4. Sequences and Prediction/Week 3.1 - Finding an optimal learning rate for a RNN.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "S+P Week 3 Lesson 2 - RNN.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "D1J15Vh_1Jih",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 544
        },
        "outputId": "d2c4f689-6c44-416d-b5ca-cb64fa9bd808"
      },
      "source": [
        "!pip install tensorflow==2.0.0-beta1"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting tensorflow==2.0.0-beta1\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/29/6c/2c9a5c4d095c63c2fb37d20def0e4f92685f7aee9243d6aae25862694fd1/tensorflow-2.0.0b1-cp36-cp36m-manylinux1_x86_64.whl (87.9MB)\n",
            "\u001b[K     |████████████████████████████████| 87.9MB 1.3MB/s \n",
            "\u001b[?25hRequirement already satisfied: numpy<2.0,>=1.14.5 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0-beta1) (1.16.4)\n",
            "Requirement already satisfied: gast>=0.2.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0-beta1) (0.2.2)\n",
            "Requirement already satisfied: google-pasta>=0.1.6 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0-beta1) (0.1.7)\n",
            "Collecting tb-nightly<1.14.0a20190604,>=1.14.0a20190603 (from tensorflow==2.0.0-beta1)\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/a4/96/571b875cd81dda9d5dfa1422a4f9d749e67c0a8d4f4f0b33a4e5f5f35e27/tb_nightly-1.14.0a20190603-py3-none-any.whl (3.1MB)\n",
            "\u001b[K     |████████████████████████████████| 3.1MB 28.8MB/s \n",
            "\u001b[?25hRequirement already satisfied: astor>=0.6.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0-beta1) (0.8.0)\n",
            "Requirement already satisfied: six>=1.10.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0-beta1) (1.12.0)\n",
            "Requirement already satisfied: grpcio>=1.8.6 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0-beta1) (1.15.0)\n",
            "Requirement already satisfied: keras-applications>=1.0.6 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0-beta1) (1.0.8)\n",
            "Requirement already satisfied: protobuf>=3.6.1 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0-beta1) (3.7.1)\n",
            "Requirement already satisfied: termcolor>=1.1.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0-beta1) (1.1.0)\n",
            "Requirement already satisfied: keras-preprocessing>=1.0.5 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0-beta1) (1.1.0)\n",
            "Requirement already satisfied: wheel>=0.26 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0-beta1) (0.33.4)\n",
            "Requirement already satisfied: absl-py>=0.7.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0-beta1) (0.7.1)\n",
            "Requirement already satisfied: wrapt>=1.11.1 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0-beta1) (1.11.2)\n",
            "Collecting tf-estimator-nightly<1.14.0.dev2019060502,>=1.14.0.dev2019060501 (from tensorflow==2.0.0-beta1)\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/32/dd/99c47dd007dcf10d63fd895611b063732646f23059c618a373e85019eb0e/tf_estimator_nightly-1.14.0.dev2019060501-py2.py3-none-any.whl (496kB)\n",
            "\u001b[K     |████████████████████████████████| 501kB 44.5MB/s \n",
            "\u001b[?25hRequirement already satisfied: markdown>=2.6.8 in /usr/local/lib/python3.6/dist-packages (from tb-nightly<1.14.0a20190604,>=1.14.0a20190603->tensorflow==2.0.0-beta1) (3.1.1)\n",
            "Requirement already satisfied: werkzeug>=0.11.15 in /usr/local/lib/python3.6/dist-packages (from tb-nightly<1.14.0a20190604,>=1.14.0a20190603->tensorflow==2.0.0-beta1) (0.15.5)\n",
            "Requirement already satisfied: setuptools>=41.0.0 in /usr/local/lib/python3.6/dist-packages (from tb-nightly<1.14.0a20190604,>=1.14.0a20190603->tensorflow==2.0.0-beta1) (41.0.1)\n",
            "Requirement already satisfied: h5py in /usr/local/lib/python3.6/dist-packages (from keras-applications>=1.0.6->tensorflow==2.0.0-beta1) (2.8.0)\n",
            "Installing collected packages: tb-nightly, tf-estimator-nightly, tensorflow\n",
            "  Found existing installation: tensorflow 1.14.0\n",
            "    Uninstalling tensorflow-1.14.0:\n",
            "      Successfully uninstalled tensorflow-1.14.0\n",
            "Successfully installed tb-nightly-1.14.0a20190603 tensorflow-2.0.0b1 tf-estimator-nightly-1.14.0.dev2019060501\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BOjujz601HcS",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "34a3713c-6ddc-4653-e901-0844979604a7"
      },
      "source": [
        "import tensorflow as tf\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "print(tf.__version__)"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "2.0.0-beta1\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "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(True)\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 < 0.4,\n",
        "                    np.cos(season_time * 2 * np.pi),\n",
        "                    1 / np.exp(3 * 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(4 * 365 + 1, dtype=\"float32\")\n",
        "baseline = 10\n",
        "series = trend(time, 0.1)  \n",
        "baseline = 10\n",
