Copy_of_S+P_Week_3_Exercise_Answer.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "name": "Copy of S+P Week 3 Exercise Answer.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true,
      "include_colab_link": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/RhuthuHegde/Tensorflow-coursera/blob/master/Copy_of_S%2BP_Week_3_Exercise_Answer.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zX4Kg8DUTKWO",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#@title Licensed under the Apache License, Version 2.0 (the \"License\");\n",
        "# you may not use this file except in compliance with the License.\n",
        "# You may obtain a copy of the License at\n",
        "#\n",
        "# https://www.apache.org/licenses/LICENSE-2.0\n",
        "#\n",
        "# Unless required by applicable law or agreed to in writing, software\n",
        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
        "# See the License for the specific language governing permissions and\n",
        "# limitations under the License."
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "D1J15Vh_1Jih",
        "colab": {}
      },
      "source": [
        "!pip install tf-nightly-2.0-preview\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "BOjujz601HcS",
        "colab": {}
      },
      "source": [
        "import tensorflow as tf\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "print(tf.__version__)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "Zswl7jRtGzkk",
        "colab": {}
      },
      "source": [
        "def plot_series(time, series, format=\"-\", start=0, end=None):\n",
        "    plt.plot(time[start:end], series[start:end], format)\n",
        "    plt.xlabel(\"Time\")\n",
        "    plt.ylabel(\"Value\")\n",
        "    plt.grid(False)\n",
        "\n",
        "def trend(time, slope=0):\n",
        "    return slope * time\n",
        "\n",
        "def seasonal_pattern(season_time):\n",
        "    \"\"\"Just an arbitrary pattern, you can change it if you wish\"\"\"\n",
        "    return np.where(season_time < 0.1,\n",
        "                    np.cos(season_time * 6 * np.pi),\n",
        "                    2 / np.exp(9 * season_time))\n",
        "\n",
        "def seasonality(time, period, amplitude=1, phase=0):\n",
        "    \"\"\"Repeats the same pattern at each period\"\"\"\n",
        "    season_time = ((time + phase) % period) / period\n",
        "    return amplitude * seasonal_pattern(season_time)\n",
        "\n",
        "def noise(time, noise_level=1, seed=None):\n",
        "    rnd = np.random.RandomState(seed)\n",
        "    return rnd.randn(len(time)) * noise_level\n",
        "\n",
        "time = np.arange(10 * 365 + 1, dtype=\"float32\")\n",
        "baseline = 10\n",
        "series = trend(time, 0.1)  \n",
        "baseline = 10\n",
        "amplitude = 40\n",
        "slope = 0.005\n",
        "noise_level = 3\n",
        "\n",
        "# Create the series\n",
        "series = baseline + trend(time, slope) + seasonality(time, period=365, amplitude=amplitude)\n",
        "# Update with noise\n",
        "series += noise(time, noise_level, seed=51)\n",
        "\n",
        "split_time = 3000\n",
        "time_train = time[:split_time]\n",
        "x_train = series[:split_time]\n",
        "time_valid = time[split_time:]\n",
        "x_valid = series[split_time:]\n",
        "\n",
        "window_size = 20\n",
        "batch_size = 32\n",
        "shuffle_buffer_size = 1000\n",
        "\n",
        "plot_series(time, series)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "4sTTIOCbyShY",
        "colab": {}
      },
      "source": [
        "def windowed_dataset(series, window_size, batch_size, shuffle_buffer):\n",
        "  dataset = tf.data.Dataset.from_tensor_slices(series)\n",
        "  dataset = dataset.window(window_size + 1, shift=1, drop_remainder=True)\n",
        "  dataset = dataset.flat_map(lambda window: window.batch(window_size + 1))\n",
        "  dataset = dataset.shuffle(shuffle_buffer).map(lambda window: (window[:-1], window[-1]))\n",
        "  dataset = dataset.batch(batch_size).prefetch(1)\n",
        "  return dataset"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "A1Hl39rklkLm",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "2593df96-7e3a-434e-97a0-c251b0716f60"
      },
      "source": [
        "tf.keras.backend.clear_session()\n",
        "tf.random.set_seed(51)\n",
        "np.random.seed(51)\n",
        "\n",
        "tf.keras.backend.clear_session()\n",
        "dataset = windowed_dataset(x_train, window_size, batch_size, shuffle_buffer_size)\n",
        "\n",
        "model = tf.keras.models.Sequential([\n",
        "  tf.keras.layers.Lambda(lambda x: tf.expand_dims(x, axis=-1),\n",
        "                      input_shape=[None]),\n",
        "  tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(32, return_sequences=True)),\n",
        "  tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(32)),\n",
        "  tf.keras.layers.Dense(1),\n",
        "  tf.keras.layers.Lambda(lambda x: x * 10.0)\n",
        "])\n",
        "\n",
        "lr_schedule = tf.keras.callbacks.LearningRateScheduler(\n",
        "    lambda epoch: 1e-8 * 10**(epoch / 20))\n",
        "optimizer = tf.keras.optimizers.SGD(lr=1e-8, momentum=0.9)\n",
        "model.compile(loss=tf.keras.losses.Huber(),\n",
        "              optimizer=optimizer,\n",
        "              metrics=[\"mae\"])\n",
        "history = model.fit(dataset, epochs=100, callbacks=[lr_schedule])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/100\n",
            "94/94 [==============================] - 7s 78ms/step - loss: 20.5121 - mae: 20.8828\n",
            "Epoch 2/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.3433 - mae: 20.8543\n",
            "Epoch 3/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.3187 - mae: 20.8218\n",
            "Epoch 4/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.2791 - mae: 20.7851\n",
            "Epoch 5/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.2381 - mae: 20.7436\n",
            "Epoch 6/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.1996 - mae: 20.6964\n",
            "Epoch 7/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.1450 - mae: 20.6429\n",
            "Epoch 8/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.0881 - mae: 20.5822\n",
            "Epoch 9/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.0125 - mae: 20.5140\n",
            "Epoch 10/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.9405 - mae: 20.4385\n",
            "Epoch 11/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.8561 - mae: 20.3585\n",
            "Epoch 12/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.7634 - mae: 20.2778\n",
            "Epoch 13/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.6917 - mae: 20.1967\n",
            "Epoch 14/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.6044 - mae: 20.1109\n",
            "Epoch 15/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.5195 - mae: 20.0163\n",
            "Epoch 16/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.4093 - mae: 19.9109\n",
            "Epoch 17/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.2939 - mae: 19.7932\n",
            "Epoch 18/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.1644 - mae: 19.6618\n",
            "Epoch 19/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.0151 - mae: 19.5149\n",
            "Epoch 20/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 18.8512 - mae: 19.3502\n",
            "Epoch 21/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 18.6664 - mae: 19.1656\n",
            "Epoch 22/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 18.4469 - mae: 18.9582\n",
            "Epoch 23/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 18.2335 - mae: 18.7249\n",
            "Epoch 24/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 17.9459 - mae: 18.4622\n",
            "Epoch 25/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 17.6673 - mae: 18.1658\n",
            "Epoch 26/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 17.3269 - mae: 17.8312\n",
            "Epoch 27/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 16.9499 - mae: 17.4553\n",
            "Epoch 28/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 16.5333 - mae: 17.0354\n",
            "Epoch 29/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 16.0706 - mae: 16.5711\n",
            "Epoch 30/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 15.5682 - mae: 16.0634\n",
            "Epoch 31/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 15.0123 - mae: 15.5110\n",
            "Epoch 32/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 14.4271 - mae: 14.9139\n",
