJzJo1iwULFpCWllZr/eDBg3E6nbXO/aZNm9i9e7fOfT1dfPHFrF27llWrVoWmIUOGMH78+NC8znXDGDFixDGPy2/evJnOnTsDkJaWRnJycq1z7fP5WLJkic51PRUXF2Oz1f6pstvtBAIBQOe6sdTlvGZkZJCXl8eKFStCbRYsWEAgEGDo0KFnXsQZd7NtId59913T7XabM2bMMDds2GDeddddZlxcnJmVlWV1aWHt3nvvNb1er/nFF1+YBw4cCE3FxcWhNvfcc4+ZmppqLliwwFy+fLmZkZFhZmRkWFh1y1HzKR7T1LluKEuXLjUdDof59NNPm1u2bDFnzpxpRkVFmW+//XaozbPPPmvGxcWZ//3vf801a9aYV155pR59PQ233HKL2aFDh9Bjxu+//77Ztm1b85e//GWojc716SkoKDBXrlxprly50gTMF154wVy5cqW5a9cu0zTrdl4vvfRSc9CgQeaSJUvMr7/+2uzRo4ceM24Mf/rTn8zU1FTT5XKZ5557rrl48WKrSwp7wHGn6dOnh9qUlJSYP//5z802bdqYUVFR5tVXX20eOHDAuqJbkO8HFJ3rhvPhhx+affv2Nd1ut9m7d2/zjTfeqLU+EAiYv/3tb82kpCTT7XabF198sblp0yaLqg1fPp/PfOCBB8zU1FQzIiLC7Nq1q/mb3/zGLCsrC7XRuT49n3/++XH/+3zLLbeYplm383r48GHzhhtuMGNiYkyPx2PedtttZkFBQYPUZ5hmjeH4RERERJoB9UERERGRZkcBRURERJodBRQRERFpdhRQREREpNlRQBEREZFmRwFFREREmh0FFBEREWl2FFBERESk2VFAERERkWZHAUVERESaHQUUERERaXYUUERERKTZ+f/HPEiAz6Js4QAAAABJRU5ErkJggg==\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# insurance data"
      ],
      "metadata": {
        "id": "U_k-o1b1FA5u"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "insurance = pd.read_csv('insurance.csv')\n",
        "df = insurance.copy()\n",
        "df.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "6Un-FpsA5BqW",
        "outputId": "8e69cdaf-1185-4b0f-f237-23d3e6b31d86"
      },
      "execution_count": 115,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "   age     sex     bmi  children smoker     region      charges\n",
              "0   19  female  27.900         0    yes  southwest  16884.92400\n",
              "1   18    male  33.770         1     no  southeast   1725.55230\n",
              "2   28    male  33.000         3     no  southeast   4449.46200\n",
              "3   33    male  22.705         0     no  northwest  21984.47061\n",
              "4   32    male  28.880         0     no  northwest   3866.85520"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-27f3deba-3e38-494e-ae09-b13ff8614f3f\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>age</th>\n",
              "      <th>sex</th>\n",
              "      <th>bmi</th>\n",
              "      <th>children</th>\n",
              "      <th>smoker</th>\n",
              "      <th>region</th>\n",
              "      <th>charges</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>19</td>\n",
              "      <td>female</td>\n",
              "      <td>27.900</td>\n",
              "      <td>0</td>\n",
              "      <td>yes</td>\n",
              "      <td>southwest</td>\n",
              "      <td>16884.92400</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>18</td>\n",
              "      <td>male</td>\n",
              "      <td>33.770</td>\n",
              "      <td>1</td>\n",
              "      <td>no</td>\n",
              "      <td>southeast</td>\n",
              "      <td>1725.55230</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>28</td>\n",
              "      <td>male</td>\n",
              "      <td>33.000</td>\n",
              "      <td>3</td>\n",
              "      <td>no</td>\n",
              "      <td>southeast</td>\n",
              "      <td>4449.46200</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>33</td>\n",
              "      <td>male</td>\n",
              "      <td>22.705</td>\n",
              "      <td>0</td>\n",
              "      <td>no</td>\n",
              "      <td>northwest</td>\n",
              "      <td>21984.47061</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>32</td>\n",
              "      <td>male</td>\n",
              "      <td>28.880</td>\n",
              "      <td>0</td>\n",
              "      <td>no</td>\n",
              "      <td>northwest</td>\n",
              "      <td>3866.85520</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-27f3deba-3e38-494e-ae09-b13ff8614f3f')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-27f3deba-3e38-494e-ae09-b13ff8614f3f button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-27f3deba-3e38-494e-ae09-b13ff8614f3f');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 115
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import tensorflow as tf\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt"
      ],
      "metadata": {
        "id": "PBjp6t9n5hlZ"
      },
      "execution_count": 116,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df.info()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "e2cNZBx_6SPw",
        "outputId": "177fa779-c02d-4141-991b-0d0529e37ff1"
      },
      "execution_count": 117,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 1338 entries, 0 to 1337\n",
            "Data columns (total 7 columns):\n",
            " #   Column    Non-Null Count  Dtype  \n",
            "---  ------    --------------  -----  \n",
            " 0   age       1338 non-null   int64  \n",