        "amplitude = 40\n",
        "slope = 0.05\n",
        "noise_level = 5\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=42)\n",
        "\n",
        "split_time = 1000\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"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4sTTIOCbyShY",
        "colab_type": "code",
        "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": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eENErLuUrCa2",
        "colab_type": "text"
      },
      "source": [
        "## Let's try training a simple 2 layers RNN with different learning rates."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "L4nblWkqg1NL",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "774a1138-7c7d-4f3f-f682-50ead7291cd4"
      },
      "source": [
        "tf.keras.backend.clear_session()\n",
        "tf.random.set_seed(51)\n",
        "np.random.seed(51)\n",
        "\n",
        "train_set = windowed_dataset(x_train, window_size, batch_size=128, shuffle_buffer=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.SimpleRNN(40, return_sequences=True),\n",
        "  tf.keras.layers.SimpleRNN(40),\n",
        "  tf.keras.layers.Dense(1),\n",
        "  tf.keras.layers.Lambda(lambda x: x * 100.0)\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(train_set, epochs=100, callbacks=[lr_schedule])"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/100\n",
            "8/8 [==============================] - 2s 204ms/step - loss: 195.6164 - mae: 196.0720\n",
            "Epoch 2/100\n",
            "8/8 [==============================] - 0s 27ms/step - loss: 194.8099 - mae: 195.2815\n",
            "Epoch 3/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 193.5298 - mae: 194.0093\n",
            "Epoch 4/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 191.9202 - mae: 192.4067\n",
            "Epoch 5/100\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 190.0269 - mae: 190.5204\n",
            "Epoch 6/100\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 187.8482 - mae: 188.3494\n",
            "Epoch 7/100\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 185.3580 - mae: 185.8679\n",
            "Epoch 8/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 182.5147 - mae: 183.0343\n",
            "Epoch 9/100\n",
            "8/8 [==============================] - 0s 27ms/step - loss: 179.2627 - mae: 179.7937\n",
            "Epoch 10/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 175.5317 - mae: 176.0756\n",
            "Epoch 11/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 171.2309 - mae: 171.7901\n",
            "Epoch 12/100\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 166.2408 - mae: 166.8182\n",
            "Epoch 13/100\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 160.3970 - mae: 160.9967\n",
            "Epoch 14/100\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 153.4630 - mae: 154.0908\n",
            "Epoch 15/100\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 145.0824 - mae: 145.7472\n",
            "Epoch 16/100\n",
            "8/8 [==============================] - 0s 27ms/step - loss: 134.7148 - mae: 135.4294\n",
            "Epoch 17/100\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 121.5979 - mae: 122.3784\n",
            "Epoch 18/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 104.8339 - mae: 105.6960\n",
            "Epoch 19/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 83.4423 - mae: 84.4015\n",
            "Epoch 20/100\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 56.6354 - mae: 57.6680\n",
            "Epoch 21/100\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 31.8977 - mae: 32.5999\n",
            "Epoch 22/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 24.1295 - mae: 24.6321\n",
            "Epoch 23/100\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 23.0485 - mae: 23.5755\n",
            "Epoch 24/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 20.7063 - mae: 21.2764\n",
            "Epoch 25/100\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 16.5913 - mae: 17.1682\n",
            "Epoch 26/100\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 13.5437 - mae: 14.1215\n",
            "Epoch 27/100\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 10.4890 - mae: 11.0766\n",
            "Epoch 28/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 8.5560 - mae: 9.1199\n",
            "Epoch 29/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 7.8495 - mae: 8.4057\n",
            "Epoch 30/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 7.8732 - mae: 8.4259\n",
            "Epoch 31/100\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 7.7311 - mae: 8.2862\n",
            "Epoch 32/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 7.6639 - mae: 8.2168\n",
            "Epoch 33/100\n",
            "8/8 [==============================] - 0s 29ms/step - loss: 7.5269 - mae: 8.0793\n",
            "Epoch 34/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 7.4704 - mae: 8.0246\n",
            "Epoch 35/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 7.4114 - mae: 7.9635\n",
            "Epoch 36/100\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 7.3461 - mae: 7.8977\n",