            "Epoch 33/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 13.7694 - mae: 14.2724\n",
            "Epoch 34/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 13.0996 - mae: 13.5898\n",
            "Epoch 35/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 12.3900 - mae: 12.8765\n",
            "Epoch 36/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 11.6543 - mae: 12.1504\n",
            "Epoch 37/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 10.9283 - mae: 11.4243\n",
            "Epoch 38/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 10.2272 - mae: 10.7178\n",
            "Epoch 39/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 9.5821 - mae: 10.0603\n",
            "Epoch 40/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 8.9960 - mae: 9.4779\n",
            "Epoch 41/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 8.4972 - mae: 8.9765\n",
            "Epoch 42/100\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 8.0762 - mae: 8.5570\n",
            "Epoch 43/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 7.7360 - mae: 8.2183\n",
            "Epoch 44/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 7.4835 - mae: 7.9602\n",
            "Epoch 45/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 7.2636 - mae: 7.7630\n",
            "Epoch 46/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 7.1312 - mae: 7.6062\n",
            "Epoch 47/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 6.9897 - mae: 7.4805\n",
            "Epoch 48/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 6.8880 - mae: 7.3679\n",
            "Epoch 49/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 6.7770 - mae: 7.2633\n",
            "Epoch 50/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 6.6769 - mae: 7.1507\n",
            "Epoch 51/100\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 6.5502 - mae: 7.0284\n",
            "Epoch 52/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 6.3442 - mae: 6.8165\n",
            "Epoch 53/100\n",
            "94/94 [==============================] - 3s 29ms/step - loss: 5.8917 - mae: 6.3590\n",
            "Epoch 54/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 5.7040 - mae: 6.1773\n",
            "Epoch 55/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 5.4852 - mae: 5.9563\n",
            "Epoch 56/100\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 5.3039 - mae: 5.7832\n",
            "Epoch 57/100\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 5.1262 - mae: 5.5890\n",
            "Epoch 58/100\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 5.0326 - mae: 5.4962\n",
            "Epoch 59/100\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 4.8506 - mae: 5.3327\n",
            "Epoch 60/100\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 4.7293 - mae: 5.1932\n",
            "Epoch 61/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 4.6898 - mae: 5.1584\n",
            "Epoch 62/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 4.5983 - mae: 5.0650\n",
            "Epoch 63/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 4.5778 - mae: 5.0418\n",
            "Epoch 64/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 4.5373 - mae: 5.0002\n",
            "Epoch 65/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 4.4130 - mae: 4.8760\n",
            "Epoch 66/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 4.3343 - mae: 4.7969\n",
            "Epoch 67/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 4.2921 - mae: 4.7616\n",
            "Epoch 68/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 4.0665 - mae: 4.5317\n",
            "Epoch 69/100\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 3.9836 - mae: 4.4448\n",
            "Epoch 70/100\n",
            "94/94 [==============================] - 3s 28ms/step - loss: 4.0949 - mae: 4.5624\n",
            "Epoch 71/100\n",
            "94/94 [==============================] - 2s 27ms/step - loss: 3.8891 - mae: 4.3498\n",
            "Epoch 72/100\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 3.8217 - mae: 4.2836\n",
            "Epoch 73/100\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 3.7562 - mae: 4.2204\n",
            "Epoch 74/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 3.7745 - mae: 4.2371\n",
            "Epoch 75/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 3.5955 - mae: 4.0628\n",
            "Epoch 76/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 3.4692 - mae: 3.9326\n",
            "Epoch 77/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 3.3973 - mae: 3.8603\n",
            "Epoch 78/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 3.4243 - mae: 3.8863\n",
            "Epoch 79/100\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 3.2619 - mae: 3.7284\n",
            "Epoch 80/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 3.2285 - mae: 3.6944\n",
            "Epoch 81/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 3.2978 - mae: 3.7619\n",
            "Epoch 82/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 3.0761 - mae: 3.5356\n",
            "Epoch 83/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 3.0100 - mae: 3.4741\n",
            "Epoch 84/100\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 3.0222 - mae: 3.4879\n",
            "Epoch 85/100\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 3.0233 - mae: 3.4874\n",
            "Epoch 86/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 3.0414 - mae: 3.5078\n",
            "Epoch 87/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 3.0413 - mae: 3.5020\n",
            "Epoch 88/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 3.1343 - mae: 3.6007\n",
            "Epoch 89/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 2.8822 - mae: 3.3485\n",
            "Epoch 90/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 3.1834 - mae: 3.6550\n",
            "Epoch 91/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 3.1859 - mae: 3.6529\n",
            "Epoch 92/100\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 2.9201 - mae: 3.3816\n",
            "Epoch 93/100\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 3.1241 - mae: 3.5852\n",
            "Epoch 94/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 2.9063 - mae: 3.3682\n",
            "Epoch 95/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 2.9250 - mae: 3.3932\n",
            "Epoch 96/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 3.0122 - mae: 3.4718\n",
            "Epoch 97/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 2.8837 - mae: 3.3477\n",
            "Epoch 98/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 2.8811 - mae: 3.3422\n",
            "Epoch 99/100\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 2.8267 - mae: 3.2997\n",
            "Epoch 100/100\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 2.9102 - mae: 3.3770\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "AkBsrsXMzoWR",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 290
        },
        "outputId": "6b8bde31-29c3-4869-af3a-b3f9756e5cd0"
      },
      "source": [
        "plt.semilogx(history.history[\"lr\"], history.history[\"loss\"])\n",
        "plt.axis([1e-8, 1e-4, 0, 30])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(1e-08, 0.0001, 0.0, 30.0)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 7
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "4uh-97bpLZCA",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6966e471-5869-43df-bbf3-40224a7c8f62"
      },
      "source": [
        "tf.keras.backend.clear_session()\n",
        "tf.random.set_seed(51)\n",
        "np.random.seed(51)\n",
        "\n",
        "tf.keras.backend.clear_session()\n",
        "dataset = windowed_dataset(x_train, window_size, batch_size, shuffle_buffer_size)\n",
        "\n",
        "model = tf.keras.models.Sequential([\n",
        "  tf.keras.layers.Lambda(lambda x: tf.expand_dims(x, axis=-1),\n",
        "                      input_shape=[None]),\n",
        "   tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(32, return_sequences=True)),\n",
        "  tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(32)),\n",
        "  tf.keras.layers.Dense(1),\n",
        "  tf.keras.layers.Lambda(lambda x: x * 100.0)\n",
        "])\n",
        "\n",
        "\n",
        "model.compile(loss=\"mse\", optimizer=tf.keras.optimizers.SGD(lr=1e-5, momentum=0.9),metrics=[\"mae\"])\n",
        "history = model.fit(dataset,epochs=500,verbose=1)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/500\n",
            "94/94 [==============================] - 7s 76ms/step - loss: 317.3262 - mae: 10.6099\n",
            "Epoch 2/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 38.9353 - mae: 4.1943\n",