            " 1   sex       1338 non-null   object \n",
            " 2   bmi       1338 non-null   float64\n",
            " 3   children  1338 non-null   int64  \n",
            " 4   smoker    1338 non-null   object \n",
            " 5   region    1338 non-null   object \n",
            " 6   charges   1338 non-null   float64\n",
            "dtypes: float64(2), int64(2), object(3)\n",
            "memory usage: 73.3+ KB\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df.isna().sum()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "29I1WuOS6aaQ",
        "outputId": "75c89374-a7d8-4cf6-92aa-5533bd2164e2"
      },
      "execution_count": 119,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "age         0\n",
              "sex         0\n",
              "bmi         0\n",
              "children    0\n",
              "smoker      0\n",
              "region      0\n",
              "charges     0\n",
              "dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 119
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df['region'].value_counts()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "PYuoDOLU7HNh",
        "outputId": "5e23e9ae-91f9-403b-8f9d-84204b10f376"
      },
      "execution_count": 125,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "southeast    364\n",
              "southwest    325\n",
              "northwest    325\n",
              "northeast    324\n",
              "Name: region, dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 125
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.preprocessing import LabelEncoder\n",
        "\n",
        "le = LabelEncoder()\n",
        "df['sex'] = le.fit_transform(df['sex'])\n",
        "df['smoker'] = le.fit_transform(df['smoker'])\n",
        "df['region'] = le.fit_transform(df['region'])\n",
        "df['']"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nzoF-ct97Wny",
        "outputId": "6fa59216-4027-478d-a819-42fda32be693"
      },
      "execution_count": 141,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0       3\n",
              "1       2\n",
              "2       2\n",
              "3       1\n",
              "4       1\n",
              "       ..\n",
              "1333    1\n",
              "1334    0\n",
              "1335    2\n",
              "1336    3\n",
              "1337    1\n",
              "Name: region, Length: 1338, dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 141
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.preprocessing import StandardScaler\n",
        "X = df.iloc[:,:6]\n",
        "y = df['charges']"
      ],
      "metadata": {
        "id": "XqQF25tM6cci"
      },
      "execution_count": 127,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)"
      ],
      "metadata": {
        "id": "IGrP42Q278Vb"
      },
      "execution_count": 128,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.preprocessing import StandardScaler\n",
        "\n",
        "ss = StandardScaler()\n",
        "X_train_s = ss.fit_transform(X_train)\n",
        "X_test_s = ss.transform(X_test)"
      ],
      "metadata": {
        "id": "SE0pT5ad8N7A"
      },
      "execution_count": 129,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow import keras\n",
        "from keras import layers\n",
        "\n",
        "model = keras.Sequential([\n",
        "    layers.Dense(units=32, activation='relu', input_shape=(6,)),\n",
        "    layers.Dense(units=16, activation='relu'),\n",
        "    layers.Dense(units=1) #종속변수가 charges로 하나이므로 출력 layer는 1\n",
        "])"
      ],
      "metadata": {
        "id": "-h8krosC8SJ6"
      },
      "execution_count": 130,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rNvr2DxS8Sng",
        "outputId": "81229b85-14ae-46e6-c2ea-25f6f96553e5"
      },
      "execution_count": 131,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_3\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " dense_22 (Dense)            (None, 32)                224       \n",
            "                                                                 \n",
            " dense_23 (Dense)            (None, 16)                528       \n",
            "                                                                 \n",
            " dense_24 (Dense)            (None, 1)                 17        \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 769\n",
            "Trainable params: 769\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        " model.compile(\n",
        "    optimizer='adam',\n",
        "    loss='mse',\n",
        "    metrics=['mse', 'mae']\n",
        " )"
      ],
      "metadata": {
        "id": "kdbvMCbt8jNv"
      },
      "execution_count": 139,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "epochs = 100\n",