            "Epoch 37/100\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 7.3054 - mae: 7.8591\n",
            "Epoch 38/100\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 7.2857 - mae: 7.8411\n",
            "Epoch 39/100\n",
            "8/8 [==============================] - 0s 28ms/step - loss: 7.2737 - mae: 7.8307\n",
            "Epoch 40/100\n",
            "8/8 [==============================] - 0s 27ms/step - loss: 7.2700 - mae: 7.8277\n",
            "Epoch 41/100\n",
            "8/8 [==============================] - 0s 27ms/step - loss: 7.2869 - mae: 7.8469\n",
            "Epoch 42/100\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 7.3024 - mae: 7.8680\n",
            "Epoch 43/100\n",
            "8/8 [==============================] - 0s 27ms/step - loss: 7.2949 - mae: 7.8550\n",
            "Epoch 44/100\n",
            "8/8 [==============================] - 0s 27ms/step - loss: 7.1785 - mae: 7.7332\n",
            "Epoch 45/100\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 6.9266 - mae: 7.4823\n",
            "Epoch 46/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 6.8034 - mae: 7.3491\n",
            "Epoch 47/100\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 6.6528 - mae: 7.2075\n",
            "Epoch 48/100\n",
            "8/8 [==============================] - 0s 27ms/step - loss: 6.5724 - mae: 7.1271\n",
            "Epoch 49/100\n",
            "8/8 [==============================] - 0s 27ms/step - loss: 6.5298 - mae: 7.0895\n",
            "Epoch 50/100\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 6.5122 - mae: 7.0751\n",
            "Epoch 51/100\n",
            "8/8 [==============================] - 0s 28ms/step - loss: 6.5237 - mae: 7.0877\n",
            "Epoch 52/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 6.4563 - mae: 7.0044\n",
            "Epoch 53/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 6.5617 - mae: 7.1326\n",
            "Epoch 54/100\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 6.5190 - mae: 7.0468\n",
            "Epoch 55/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 7.1115 - mae: 7.6400\n",
            "Epoch 56/100\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 6.9174 - mae: 7.4577\n",
            "Epoch 57/100\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 7.3977 - mae: 7.9785\n",
            "Epoch 58/100\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 8.0083 - mae: 8.4784\n",
            "Epoch 59/100\n",
            "8/8 [==============================] - 0s 27ms/step - loss: 7.3391 - mae: 7.9236\n",
            "Epoch 60/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 7.4138 - mae: 7.7832\n",
            "Epoch 61/100\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 11.8319 - mae: 12.5305\n",
            "Epoch 62/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 10.0028 - mae: 10.5460\n",
            "Epoch 63/100\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 7.3712 - mae: 7.9257\n",
            "Epoch 64/100\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 7.3975 - mae: 7.9864\n",
            "Epoch 65/100\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 6.6837 - mae: 7.2028\n",
            "Epoch 66/100\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 6.6753 - mae: 7.2361\n",
            "Epoch 67/100\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 6.6805 - mae: 7.2388\n",
            "Epoch 68/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 7.5913 - mae: 8.1123\n",
            "Epoch 69/100\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 7.2316 - mae: 7.8188\n",
            "Epoch 70/100\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 6.6379 - mae: 7.1672\n",
            "Epoch 71/100\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 7.0936 - mae: 7.5507\n",
            "Epoch 72/100\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 12.7947 - mae: 13.1296\n",
            "Epoch 73/100\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 14.6150 - mae: 15.2372\n",
            "Epoch 74/100\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 9.7449 - mae: 10.3792\n",
            "Epoch 75/100\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 17.4348 - mae: 17.7422\n",
            "Epoch 76/100\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 15.3940 - mae: 16.0708\n",
            "Epoch 77/100\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 12.7528 - mae: 13.3536\n",
            "Epoch 78/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 11.9877 - mae: 12.6148\n",
            "Epoch 79/100\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 13.3430 - mae: 13.4249\n",
            "Epoch 80/100\n",
            "8/8 [==============================] - 0s 28ms/step - loss: 19.7446 - mae: 20.0583\n",
            "Epoch 81/100\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 17.2401 - mae: 17.9492\n",
            "Epoch 82/100\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 13.5173 - mae: 13.8270\n",
            "Epoch 83/100\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 17.2959 - mae: 17.7906\n",
            "Epoch 84/100\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 16.3584 - mae: 16.9491\n",