            "Epoch 3/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 33.3070 - mae: 3.9061\n",
            "Epoch 4/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 28.7480 - mae: 3.5994\n",
            "Epoch 5/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 31.5610 - mae: 3.8953\n",
            "Epoch 6/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 27.5148 - mae: 3.5512\n",
            "Epoch 7/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 26.3879 - mae: 3.4617\n",
            "Epoch 8/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 31.0663 - mae: 3.8562\n",
            "Epoch 9/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 28.1705 - mae: 3.6017\n",
            "Epoch 10/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 29.6219 - mae: 3.8047\n",
            "Epoch 11/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 28.1569 - mae: 3.6431\n",
            "Epoch 12/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 28.6770 - mae: 3.7158\n",
            "Epoch 13/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 25.2688 - mae: 3.3899\n",
            "Epoch 14/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 28.0231 - mae: 3.6522\n",
            "Epoch 15/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 27.5249 - mae: 3.6053\n",
            "Epoch 16/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 26.1767 - mae: 3.4643\n",
            "Epoch 17/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 28.1173 - mae: 3.6475\n",
            "Epoch 18/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 24.3173 - mae: 3.3000\n",
            "Epoch 19/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 26.9995 - mae: 3.5481\n",
            "Epoch 20/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 24.9156 - mae: 3.3942\n",
            "Epoch 21/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 26.0637 - mae: 3.4917\n",
            "Epoch 22/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 26.2883 - mae: 3.4825\n",
            "Epoch 23/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 24.2487 - mae: 3.2791\n",
            "Epoch 24/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 27.2745 - mae: 3.5988\n",
            "Epoch 25/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 25.8298 - mae: 3.4867\n",
            "Epoch 26/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 25.1074 - mae: 3.4117\n",
            "Epoch 27/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 25.7197 - mae: 3.4523\n",
            "Epoch 28/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 28.6707 - mae: 3.6757\n",
            "Epoch 29/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 26.9537 - mae: 3.5873\n",
            "Epoch 30/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 23.8332 - mae: 3.3018\n",
            "Epoch 31/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 23.7705 - mae: 3.2718\n",
            "Epoch 32/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 25.7472 - mae: 3.4593\n",
            "Epoch 33/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 29.6355 - mae: 3.9179\n",
            "Epoch 34/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 26.4020 - mae: 3.5818\n",
            "Epoch 35/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 27.4128 - mae: 3.6328\n",
            "Epoch 36/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 26.2063 - mae: 3.5773\n",
            "Epoch 37/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 26.7745 - mae: 3.5416\n",
            "Epoch 38/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 24.3039 - mae: 3.3331\n",
            "Epoch 39/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 26.0108 - mae: 3.5462\n",
            "Epoch 40/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 25.8497 - mae: 3.4804\n",
            "Epoch 41/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 26.2726 - mae: 3.4989\n",
            "Epoch 42/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 24.2589 - mae: 3.3588\n",
            "Epoch 43/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 25.3682 - mae: 3.4576\n",
            "Epoch 44/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 24.0836 - mae: 3.3332\n",
            "Epoch 45/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 23.5319 - mae: 3.2813\n",
            "Epoch 46/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 26.7626 - mae: 3.5958\n",
            "Epoch 47/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 25.5302 - mae: 3.4779\n",
            "Epoch 48/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 27.7386 - mae: 3.6840\n",
            "Epoch 49/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 25.3694 - mae: 3.4737\n",
            "Epoch 50/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 24.5315 - mae: 3.3462\n",
            "Epoch 51/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 23.5464 - mae: 3.3042\n",
            "Epoch 52/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 24.3952 - mae: 3.3477\n",
            "Epoch 53/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 24.8206 - mae: 3.4125\n",
            "Epoch 54/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 24.9364 - mae: 3.4056\n",
            "Epoch 55/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 23.2811 - mae: 3.2436\n",
            "Epoch 56/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 23.2605 - mae: 3.2389\n",
            "Epoch 57/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 23.5210 - mae: 3.2939\n",
            "Epoch 58/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 23.6461 - mae: 3.2941\n",
            "Epoch 59/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 24.1542 - mae: 3.3858\n",
            "Epoch 60/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 23.3212 - mae: 3.3041\n",
            "Epoch 61/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 23.5337 - mae: 3.2671\n",
            "Epoch 62/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 24.0404 - mae: 3.3418\n",
            "Epoch 63/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 23.8666 - mae: 3.3399\n",
            "Epoch 64/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 24.0127 - mae: 3.3100\n",
            "Epoch 65/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 24.0886 - mae: 3.3507\n",
            "Epoch 66/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 23.1626 - mae: 3.2586\n",
            "Epoch 67/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 24.5026 - mae: 3.3616\n",
            "Epoch 68/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 22.6044 - mae: 3.1878\n",
            "Epoch 69/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 23.4894 - mae: 3.2544\n",
            "Epoch 70/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 24.4824 - mae: 3.4014\n",
            "Epoch 71/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 22.7573 - mae: 3.2208\n",
            "Epoch 72/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 22.7916 - mae: 3.2219\n",
            "Epoch 73/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 22.8973 - mae: 3.2489\n",
            "Epoch 74/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 23.4484 - mae: 3.2801\n",
            "Epoch 75/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 23.7248 - mae: 3.2865\n",
            "Epoch 76/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 23.2224 - mae: 3.2506\n",
            "Epoch 77/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 23.4686 - mae: 3.2887\n",
            "Epoch 78/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 24.3428 - mae: 3.3854\n",
            "Epoch 79/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 22.0390 - mae: 3.1385\n",
            "Epoch 80/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 22.7838 - mae: 3.2032\n",
            "Epoch 81/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 22.6706 - mae: 3.2023\n",
            "Epoch 82/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 23.1920 - mae: 3.2569\n",
            "Epoch 83/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 23.9348 - mae: 3.3663\n",
            "Epoch 84/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 25.2038 - mae: 3.5488\n",
            "Epoch 85/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 22.9530 - mae: 3.2692\n",
            "Epoch 86/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.7926 - mae: 3.1244\n",
            "Epoch 87/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 22.6918 - mae: 3.1994\n",
            "Epoch 88/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 24.2488 - mae: 3.4070\n",
            "Epoch 89/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.9653 - mae: 3.1387\n",