        "batch_size = 12\n",
        "\n",
        "model.fit(\n",
        "    X_train_s, y_train,\n",
        "    epochs = epochs,\n",
        "    batch_size = batch_size,\n",
        "    validation_split=0.2,\n",
        "    verbose = 1\n",
        ")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "v7vtDl1m8n3x",
        "outputId": "8bcb0294-58e2-44f9-b727-d83f65be72d1"
      },
      "execution_count": 140,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "72/72 [==============================] - 1s 6ms/step - loss: 35643056.0000 - mse: 35643056.0000 - mae: 4101.2280 - val_loss: 32437090.0000 - val_mse: 32437090.0000 - val_mae: 3769.0503\n",
            "Epoch 2/100\n",
            "72/72 [==============================] - 0s 2ms/step - loss: 35599616.0000 - mse: 35599616.0000 - mae: 4095.2283 - val_loss: 32455398.0000 - val_mse: 32455398.0000 - val_mae: 3767.0845\n",
            "Epoch 3/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35576588.0000 - mse: 35576588.0000 - mae: 4092.0000 - val_loss: 32446852.0000 - val_mse: 32446852.0000 - val_mae: 3767.5415\n",
            "Epoch 4/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35562032.0000 - mse: 35562032.0000 - mae: 4093.2961 - val_loss: 32461888.0000 - val_mse: 32461888.0000 - val_mae: 3768.6477\n",
            "Epoch 5/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35514304.0000 - mse: 35514304.0000 - mae: 4088.2383 - val_loss: 32468710.0000 - val_mse: 32468710.0000 - val_mae: 3769.9871\n",
            "Epoch 6/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35510096.0000 - mse: 35510100.0000 - mae: 4087.8716 - val_loss: 32467220.0000 - val_mse: 32467220.0000 - val_mae: 3771.7568\n",
            "Epoch 7/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35486828.0000 - mse: 35486824.0000 - mae: 4082.0330 - val_loss: 32504650.0000 - val_mse: 32504650.0000 - val_mae: 3769.2964\n",
            "Epoch 8/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35474460.0000 - mse: 35474464.0000 - mae: 4074.1594 - val_loss: 32538132.0000 - val_mse: 32538132.0000 - val_mae: 3768.4807\n",
            "Epoch 9/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35427052.0000 - mse: 35427052.0000 - mae: 4077.7693 - val_loss: 32509102.0000 - val_mse: 32509102.0000 - val_mae: 3769.1443\n",
            "Epoch 10/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35424612.0000 - mse: 35424612.0000 - mae: 4077.5164 - val_loss: 32526974.0000 - val_mse: 32526974.0000 - val_mae: 3768.9124\n",
            "Epoch 11/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35382896.0000 - mse: 35382892.0000 - mae: 4071.5464 - val_loss: 32553460.0000 - val_mse: 32553460.0000 - val_mae: 3767.9924\n",
            "Epoch 12/100\n",
            "72/72 [==============================] - 0s 4ms/step - loss: 35378104.0000 - mse: 35378104.0000 - mae: 4071.1499 - val_loss: 32600086.0000 - val_mse: 32600086.0000 - val_mae: 3772.2385\n",
            "Epoch 13/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35355332.0000 - mse: 35355332.0000 - mae: 4073.9658 - val_loss: 32520694.0000 - val_mse: 32520694.0000 - val_mae: 3771.1536\n",
            "Epoch 14/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35338124.0000 - mse: 35338124.0000 - mae: 4071.1895 - val_loss: 32554608.0000 - val_mse: 32554608.0000 - val_mae: 3770.8440\n",
            "Epoch 15/100\n",
            "72/72 [==============================] - 0s 4ms/step - loss: 35318984.0000 - mse: 35318984.0000 - mae: 4070.3904 - val_loss: 32559118.0000 - val_mse: 32559118.0000 - val_mae: 3772.3438\n",
            "Epoch 16/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35293372.0000 - mse: 35293372.0000 - mae: 4072.9299 - val_loss: 32501274.0000 - val_mse: 32501274.0000 - val_mae: 3772.1877\n",
            "Epoch 17/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35309964.0000 - mse: 35309964.0000 - mae: 4074.8381 - val_loss: 32522230.0000 - val_mse: 32522230.0000 - val_mae: 3772.9668\n",
            "Epoch 18/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35257732.0000 - mse: 35257732.0000 - mae: 4065.5234 - val_loss: 32564346.0000 - val_mse: 32564346.0000 - val_mae: 3770.6814\n",
            "Epoch 19/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35240348.0000 - mse: 35240348.0000 - mae: 4069.7759 - val_loss: 32523650.0000 - val_mse: 32523650.0000 - val_mae: 3775.4260\n",
            "Epoch 20/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35220328.0000 - mse: 35220328.0000 - mae: 4070.0540 - val_loss: 32513862.0000 - val_mse: 32513862.0000 - val_mae: 3773.1240\n",
            "Epoch 21/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35211160.0000 - mse: 35211160.0000 - mae: 4065.7891 - val_loss: 32529334.0000 - val_mse: 32529334.0000 - val_mae: 3774.8264\n",