            "Epoch 85/100\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 15.8124 - mae: 15.9968\n",
            "Epoch 86/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 18.1265 - mae: 18.3796\n",
            "Epoch 87/100\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 21.2566 - mae: 21.0903\n",
            "Epoch 88/100\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 28.0360 - mae: 28.8328\n",
            "Epoch 89/100\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 16.1386 - mae: 16.8259\n",
            "Epoch 90/100\n",
            "8/8 [==============================] - 0s 29ms/step - loss: 31.9562 - mae: 33.1344\n",
            "Epoch 91/100\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 38.5418 - mae: 39.4129\n",
            "Epoch 92/100\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 37.0045 - mae: 37.8386\n",
            "Epoch 93/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 16.3824 - mae: 16.9966\n",
            "Epoch 94/100\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 41.7508 - mae: 42.1462\n",
            "Epoch 95/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 70.8838 - mae: 71.4910\n",
            "Epoch 96/100\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 75.2327 - mae: 75.2848\n",
            "Epoch 97/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 69.7161 - mae: 69.5044\n",
            "Epoch 98/100\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 69.4055 - mae: 69.3081\n",
            "Epoch 99/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 95.2281 - mae: 95.0477\n",
            "Epoch 100/100\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 49.8335 - mae: 50.2340\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5He3pp-Hj758",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 290
        },
        "outputId": "980a4a70-5fd8-413e-aa1f-ebc56a8b20fc"
      },
      "source": [
        "plt.semilogx(history.history[\"lr\"], history.history[\"loss\"])\n",
        "plt.axis([1e-8, 1e-4, 0, 30])"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[1e-08, 0.0001, 0, 30]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 15
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAEACAYAAABfxaZOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dgeodOuerdEE",
        "colab_type": "text"
      },
      "source": [
        "## 5*10e-5 seems to be a reasonable learning rate, hence it gets chosen."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6y1KMowRkHkC",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "eaaf51d1-7bd5-4485-8d9b-871d0ad80ce1"
      },
      "source": [
        "tf.keras.backend.clear_session()\n",
        "tf.random.set_seed(51)\n",
        "np.random.seed(51)\n",
        "\n",
        "dataset = windowed_dataset(x_train, window_size, batch_size=128, shuffle_buffer=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.SimpleRNN(40, return_sequences=True),\n",
        "  tf.keras.layers.SimpleRNN(40),\n",
        "  tf.keras.layers.Dense(1),\n",
        "  tf.keras.layers.Lambda(lambda x: x * 100.0)\n",
        "])\n",
        "\n",
        "optimizer = tf.keras.optimizers.SGD(lr=5e-5, momentum=0.9)\n",
        "model.compile(loss=tf.keras.losses.Huber(),\n",
        "              optimizer=optimizer,\n",
        "              metrics=[\"mae\"])\n",
        "history = model.fit(dataset,epochs=400)"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/400\n",
            "8/8 [==============================] - 2s 208ms/step - loss: 80.8704 - mae: 83.5782\n",
            "Epoch 2/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 23.3276 - mae: 24.1570\n",
            "Epoch 3/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 17.8614 - mae: 18.4302\n",
            "Epoch 4/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 11.9517 - mae: 12.5089\n",
            "Epoch 5/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 9.9229 - mae: 10.1368\n",
            "Epoch 6/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 13.7531 - mae: 14.3888\n",
            "Epoch 7/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 9.6444 - mae: 10.0865\n",
            "Epoch 8/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 11.9429 - mae: 12.3598\n",
            "Epoch 9/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 13.2042 - mae: 13.7624\n",
            "Epoch 10/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 9.5772 - mae: 10.0502\n",
            "Epoch 11/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 6.6082 - mae: 7.0915\n",
            "Epoch 12/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 8.6432 - mae: 8.9645\n",
            "Epoch 13/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 9.8198 - mae: 10.2675\n",
            "Epoch 14/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 7.9072 - mae: 8.4658\n",
            "Epoch 15/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 9.5199 - mae: 10.0401\n",
            "Epoch 16/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 15.4967 - mae: 16.1459\n",
            "Epoch 17/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 11.4394 - mae: 12.0175\n",
            "Epoch 18/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 8.0569 - mae: 8.5736\n",