            "Epoch 90/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 23.3707 - mae: 3.2741\n",
            "Epoch 91/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 22.7227 - mae: 3.2437\n",
            "Epoch 92/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 23.0548 - mae: 3.2601\n",
            "Epoch 93/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 23.2359 - mae: 3.2594\n",
            "Epoch 94/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 22.1729 - mae: 3.1581\n",
            "Epoch 95/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 22.6759 - mae: 3.2267\n",
            "Epoch 96/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 22.4012 - mae: 3.1823\n",
            "Epoch 97/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 22.6928 - mae: 3.2488\n",
            "Epoch 98/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 22.1256 - mae: 3.1941\n",
            "Epoch 99/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 23.1751 - mae: 3.3251\n",
            "Epoch 100/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 25.8207 - mae: 3.5944\n",
            "Epoch 101/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 22.5206 - mae: 3.2518\n",
            "Epoch 102/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 23.1137 - mae: 3.3012\n",
            "Epoch 103/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 23.0140 - mae: 3.2771\n",
            "Epoch 104/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 22.0115 - mae: 3.1517\n",
            "Epoch 105/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.9506 - mae: 3.2147\n",
            "Epoch 106/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 28.9291 - mae: 3.8518\n",
            "Epoch 107/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 23.9582 - mae: 3.3804\n",
            "Epoch 108/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.9490 - mae: 3.1499\n",
            "Epoch 109/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.6829 - mae: 3.1226\n",
            "Epoch 110/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 23.2419 - mae: 3.3075\n",
            "Epoch 111/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 22.7227 - mae: 3.2563\n",
            "Epoch 112/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.5312 - mae: 3.1077\n",
            "Epoch 113/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 22.4939 - mae: 3.2141\n",
            "Epoch 114/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.6366 - mae: 3.1183\n",
            "Epoch 115/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.7408 - mae: 3.1753\n",
            "Epoch 116/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.8358 - mae: 3.1441\n",
            "Epoch 117/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 22.0356 - mae: 3.1609\n",
            "Epoch 118/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.6947 - mae: 3.1387\n",
            "Epoch 119/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 22.0938 - mae: 3.1944\n",
            "Epoch 120/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.7132 - mae: 3.1478\n",
            "Epoch 121/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 21.8184 - mae: 3.1377\n",
            "Epoch 122/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 22.4340 - mae: 3.2195\n",
            "Epoch 123/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.3128 - mae: 3.1139\n",
            "Epoch 124/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 22.1348 - mae: 3.2031\n",
            "Epoch 125/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.4906 - mae: 3.1660\n",
            "Epoch 126/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 23.4230 - mae: 3.3704\n",
            "Epoch 127/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 22.0998 - mae: 3.1835\n",
            "Epoch 128/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 22.1304 - mae: 3.1994\n",
            "Epoch 129/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 22.5770 - mae: 3.2457\n",
            "Epoch 130/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 22.1173 - mae: 3.1964\n",
            "Epoch 131/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 22.3328 - mae: 3.2105\n",
            "Epoch 132/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.1190 - mae: 3.1008\n",
            "Epoch 133/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.6184 - mae: 3.0349\n",
            "Epoch 134/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.6953 - mae: 3.0553\n",
            "Epoch 135/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.2022 - mae: 3.1293\n",
            "Epoch 136/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 24.0297 - mae: 3.3659\n",
            "Epoch 137/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.7213 - mae: 3.1417\n",
            "Epoch 138/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.3158 - mae: 3.1353\n",
            "Epoch 139/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.7193 - mae: 3.1672\n",
            "Epoch 140/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.5687 - mae: 3.0914\n",
            "Epoch 141/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 25.4408 - mae: 3.5182\n",
            "Epoch 142/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.1031 - mae: 3.0611\n",
            "Epoch 143/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.3866 - mae: 3.0218\n",
            "Epoch 144/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.6398 - mae: 3.1484\n",
            "Epoch 145/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.2980 - mae: 3.0999\n",
            "Epoch 146/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 22.4927 - mae: 3.2549\n",
            "Epoch 147/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.4735 - mae: 3.1107\n",
            "Epoch 148/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.1820 - mae: 3.1274\n",
            "Epoch 149/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 22.2475 - mae: 3.2606\n",
            "Epoch 150/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 23.2213 - mae: 3.3480\n",
            "Epoch 151/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.8757 - mae: 3.2155\n",
            "Epoch 152/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 22.1125 - mae: 3.1725\n",
            "Epoch 153/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.4655 - mae: 3.1033\n",
            "Epoch 154/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.0159 - mae: 3.1034\n",
            "Epoch 155/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 23.3581 - mae: 3.3188\n",
            "Epoch 156/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.6520 - mae: 3.0547\n",
            "Epoch 157/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 23.5356 - mae: 3.3751\n",
            "Epoch 158/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.5494 - mae: 3.0337\n",
            "Epoch 159/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.3420 - mae: 3.1373\n",
            "Epoch 160/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.8242 - mae: 3.2140\n",
            "Epoch 161/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.6051 - mae: 3.0162\n",
            "Epoch 162/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 22.5058 - mae: 3.2159\n",
            "Epoch 163/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.8006 - mae: 3.0535\n",
            "Epoch 164/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.9514 - mae: 3.0550\n",
            "Epoch 165/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 20.6775 - mae: 3.1122\n",
            "Epoch 166/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 22.3590 - mae: 3.2513\n",
            "Epoch 167/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.9460 - mae: 3.0751\n",
            "Epoch 168/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.7825 - mae: 3.0795\n",
            "Epoch 169/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.3080 - mae: 3.1363\n",
            "Epoch 170/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.8807 - mae: 3.0765\n",
            "Epoch 171/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.2983 - mae: 3.0223\n",
            "Epoch 172/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.8468 - mae: 3.0938\n",
            "Epoch 173/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.3355 - mae: 3.0033\n",
            "Epoch 174/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.4376 - mae: 3.0208\n",
            "Epoch 175/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.7841 - mae: 3.0750\n",
            "Epoch 176/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.7601 - mae: 3.0844\n",
            "Epoch 177/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.9556 - mae: 3.0946\n",