            "Epoch 22/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35179524.0000 - mse: 35179524.0000 - mae: 4064.6277 - val_loss: 32541728.0000 - val_mse: 32541728.0000 - val_mae: 3774.8984\n",
            "Epoch 23/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35169076.0000 - mse: 35169076.0000 - mae: 4069.3047 - val_loss: 32526808.0000 - val_mse: 32526808.0000 - val_mae: 3774.8345\n",
            "Epoch 24/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35169156.0000 - mse: 35169156.0000 - mae: 4069.0369 - val_loss: 32509216.0000 - val_mse: 32509216.0000 - val_mae: 3776.0159\n",
            "Epoch 25/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35129352.0000 - mse: 35129352.0000 - mae: 4060.3660 - val_loss: 32541248.0000 - val_mse: 32541248.0000 - val_mae: 3774.9644\n",
            "Epoch 26/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35119700.0000 - mse: 35119700.0000 - mae: 4062.2769 - val_loss: 32582676.0000 - val_mse: 32582676.0000 - val_mae: 3773.0884\n",
            "Epoch 27/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35113596.0000 - mse: 35113596.0000 - mae: 4066.6240 - val_loss: 32506774.0000 - val_mse: 32506774.0000 - val_mae: 3773.2673\n",
            "Epoch 28/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35106300.0000 - mse: 35106300.0000 - mae: 4067.3882 - val_loss: 32532802.0000 - val_mse: 32532802.0000 - val_mae: 3777.7803\n",
            "Epoch 29/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35073896.0000 - mse: 35073896.0000 - mae: 4066.0947 - val_loss: 32502226.0000 - val_mse: 32502226.0000 - val_mae: 3777.6677\n",
            "Epoch 30/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35050368.0000 - mse: 35050368.0000 - mae: 4062.2205 - val_loss: 32546856.0000 - val_mse: 32546854.0000 - val_mae: 3776.1772\n",
            "Epoch 31/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35030824.0000 - mse: 35030820.0000 - mae: 4061.8467 - val_loss: 32521470.0000 - val_mse: 32521470.0000 - val_mae: 3776.3398\n",
            "Epoch 32/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35015592.0000 - mse: 35015592.0000 - mae: 4061.1672 - val_loss: 32523480.0000 - val_mse: 32523480.0000 - val_mae: 3775.2087\n",
            "Epoch 33/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 35011724.0000 - mse: 35011724.0000 - mae: 4060.2383 - val_loss: 32533870.0000 - val_mse: 32533870.0000 - val_mae: 3776.0962\n",
            "Epoch 34/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34994156.0000 - mse: 34994156.0000 - mae: 4053.6133 - val_loss: 32568870.0000 - val_mse: 32568870.0000 - val_mae: 3773.9619\n",
            "Epoch 35/100\n",
            "72/72 [==============================] - 0s 4ms/step - loss: 34965072.0000 - mse: 34965072.0000 - mae: 4055.4907 - val_loss: 32487294.0000 - val_mse: 32487294.0000 - val_mae: 3774.6909\n",
            "Epoch 36/100\n",
            "72/72 [==============================] - 0s 6ms/step - loss: 34992836.0000 - mse: 34992836.0000 - mae: 4055.8677 - val_loss: 32493408.0000 - val_mse: 32493408.0000 - val_mae: 3774.5283\n",
            "Epoch 37/100\n",
            "72/72 [==============================] - 0s 4ms/step - loss: 34945752.0000 - mse: 34945752.0000 - mae: 4054.9866 - val_loss: 32540696.0000 - val_mse: 32540696.0000 - val_mae: 3772.7847\n",
            "Epoch 38/100\n",
            "72/72 [==============================] - 0s 4ms/step - loss: 34920936.0000 - mse: 34920936.0000 - mae: 4050.6340 - val_loss: 32479666.0000 - val_mse: 32479666.0000 - val_mae: 3770.1370\n",
            "Epoch 39/100\n",
            "72/72 [==============================] - 0s 4ms/step - loss: 34908312.0000 - mse: 34908312.0000 - mae: 4049.5388 - val_loss: 32525534.0000 - val_mse: 32525534.0000 - val_mae: 3771.2705\n",
            "Epoch 40/100\n",
            "72/72 [==============================] - 0s 4ms/step - loss: 34893680.0000 - mse: 34893680.0000 - mae: 4047.6907 - val_loss: 32502898.0000 - val_mse: 32502898.0000 - val_mae: 3769.2185\n",
            "Epoch 41/100\n",
            "72/72 [==============================] - 0s 4ms/step - loss: 34877028.0000 - mse: 34877028.0000 - mae: 4043.9565 - val_loss: 32500236.0000 - val_mse: 32500236.0000 - val_mae: 3771.0488\n",
            "Epoch 42/100\n",
            "72/72 [==============================] - 0s 4ms/step - loss: 34873632.0000 - mse: 34873632.0000 - mae: 4048.1860 - val_loss: 32502356.0000 - val_mse: 32502356.0000 - val_mae: 3774.6765\n",
            "Epoch 43/100\n",
            "72/72 [==============================] - 0s 4ms/step - loss: 34846060.0000 - mse: 34846060.0000 - mae: 4048.9302 - val_loss: 32458662.0000 - val_mse: 32458662.0000 - val_mae: 3770.3481\n",
            "Epoch 44/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34822300.0000 - mse: 34822304.0000 - mae: 4046.9602 - val_loss: 32501270.0000 - val_mse: 32501270.0000 - val_mae: 3773.8418\n",