            "Epoch 178/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.4968 - mae: 3.0316\n",
            "Epoch 179/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.3949 - mae: 3.1202\n",
            "Epoch 180/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.6790 - mae: 3.0304\n",
            "Epoch 181/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.6405 - mae: 3.1429\n",
            "Epoch 182/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.2671 - mae: 3.0019\n",
            "Epoch 183/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.3943 - mae: 3.0963\n",
            "Epoch 184/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.7635 - mae: 3.0593\n",
            "Epoch 185/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.6261 - mae: 3.0272\n",
            "Epoch 186/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.3650 - mae: 3.1327\n",
            "Epoch 187/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.5510 - mae: 3.1309\n",
            "Epoch 188/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.5782 - mae: 3.0383\n",
            "Epoch 189/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.1858 - mae: 3.0992\n",
            "Epoch 190/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.7974 - mae: 3.0597\n",
            "Epoch 191/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.4188 - mae: 3.0288\n",
            "Epoch 192/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.4539 - mae: 3.0496\n",
            "Epoch 193/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 20.6269 - mae: 3.0476\n",
            "Epoch 194/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.5287 - mae: 3.1358\n",
            "Epoch 195/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 22.1199 - mae: 3.2089\n",
            "Epoch 196/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.6305 - mae: 3.0440\n",
            "Epoch 197/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.9845 - mae: 3.0994\n",
            "Epoch 198/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.9725 - mae: 3.0931\n",
            "Epoch 199/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.4455 - mae: 3.0163\n",
            "Epoch 200/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.7709 - mae: 3.1634\n",
            "Epoch 201/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.2730 - mae: 3.1014\n",
            "Epoch 202/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.4564 - mae: 3.1365\n",
            "Epoch 203/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.1658 - mae: 3.0935\n",
            "Epoch 204/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.9558 - mae: 3.0744\n",
            "Epoch 205/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.0234 - mae: 3.0957\n",
            "Epoch 206/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.2463 - mae: 2.9921\n",
            "Epoch 207/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.3708 - mae: 3.0117\n",
            "Epoch 208/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.4963 - mae: 3.0602\n",
            "Epoch 209/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.1249 - mae: 2.9802\n",
            "Epoch 210/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.2293 - mae: 2.9902\n",
            "Epoch 211/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.7548 - mae: 3.0961\n",
            "Epoch 212/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.8743 - mae: 3.2094\n",
            "Epoch 213/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.9042 - mae: 2.9957\n",
            "Epoch 214/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.0134 - mae: 2.9680\n",
            "Epoch 215/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.7670 - mae: 2.9282\n",
            "Epoch 216/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.6250 - mae: 2.9770\n",
            "Epoch 217/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 23.5419 - mae: 3.3008\n",
            "Epoch 218/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.3440 - mae: 3.0160\n",
            "Epoch 219/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.5820 - mae: 3.1639\n",
            "Epoch 220/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.1670 - mae: 3.0856\n",
            "Epoch 221/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.2020 - mae: 2.9986\n",
            "Epoch 222/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.5870 - mae: 3.0402\n",
            "Epoch 223/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.4692 - mae: 3.0013\n",
            "Epoch 224/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.9753 - mae: 2.9871\n",
            "Epoch 225/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.4182 - mae: 3.1186\n",
            "Epoch 226/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.5142 - mae: 3.0263\n",
            "Epoch 227/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.5402 - mae: 3.1369\n",
            "Epoch 228/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.3337 - mae: 3.0916\n",
            "Epoch 229/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.9939 - mae: 3.0798\n",
            "Epoch 230/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.5796 - mae: 3.0547\n",
            "Epoch 231/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 21.4227 - mae: 3.1526\n",
            "Epoch 232/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.9232 - mae: 2.9575\n",
            "Epoch 233/500\n",
            "94/94 [==============================] - 2s 27ms/step - loss: 20.7319 - mae: 3.0568\n",
            "Epoch 234/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.9026 - mae: 3.0983\n",
            "Epoch 235/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.7802 - mae: 3.1628\n",
            "Epoch 236/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 20.4386 - mae: 3.0064\n",
            "Epoch 237/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 20.8139 - mae: 3.0406\n",
            "Epoch 238/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 22.1026 - mae: 3.2343\n",
            "Epoch 239/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.4574 - mae: 3.0128\n",
            "Epoch 240/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.1084 - mae: 2.9949\n",
            "Epoch 241/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.3228 - mae: 3.0385\n",
            "Epoch 242/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.9561 - mae: 3.0892\n",
            "Epoch 243/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.8183 - mae: 3.2058\n",
            "Epoch 244/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.4473 - mae: 3.1112\n",
            "Epoch 245/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.9347 - mae: 3.0992\n",
            "Epoch 246/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.9679 - mae: 3.0822\n",
            "Epoch 247/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.1608 - mae: 3.1124\n",
            "Epoch 248/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.9588 - mae: 2.9625\n",
            "Epoch 249/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.3442 - mae: 3.0978\n",
            "Epoch 250/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.9911 - mae: 2.9634\n",
            "Epoch 251/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.2473 - mae: 3.0014\n",
            "Epoch 252/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.0059 - mae: 3.1064\n",
            "Epoch 253/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.1367 - mae: 3.0715\n",
            "Epoch 254/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.1650 - mae: 2.9903\n",
            "Epoch 255/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.0247 - mae: 2.9854\n",
            "Epoch 256/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.6701 - mae: 3.0651\n",
            "Epoch 257/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.5273 - mae: 3.0462\n",
            "Epoch 258/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.5064 - mae: 3.0352\n",
            "Epoch 259/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.8452 - mae: 2.9619\n",
            "Epoch 260/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.8976 - mae: 3.0915\n",
            "Epoch 261/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 20.6571 - mae: 3.0471\n",
            "Epoch 262/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.1181 - mae: 2.9870\n",
            "Epoch 263/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.8170 - mae: 3.0774\n",
            "Epoch 264/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.0873 - mae: 2.9974\n",
            "Epoch 265/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.5525 - mae: 3.0694\n",
            "Epoch 266/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.0156 - mae: 3.0821\n",