            "Epoch 45/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34812020.0000 - mse: 34812020.0000 - mae: 4049.8113 - val_loss: 32507100.0000 - val_mse: 32507100.0000 - val_mae: 3772.1514\n",
            "Epoch 46/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34793852.0000 - mse: 34793852.0000 - mae: 4041.2832 - val_loss: 32509326.0000 - val_mse: 32509326.0000 - val_mae: 3770.9399\n",
            "Epoch 47/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34768780.0000 - mse: 34768780.0000 - mae: 4042.0015 - val_loss: 32489878.0000 - val_mse: 32489878.0000 - val_mae: 3771.1411\n",
            "Epoch 48/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34765244.0000 - mse: 34765248.0000 - mae: 4043.7134 - val_loss: 32443210.0000 - val_mse: 32443210.0000 - val_mae: 3769.2251\n",
            "Epoch 49/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34753276.0000 - mse: 34753276.0000 - mae: 4037.7385 - val_loss: 32431742.0000 - val_mse: 32431742.0000 - val_mae: 3767.7190\n",
            "Epoch 50/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34726244.0000 - mse: 34726248.0000 - mae: 4041.0491 - val_loss: 32422898.0000 - val_mse: 32422898.0000 - val_mae: 3765.8933\n",
            "Epoch 51/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34701716.0000 - mse: 34701716.0000 - mae: 4037.1687 - val_loss: 32433032.0000 - val_mse: 32433032.0000 - val_mae: 3766.5793\n",
            "Epoch 52/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34708672.0000 - mse: 34708672.0000 - mae: 4041.7263 - val_loss: 32458946.0000 - val_mse: 32458946.0000 - val_mae: 3763.8630\n",
            "Epoch 53/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34667224.0000 - mse: 34667224.0000 - mae: 4030.9526 - val_loss: 32499772.0000 - val_mse: 32499772.0000 - val_mae: 3763.8689\n",
            "Epoch 54/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34649756.0000 - mse: 34649756.0000 - mae: 4028.0076 - val_loss: 32447136.0000 - val_mse: 32447136.0000 - val_mae: 3760.2078\n",
            "Epoch 55/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34641652.0000 - mse: 34641652.0000 - mae: 4030.6545 - val_loss: 32433636.0000 - val_mse: 32433636.0000 - val_mae: 3761.3083\n",
            "Epoch 56/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34623436.0000 - mse: 34623436.0000 - mae: 4029.8616 - val_loss: 32421116.0000 - val_mse: 32421116.0000 - val_mae: 3758.1401\n",
            "Epoch 57/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34608492.0000 - mse: 34608492.0000 - mae: 4025.8547 - val_loss: 32445844.0000 - val_mse: 32445844.0000 - val_mae: 3761.5769\n",
            "Epoch 58/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34594616.0000 - mse: 34594616.0000 - mae: 4023.6677 - val_loss: 32442292.0000 - val_mse: 32442292.0000 - val_mae: 3756.3071\n",
            "Epoch 59/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34569804.0000 - mse: 34569804.0000 - mae: 4023.5798 - val_loss: 32415342.0000 - val_mse: 32415342.0000 - val_mae: 3756.9646\n",
            "Epoch 60/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34550556.0000 - mse: 34550556.0000 - mae: 4023.4900 - val_loss: 32384612.0000 - val_mse: 32384612.0000 - val_mae: 3755.4639\n",
            "Epoch 61/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34539928.0000 - mse: 34539928.0000 - mae: 4021.6182 - val_loss: 32368848.0000 - val_mse: 32368852.0000 - val_mae: 3755.5759\n",
            "Epoch 62/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34518640.0000 - mse: 34518640.0000 - mae: 4019.8108 - val_loss: 32344676.0000 - val_mse: 32344676.0000 - val_mae: 3751.2693\n",
            "Epoch 63/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34526300.0000 - mse: 34526300.0000 - mae: 4014.9663 - val_loss: 32371264.0000 - val_mse: 32371264.0000 - val_mae: 3752.6353\n",
            "Epoch 64/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34508256.0000 - mse: 34508256.0000 - mae: 4019.4619 - val_loss: 32353802.0000 - val_mse: 32353802.0000 - val_mae: 3751.9766\n",
            "Epoch 65/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34476136.0000 - mse: 34476136.0000 - mae: 4019.4746 - val_loss: 32326014.0000 - val_mse: 32326010.0000 - val_mae: 3751.3469\n",
            "Epoch 66/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34456908.0000 - mse: 34456908.0000 - mae: 4019.0793 - val_loss: 32322318.0000 - val_mse: 32322314.0000 - val_mae: 3751.9836\n",
            "Epoch 67/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34437272.0000 - mse: 34437272.0000 - mae: 4017.1548 - val_loss: 32307010.0000 - val_mse: 32307010.0000 - val_mae: 3748.7732\n",
            "Epoch 68/100\n",
            "72/72 [==============================] - 0s 4ms/step - loss: 34409208.0000 - mse: 34409208.0000 - mae: 4014.2261 - val_loss: 32323428.0000 - val_mse: 32323428.0000 - val_mae: 3748.6072\n",
            "Epoch 69/100\n",
            "72/72 [==============================] - 0s 4ms/step - loss: 34405236.0000 - mse: 34405236.0000 - mae: 4009.7263 - val_loss: 32305086.0000 - val_mse: 32305086.0000 - val_mae: 3746.3789\n",
            "Epoch 70/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34392388.0000 - mse: 34392388.0000 - mae: 4008.4038 - val_loss: 32362664.0000 - val_mse: 32362664.0000 - val_mae: 3744.1062\n",
            "Epoch 71/100\n",
            "72/72 [==============================] - 0s 4ms/step - loss: 34376004.0000 - mse: 34376004.0000 - mae: 4005.8945 - val_loss: 32312304.0000 - val_mse: 32312304.0000 - val_mae: 3743.3228\n",
            "Epoch 72/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34372764.0000 - mse: 34372764.0000 - mae: 4008.6375 - val_loss: 32329702.0000 - val_mse: 32329702.0000 - val_mae: 3744.4548\n",
            "Epoch 73/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34356068.0000 - mse: 34356068.0000 - mae: 4007.8440 - val_loss: 32305140.0000 - val_mse: 32305140.0000 - val_mae: 3742.9500\n",
            "Epoch 74/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34342932.0000 - mse: 34342932.0000 - mae: 4008.4485 - val_loss: 32284988.0000 - val_mse: 32284988.0000 - val_mae: 3744.9375\n",
            "Epoch 75/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34319348.0000 - mse: 34319348.0000 - mae: 4008.0747 - val_loss: 32238716.0000 - val_mse: 32238716.0000 - val_mae: 3740.2419\n",
            "Epoch 76/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34301456.0000 - mse: 34301456.0000 - mae: 4001.4497 - val_loss: 32289732.0000 - val_mse: 32289732.0000 - val_mae: 3737.2559\n",
            "Epoch 77/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34290424.0000 - mse: 34290424.0000 - mae: 4001.0745 - val_loss: 32253340.0000 - val_mse: 32253340.0000 - val_mae: 3738.4705\n",
            "Epoch 78/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34264504.0000 - mse: 34264504.0000 - mae: 3996.9756 - val_loss: 32244258.0000 - val_mse: 32244258.0000 - val_mae: 3734.9480\n",
            "Epoch 79/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34248688.0000 - mse: 34248688.0000 - mae: 3997.2629 - val_loss: 32254284.0000 - val_mse: 32254284.0000 - val_mae: 3737.7319\n",
            "Epoch 80/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34233392.0000 - mse: 34233392.0000 - mae: 3992.5068 - val_loss: 32272946.0000 - val_mse: 32272946.0000 - val_mae: 3734.4490\n",
            "Epoch 81/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34210540.0000 - mse: 34210540.0000 - mae: 3994.3955 - val_loss: 32251968.0000 - val_mse: 32251968.0000 - val_mae: 3738.8135\n",
            "Epoch 82/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34185020.0000 - mse: 34185020.0000 - mae: 3992.8845 - val_loss: 32234864.0000 - val_mse: 32234864.0000 - val_mae: 3735.8140\n",
            "Epoch 83/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34186064.0000 - mse: 34186064.0000 - mae: 3998.4834 - val_loss: 32191858.0000 - val_mse: 32191858.0000 - val_mae: 3736.3081\n",
            "Epoch 84/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34171436.0000 - mse: 34171436.0000 - mae: 3991.8149 - val_loss: 32160600.0000 - val_mse: 32160600.0000 - val_mae: 3731.0786\n",
            "Epoch 85/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34144144.0000 - mse: 34144144.0000 - mae: 3993.1235 - val_loss: 32201444.0000 - val_mse: 32201444.0000 - val_mae: 3731.6736\n",
            "Epoch 86/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34137520.0000 - mse: 34137520.0000 - mae: 3988.6274 - val_loss: 32237766.0000 - val_mse: 32237766.0000 - val_mae: 3729.1567\n",
            "Epoch 87/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 34117656.0000 - mse: 34117656.0000 - mae: 3984.3301 - val_loss: 32170268.0000 - val_mse: 32170268.0000 - val_mae: 3728.8901\n",
            "Epoch 88/100\n",
            "72/72 [==============================] - 0s 5ms/step - loss: 34102224.0000 - mse: 34102224.0000 - mae: 3984.4629 - val_loss: 32175326.0000 - val_mse: 32175326.0000 - val_mae: 3726.5225\n",
            "Epoch 89/100\n",
            "72/72 [==============================] - 0s 5ms/step - loss: 34108488.0000 - mse: 34108488.0000 - mae: 3987.9016 - val_loss: 32191220.0000 - val_mse: 32191220.0000 - val_mae: 3727.4773\n",
            "Epoch 90/100\n",
            "72/72 [==============================] - 0s 5ms/step - loss: 34062956.0000 - mse: 34062956.0000 - mae: 3980.6345 - val_loss: 32187040.0000 - val_mse: 32187040.0000 - val_mae: 3728.6245\n",
            "Epoch 91/100\n",
            "72/72 [==============================] - 0s 4ms/step - loss: 34060720.0000 - mse: 34060720.0000 - mae: 3977.6746 - val_loss: 32199258.0000 - val_mse: 32199258.0000 - val_mae: 3723.5095\n",