            "Epoch 267/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.1246 - mae: 3.0207\n",
            "Epoch 268/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.9991 - mae: 2.9717\n",
            "Epoch 269/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.1857 - mae: 3.0961\n",
            "Epoch 270/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.6114 - mae: 3.0436\n",
            "Epoch 271/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 22.1501 - mae: 3.1970\n",
            "Epoch 272/500\n",
            "94/94 [==============================] - 2s 23ms/step - loss: 20.4819 - mae: 3.0496\n",
            "Epoch 273/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.9794 - mae: 3.1880\n",
            "Epoch 274/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.8582 - mae: 3.0861\n",
            "Epoch 275/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.0370 - mae: 3.0983\n",
            "Epoch 276/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.0262 - mae: 2.9829\n",
            "Epoch 277/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.0799 - mae: 2.9702\n",
            "Epoch 278/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.1167 - mae: 3.0781\n",
            "Epoch 279/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.0077 - mae: 2.9666\n",
            "Epoch 280/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.0625 - mae: 2.9747\n",
            "Epoch 281/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.5950 - mae: 2.9224\n",
            "Epoch 282/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.8707 - mae: 3.1632\n",
            "Epoch 283/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.8121 - mae: 2.9586\n",
            "Epoch 284/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.3143 - mae: 3.1212\n",
            "Epoch 285/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.5860 - mae: 3.1574\n",
            "Epoch 286/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.7951 - mae: 2.9641\n",
            "Epoch 287/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.5888 - mae: 3.0348\n",
            "Epoch 288/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.7494 - mae: 3.0653\n",
            "Epoch 289/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.5110 - mae: 3.0347\n",
            "Epoch 290/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.8093 - mae: 3.0720\n",
            "Epoch 291/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.7507 - mae: 2.9558\n",
            "Epoch 292/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.0639 - mae: 2.9668\n",
            "Epoch 293/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.4937 - mae: 3.0395\n",
            "Epoch 294/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.6059 - mae: 3.0069\n",
            "Epoch 295/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.1725 - mae: 2.9974\n",
            "Epoch 296/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.6574 - mae: 2.9404\n",
            "Epoch 297/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.7072 - mae: 2.9829\n",
            "Epoch 298/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.3184 - mae: 3.0962\n",
            "Epoch 299/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.0535 - mae: 2.9906\n",
            "Epoch 300/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.7955 - mae: 3.0955\n",
            "Epoch 301/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.2746 - mae: 3.0123\n",
            "Epoch 302/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 19.8877 - mae: 2.9711\n",
            "Epoch 303/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.3241 - mae: 3.0003\n",
            "Epoch 304/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.2663 - mae: 3.0076\n",
            "Epoch 305/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.7522 - mae: 3.0729\n",
            "Epoch 306/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.7437 - mae: 3.1586\n",
            "Epoch 307/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.6857 - mae: 2.9797\n",
            "Epoch 308/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.2868 - mae: 3.0173\n",
            "Epoch 309/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.3782 - mae: 3.0052\n",
            "Epoch 310/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.3229 - mae: 3.0155\n",
            "Epoch 311/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.9761 - mae: 3.0868\n",
            "Epoch 312/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.4400 - mae: 2.9714\n",
            "Epoch 313/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.0927 - mae: 3.1223\n",
            "Epoch 314/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.7753 - mae: 2.9488\n",
            "Epoch 315/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.6573 - mae: 3.0313\n",
            "Epoch 316/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.9878 - mae: 2.9977\n",
            "Epoch 317/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.7621 - mae: 3.0699\n",
            "Epoch 318/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.9450 - mae: 3.1137\n",
            "Epoch 319/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.1431 - mae: 2.9906\n",
            "Epoch 320/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.2659 - mae: 3.0402\n",
            "Epoch 321/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 23.2152 - mae: 3.3197\n",
            "Epoch 322/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.1233 - mae: 2.9775\n",
            "Epoch 323/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.9473 - mae: 2.9661\n",
            "Epoch 324/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.4463 - mae: 3.0495\n",
            "Epoch 325/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 22.2523 - mae: 3.2171\n",
            "Epoch 326/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.0040 - mae: 2.9922\n",
            "Epoch 327/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.8223 - mae: 2.9515\n",
            "Epoch 328/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.7517 - mae: 2.9564\n",
            "Epoch 329/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.6749 - mae: 3.0606\n",
            "Epoch 330/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 22.6844 - mae: 3.2479\n",
            "Epoch 331/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 20.3591 - mae: 3.0185\n",
            "Epoch 332/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.1604 - mae: 2.9966\n",
            "Epoch 333/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.0697 - mae: 2.9834\n",
            "Epoch 334/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.6091 - mae: 3.0269\n",
            "Epoch 335/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.9800 - mae: 3.0924\n",
            "Epoch 336/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.0807 - mae: 3.0260\n",
            "Epoch 337/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.1579 - mae: 3.0029\n",
            "Epoch 338/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.1274 - mae: 2.9751\n",
            "Epoch 339/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.6075 - mae: 2.9169\n",
            "Epoch 340/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.3532 - mae: 3.0045\n",
            "Epoch 341/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.5838 - mae: 3.0369\n",
            "Epoch 342/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.7253 - mae: 2.9437\n",
            "Epoch 343/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.5685 - mae: 2.9311\n",
            "Epoch 344/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.8505 - mae: 3.0955\n",
            "Epoch 345/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.1594 - mae: 2.9807\n",
            "Epoch 346/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 20.4039 - mae: 3.0039\n",
            "Epoch 347/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.5280 - mae: 3.0342\n",
            "Epoch 348/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.0031 - mae: 2.9744\n",
            "Epoch 349/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.6931 - mae: 2.9667\n",
            "Epoch 350/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.0636 - mae: 2.9838\n",
            "Epoch 351/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.3943 - mae: 3.0069\n",
            "Epoch 352/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.0978 - mae: 2.9871\n",
            "Epoch 353/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.7076 - mae: 3.0415\n",
            "Epoch 354/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.1977 - mae: 2.9892\n",