            "Epoch 92/100\n",
            "72/72 [==============================] - 0s 4ms/step - loss: 34041272.0000 - mse: 34041272.0000 - mae: 3978.2429 - val_loss: 32150068.0000 - val_mse: 32150068.0000 - val_mae: 3727.9976\n",
            "Epoch 93/100\n",
            "72/72 [==============================] - 0s 4ms/step - loss: 34078304.0000 - mse: 34078304.0000 - mae: 3978.5176 - val_loss: 32167420.0000 - val_mse: 32167420.0000 - val_mae: 3723.9475\n",
            "Epoch 94/100\n",
            "72/72 [==============================] - 0s 4ms/step - loss: 34015872.0000 - mse: 34015872.0000 - mae: 3975.6443 - val_loss: 32170512.0000 - val_mse: 32170512.0000 - val_mae: 3724.4585\n",
            "Epoch 95/100\n",
            "72/72 [==============================] - 0s 4ms/step - loss: 33977000.0000 - mse: 33977000.0000 - mae: 3975.5598 - val_loss: 32130266.0000 - val_mse: 32130266.0000 - val_mae: 3720.7056\n",
            "Epoch 96/100\n",
            "72/72 [==============================] - 0s 4ms/step - loss: 33960252.0000 - mse: 33960252.0000 - mae: 3972.6494 - val_loss: 32141814.0000 - val_mse: 32141814.0000 - val_mae: 3720.5447\n",
            "Epoch 97/100\n",
            "72/72 [==============================] - 0s 4ms/step - loss: 33952248.0000 - mse: 33952248.0000 - mae: 3969.5969 - val_loss: 32168620.0000 - val_mse: 32168620.0000 - val_mae: 3716.4087\n",
            "Epoch 98/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 33954984.0000 - mse: 33954984.0000 - mae: 3966.5540 - val_loss: 32150868.0000 - val_mse: 32150868.0000 - val_mae: 3716.2419\n",
            "Epoch 99/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 33919800.0000 - mse: 33919800.0000 - mae: 3962.1091 - val_loss: 32175346.0000 - val_mse: 32175346.0000 - val_mae: 3715.2646\n",
            "Epoch 100/100\n",
            "72/72 [==============================] - 0s 3ms/step - loss: 33922592.0000 - mse: 33922592.0000 - mae: 3962.6018 - val_loss: 32184494.0000 - val_mse: 32184494.0000 - val_mae: 3714.4216\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7fa003e72470>"
            ]
          },
          "metadata": {},
          "execution_count": 140
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.evaluate(X_train_s, y_train)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3lJUsJ4b8txc",
        "outputId": "8ca4bf1a-a6c3-449c-ccd2-0df929514fa2"
      },
      "execution_count": 142,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "34/34 [==============================] - 0s 2ms/step - loss: 33528010.0000 - mse: 33528010.0000 - mae: 3909.6204\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[33528010.0, 33528010.0, 3909.620361328125]"
            ]
          },
          "metadata": {},
          "execution_count": 142
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "history.history.keys()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QZfP_CC589dg",
        "outputId": "8caa6102-dc2d-42ce-be65-ae1b6a54f82c"
      },
      "execution_count": 143,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "dict_keys(['loss', 'mse', 'mae', 'val_loss', 'val_mse', 'val_mae'])"
            ]
          },
          "metadata": {},
          "execution_count": 143
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "hist = pd.DataFrame(history.history)\n",
        "hist['epoch'] = history.epoch\n",
        "hist"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 423
        },
        "id": "zlC5UAob8-bb",
        "outputId": "f2f2087f-cea4-40a3-d841-009351d92a4b"
      },
      "execution_count": 144,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "          loss         mse        mae    val_loss     val_mse    val_mae  \\\n",
              "0   584.483521  584.483521  22.733809  663.135498  663.135498  24.457890   \n",
              "1   575.533325  575.533325  22.511660  653.566956  653.566956  24.243902   \n",
              "2   565.741821  565.741821  22.257097  642.982544  642.982544  23.998941   \n",
              "3   554.583862  554.583862  21.964846  631.068298  631.068298  23.709518   \n",
              "4   542.278015  542.278015  21.624388  617.306885  617.306885  23.373186   \n",
              "..         ...         ...        ...         ...         ...        ...   \n",
              "95   16.587162   16.587162   3.044141   14.856238   14.856238   2.805383   \n",
              "96   16.511187   16.511187   3.036433   14.770385   14.770385   2.801529   \n",
              "97   16.474401   16.474401   3.032139   14.696231   14.696231   2.797776   \n",
              "98   16.360373   16.360373   3.019833   14.615176   14.615176   2.795108   \n",
              "99   16.294367   16.294367   3.015908   14.642282   14.642282   2.801860   \n",
              "\n",
              "    epoch  \n",