            "Epoch 355/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.7449 - mae: 2.9589\n",
            "Epoch 356/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.3405 - mae: 3.0127\n",
            "Epoch 357/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.2003 - mae: 2.9747\n",
            "Epoch 358/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.8145 - mae: 3.1822\n",
            "Epoch 359/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.3291 - mae: 3.0428\n",
            "Epoch 360/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.9978 - mae: 2.9913\n",
            "Epoch 361/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 20.0509 - mae: 2.9850\n",
            "Epoch 362/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.4319 - mae: 3.0322\n",
            "Epoch 363/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.4889 - mae: 3.0416\n",
            "Epoch 364/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.8475 - mae: 3.0573\n",
            "Epoch 365/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.1224 - mae: 2.9910\n",
            "Epoch 366/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.0362 - mae: 2.9773\n",
            "Epoch 367/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.5231 - mae: 3.0469\n",
            "Epoch 368/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.4225 - mae: 3.1346\n",
            "Epoch 369/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.6891 - mae: 3.0480\n",
            "Epoch 370/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 19.5874 - mae: 2.9596\n",
            "Epoch 371/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 20.2155 - mae: 2.9779\n",
            "Epoch 372/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 20.1343 - mae: 2.9670\n",
            "Epoch 373/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 19.7456 - mae: 2.9390\n",
            "Epoch 374/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.2170 - mae: 2.9912\n",
            "Epoch 375/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.3648 - mae: 2.9194\n",
            "Epoch 376/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.6395 - mae: 3.0551\n",
            "Epoch 377/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.1821 - mae: 3.0078\n",
            "Epoch 378/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.7059 - mae: 3.0516\n",
            "Epoch 379/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.2584 - mae: 3.1327\n",
            "Epoch 380/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.2095 - mae: 3.0523\n",
            "Epoch 381/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 22.7124 - mae: 3.2430\n",
            "Epoch 382/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.8180 - mae: 2.9749\n",
            "Epoch 383/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.3940 - mae: 2.9036\n",
            "Epoch 384/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 19.9724 - mae: 2.9906\n",
            "Epoch 385/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.9418 - mae: 2.9691\n",
            "Epoch 386/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.5704 - mae: 3.0416\n",
            "Epoch 387/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.5054 - mae: 2.9368\n",
            "Epoch 388/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.4588 - mae: 3.0639\n",
            "Epoch 389/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.4305 - mae: 3.0004\n",
            "Epoch 390/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.1134 - mae: 3.1233\n",
            "Epoch 391/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.5124 - mae: 3.0642\n",
            "Epoch 392/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.1995 - mae: 2.9901\n",
            "Epoch 393/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.9790 - mae: 2.9632\n",
            "Epoch 394/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.8761 - mae: 2.9510\n",
            "Epoch 395/500\n",
            "94/94 [==============================] - 4s 38ms/step - loss: 19.9667 - mae: 2.9819\n",
            "Epoch 396/500\n",
            "94/94 [==============================] - 4s 41ms/step - loss: 19.6282 - mae: 2.9364\n",
            "Epoch 397/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.7105 - mae: 2.9394\n",
            "Epoch 398/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.4895 - mae: 3.0680\n",
            "Epoch 399/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 18.9561 - mae: 2.9555\n",
            "Epoch 400/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.7878 - mae: 3.1487\n",
            "Epoch 401/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.1668 - mae: 3.0005\n",
            "Epoch 402/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.1599 - mae: 2.9938\n",
            "Epoch 403/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.3992 - mae: 2.9283\n",
            "Epoch 404/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.9351 - mae: 2.9624\n",
            "Epoch 405/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.3716 - mae: 2.8987\n",
            "Epoch 406/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 19.6847 - mae: 2.9523\n",
            "Epoch 407/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.0284 - mae: 2.9809\n",
            "Epoch 408/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.6314 - mae: 2.9449\n",
            "Epoch 409/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 19.7809 - mae: 2.9497\n",
            "Epoch 410/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 20.8222 - mae: 3.0768\n",
            "Epoch 411/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 20.1611 - mae: 3.0151\n",
            "Epoch 412/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.5802 - mae: 2.9505\n",
            "Epoch 413/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.3051 - mae: 3.0027\n",
            "Epoch 414/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.4535 - mae: 3.0386\n",
            "Epoch 415/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.3940 - mae: 3.0407\n",
            "Epoch 416/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.9308 - mae: 3.2072\n",
            "Epoch 417/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.9365 - mae: 2.9557\n",
            "Epoch 418/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.1650 - mae: 3.0012\n",
            "Epoch 419/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.6392 - mae: 2.9308\n",
            "Epoch 420/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.9350 - mae: 2.9668\n",
            "Epoch 421/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.7880 - mae: 2.9809\n",
            "Epoch 422/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.5248 - mae: 3.0400\n",
            "Epoch 423/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 19.6624 - mae: 2.9492\n",
            "Epoch 424/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.8170 - mae: 2.9634\n",
            "Epoch 425/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.5200 - mae: 3.0568\n",
            "Epoch 426/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.2943 - mae: 3.0081\n",
            "Epoch 427/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.4329 - mae: 2.9089\n",
            "Epoch 428/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.0031 - mae: 3.0879\n",
            "Epoch 429/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.2806 - mae: 3.0220\n",
            "Epoch 430/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.5765 - mae: 2.9404\n",
            "Epoch 431/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.5960 - mae: 3.0344\n",
            "Epoch 432/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.0589 - mae: 2.9759\n",
            "Epoch 433/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.4302 - mae: 2.9241\n",
            "Epoch 434/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.4208 - mae: 3.0207\n",
            "Epoch 435/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.4783 - mae: 3.0445\n",
            "Epoch 436/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 23.0460 - mae: 3.3204\n",
            "Epoch 437/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.7113 - mae: 2.9456\n",
            "Epoch 438/500\n",
            "94/94 [==============================] - 2s 26ms/step - loss: 19.6755 - mae: 2.9465\n",
            "Epoch 439/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.2982 - mae: 2.9931\n",
            "Epoch 440/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.9068 - mae: 2.9502\n",
            "Epoch 441/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.2839 - mae: 2.8932\n",
            "Epoch 442/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.3751 - mae: 3.0006\n",