              "0       0  \n",
              "1       1  \n",
              "2       2  \n",
              "3       3  \n",
              "4       4  \n",
              "..    ...  \n",
              "95     95  \n",
              "96     96  \n",
              "97     97  \n",
              "98     98  \n",
              "99     99  \n",
              "\n",
              "[100 rows x 7 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-1a8946d8-5eb7-4a4f-8162-c4213ddfed2c\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>loss</th>\n",
              "      <th>mse</th>\n",
              "      <th>mae</th>\n",
              "      <th>val_loss</th>\n",
              "      <th>val_mse</th>\n",
              "      <th>val_mae</th>\n",
              "      <th>epoch</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>584.483521</td>\n",
              "      <td>584.483521</td>\n",
              "      <td>22.733809</td>\n",
              "      <td>663.135498</td>\n",
              "      <td>663.135498</td>\n",
              "      <td>24.457890</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>575.533325</td>\n",
              "      <td>575.533325</td>\n",
              "      <td>22.511660</td>\n",
              "      <td>653.566956</td>\n",
              "      <td>653.566956</td>\n",
              "      <td>24.243902</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>565.741821</td>\n",
              "      <td>565.741821</td>\n",
              "      <td>22.257097</td>\n",
              "      <td>642.982544</td>\n",
              "      <td>642.982544</td>\n",
              "      <td>23.998941</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>554.583862</td>\n",
              "      <td>554.583862</td>\n",
              "      <td>21.964846</td>\n",
              "      <td>631.068298</td>\n",
              "      <td>631.068298</td>\n",
              "      <td>23.709518</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>542.278015</td>\n",
              "      <td>542.278015</td>\n",
              "      <td>21.624388</td>\n",
              "      <td>617.306885</td>\n",
              "      <td>617.306885</td>\n",
              "      <td>23.373186</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>95</th>\n",
              "      <td>16.587162</td>\n",
              "      <td>16.587162</td>\n",
              "      <td>3.044141</td>\n",
              "      <td>14.856238</td>\n",
              "      <td>14.856238</td>\n",
              "      <td>2.805383</td>\n",
              "      <td>95</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>96</th>\n",
              "      <td>16.511187</td>\n",
              "      <td>16.511187</td>\n",
              "      <td>3.036433</td>\n",
              "      <td>14.770385</td>\n",
              "      <td>14.770385</td>\n",
              "      <td>2.801529</td>\n",
              "      <td>96</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>97</th>\n",
              "      <td>16.474401</td>\n",
              "      <td>16.474401</td>\n",
              "      <td>3.032139</td>\n",
              "      <td>14.696231</td>\n",
              "      <td>14.696231</td>\n",
              "      <td>2.797776</td>\n",
              "      <td>97</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>98</th>\n",
              "      <td>16.360373</td>\n",
              "      <td>16.360373</td>\n",
              "      <td>3.019833</td>\n",
              "      <td>14.615176</td>\n",
              "      <td>14.615176</td>\n",
              "      <td>2.795108</td>\n",
              "      <td>98</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>99</th>\n",
              "      <td>16.294367</td>\n",
              "      <td>16.294367</td>\n",
              "      <td>3.015908</td>\n",
              "      <td>14.642282</td>\n",
              "      <td>14.642282</td>\n",
              "      <td>2.801860</td>\n",
              "      <td>99</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>100 rows × 7 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-1a8946d8-5eb7-4a4f-8162-c4213ddfed2c')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-1a8946d8-5eb7-4a4f-8162-c4213ddfed2c button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-1a8946d8-5eb7-4a4f-8162-c4213ddfed2c');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 144
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.plot(hist['epoch'], hist['mse'], label='train_mse')\n",
        "plt.plot(hist['epoch'], hist['val_mse'], label='val_mse')\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 430
        },
        "id": "qTUb_-gJ9Bo1",
        "outputId": "5397df45-4ed3-4736-f9fc-f7d20ec50bb9"
      },
      "execution_count": 145,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    }
  ]
}