            "Epoch 443/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.8329 - mae: 2.9633\n",
            "Epoch 444/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.5714 - mae: 3.0522\n",
            "Epoch 445/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.1309 - mae: 2.9926\n",
            "Epoch 446/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.4698 - mae: 3.0502\n",
            "Epoch 447/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.2180 - mae: 3.1112\n",
            "Epoch 448/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.5009 - mae: 3.0380\n",
            "Epoch 449/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.9612 - mae: 3.0832\n",
            "Epoch 450/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.6572 - mae: 2.9375\n",
            "Epoch 451/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.7883 - mae: 2.9536\n",
            "Epoch 452/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.4756 - mae: 2.9109\n",
            "Epoch 453/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.2796 - mae: 3.0008\n",
            "Epoch 454/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.8036 - mae: 2.9661\n",
            "Epoch 455/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.9833 - mae: 3.1258\n",
            "Epoch 456/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.6702 - mae: 3.0428\n",
            "Epoch 457/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.8740 - mae: 2.9656\n",
            "Epoch 458/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.1243 - mae: 3.0023\n",
            "Epoch 459/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.7668 - mae: 2.9614\n",
            "Epoch 460/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.9171 - mae: 2.9699\n",
            "Epoch 461/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.4608 - mae: 3.0644\n",
            "Epoch 462/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.6364 - mae: 3.0729\n",
            "Epoch 463/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.4527 - mae: 3.0453\n",
            "Epoch 464/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.5904 - mae: 2.9329\n",
            "Epoch 465/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.3475 - mae: 3.0158\n",
            "Epoch 466/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.3813 - mae: 2.9131\n",
            "Epoch 467/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.2274 - mae: 3.0118\n",
            "Epoch 468/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.8662 - mae: 2.9471\n",
            "Epoch 469/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.2975 - mae: 2.8973\n",
            "Epoch 470/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.3518 - mae: 3.0041\n",
            "Epoch 471/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.9352 - mae: 2.9440\n",
            "Epoch 472/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.3748 - mae: 2.9203\n",
            "Epoch 473/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.3659 - mae: 2.9104\n",
            "Epoch 474/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.2433 - mae: 3.0085\n",
            "Epoch 475/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 20.2581 - mae: 3.0174\n",
            "Epoch 476/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.0924 - mae: 2.8786\n",
            "Epoch 477/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.4185 - mae: 2.9080\n",
            "Epoch 478/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 21.6968 - mae: 3.1532\n",
            "Epoch 479/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.8846 - mae: 2.9644\n",
            "Epoch 480/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.9139 - mae: 2.9833\n",
            "Epoch 481/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.4863 - mae: 3.0580\n",
            "Epoch 482/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.8980 - mae: 2.9887\n",
            "Epoch 483/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.7683 - mae: 2.9632\n",
            "Epoch 484/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.5037 - mae: 2.9308\n",
            "Epoch 485/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.9209 - mae: 3.0835\n",
            "Epoch 486/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.2559 - mae: 2.9104\n",
            "Epoch 487/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.5458 - mae: 2.9312\n",
            "Epoch 488/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.0033 - mae: 2.9921\n",
            "Epoch 489/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.2788 - mae: 2.9162\n",
            "Epoch 490/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.5565 - mae: 2.9309\n",
            "Epoch 491/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.8556 - mae: 2.9521\n",
            "Epoch 492/500\n",
            "94/94 [==============================] - 2s 25ms/step - loss: 19.9278 - mae: 2.9470\n",
            "Epoch 493/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 21.4225 - mae: 3.1233\n",
            "Epoch 494/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.9712 - mae: 2.9734\n",
            "Epoch 495/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.3941 - mae: 3.0106\n",
            "Epoch 496/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.2896 - mae: 2.9089\n",
            "Epoch 497/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.8270 - mae: 2.9603\n",
            "Epoch 498/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.5433 - mae: 2.9342\n",
            "Epoch 499/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 19.8258 - mae: 2.9413\n",
            "Epoch 500/500\n",
            "94/94 [==============================] - 2s 24ms/step - loss: 20.8011 - mae: 3.0503\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "icGDaND7z0ne",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 388
        },
        "outputId": "85d467a1-fcd2-47ee-e604-99bb321e0fb8"
      },
      "source": [
        "forecast = []\n",
        "results = []\n",
        "for time in range(len(series) - window_size):\n",
        "  forecast.append(model.predict(series[time:time + window_size][np.newaxis]))\n",
        "\n",
        "forecast = forecast[split_time-window_size:]\n",
        "results = np.array(forecast)[:, 0, 0]\n",
        "\n",
        "\n",
        "plt.figure(figsize=(10, 6))\n",
        "\n",
        "plot_series(time_valid, x_valid)\n",
        "plot_series(time_valid, results)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "KfPeqI7rz4LD",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "a9ab16e8-63c1-4cd7-b1a9-ee7ae4661e76"
      },
      "source": [
        "tf.keras.metrics.mean_absolute_error(x_valid, results).numpy()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "3.3967555"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "JUsdZB_tzDLe",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 607
        },
        "outputId": "76800831-5920-4305-edf6-a3ca09613ea4"
      },
      "source": [
        "import matplotlib.image  as mpimg\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "#-----------------------------------------------------------\n",
        "# Retrieve a list of list results on training and test data\n",
        "# sets for each training epoch\n",
        "#-----------------------------------------------------------\n",
        "mae=history.history['mae']\n",
        "loss=history.history['loss']\n",
        "\n",
        "epochs=range(len(loss)) # Get number of epochs\n",
        "\n",
        "#------------------------------------------------\n",
        "# Plot MAE and Loss\n",
        "#------------------------------------------------\n",
        "plt.plot(epochs, mae, 'r')\n",
        "plt.plot(epochs, loss, 'b')\n",
        "plt.title('MAE and Loss')\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.legend([\"MAE\", \"Loss\"])\n",
        "\n",
        "plt.figure()\n",
        "\n",
        "epochs_zoom = epochs[200:]\n",
        "mae_zoom = mae[200:]\n",
        "loss_zoom = loss[200:]\n",
        "\n",
        "#------------------------------------------------\n",
        "# Plot Zoomed MAE and Loss\n",
        "#------------------------------------------------\n",
        "plt.plot(epochs_zoom, mae_zoom, 'r')\n",
        "plt.plot(epochs_zoom, loss_zoom, 'b')\n",
        "plt.title('MAE and Loss')\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.legend([\"MAE\", \"Loss\"])\n",
        "\n",
        "plt.figure()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 0 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 11
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 0 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    }
  ]
}