            "Epoch 19/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 7.6973 - mae: 8.2719\n",
            "Epoch 20/400\n",
            "8/8 [==============================] - 0s 22ms/step - loss: 9.8644 - mae: 10.3930\n",
            "Epoch 21/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 8.3163 - mae: 8.8850\n",
            "Epoch 22/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 7.0390 - mae: 7.5091\n",
            "Epoch 23/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 7.3821 - mae: 7.9541\n",
            "Epoch 24/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 6.6646 - mae: 7.1299\n",
            "Epoch 25/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 6.4968 - mae: 7.0362\n",
            "Epoch 26/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 6.3651 - mae: 6.8477\n",
            "Epoch 27/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 6.4460 - mae: 6.9801\n",
            "Epoch 28/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 6.2136 - mae: 6.7264\n",
            "Epoch 29/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 5.7063 - mae: 6.2005\n",
            "Epoch 30/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 6.0476 - mae: 6.5154\n",
            "Epoch 31/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 6.3494 - mae: 6.8976\n",
            "Epoch 32/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.7394 - mae: 6.1915\n",
            "Epoch 33/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 6.4426 - mae: 6.9754\n",
            "Epoch 34/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 5.7839 - mae: 6.3072\n",
            "Epoch 35/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 6.0123 - mae: 6.4884\n",
            "Epoch 36/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 6.5899 - mae: 7.0269\n",
            "Epoch 37/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 5.9701 - mae: 6.4132\n",
            "Epoch 38/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.9299 - mae: 6.4788\n",
            "Epoch 39/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 5.6676 - mae: 6.1678\n",
            "Epoch 40/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 5.8731 - mae: 6.3763\n",
            "Epoch 41/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.5608 - mae: 6.0749\n",
            "Epoch 42/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.5548 - mae: 6.0078\n",
            "Epoch 43/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 6.0374 - mae: 6.5646\n",
            "Epoch 44/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.4264 - mae: 5.9330\n",
            "Epoch 45/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.5587 - mae: 6.0465\n",
            "Epoch 46/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.5488 - mae: 6.0709\n",
            "Epoch 47/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.4385 - mae: 5.9662\n",
            "Epoch 48/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.6475 - mae: 6.1132\n",
            "Epoch 49/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.6170 - mae: 6.0925\n",
            "Epoch 50/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 6.2286 - mae: 6.6934\n",
            "Epoch 51/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 5.8892 - mae: 6.3588\n",
            "Epoch 52/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.5168 - mae: 5.9632\n",
            "Epoch 53/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.7025 - mae: 6.2396\n",
            "Epoch 54/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 5.2553 - mae: 5.7090\n",
            "Epoch 55/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.8335 - mae: 6.3837\n",
            "Epoch 56/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 6.0695 - mae: 6.5717\n",
            "Epoch 57/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.9111 - mae: 6.2995\n",
            "Epoch 58/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 6.5007 - mae: 7.0326\n",
            "Epoch 59/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.8078 - mae: 6.2632\n",
            "Epoch 60/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.1788 - mae: 5.7026\n",
            "Epoch 61/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.3854 - mae: 5.8942\n",
            "Epoch 62/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 6.6076 - mae: 7.0712\n",
            "Epoch 63/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 6.3051 - mae: 6.7785\n",
            "Epoch 64/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.9726 - mae: 6.5013\n",
            "Epoch 65/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.3941 - mae: 5.8732\n",
            "Epoch 66/400\n",
            "8/8 [==============================] - 0s 27ms/step - loss: 5.1296 - mae: 5.6305\n",
            "Epoch 67/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.0135 - mae: 5.5158\n",
            "Epoch 68/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.3772 - mae: 5.8360\n",
            "Epoch 69/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.4106 - mae: 5.8901\n",
            "Epoch 70/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.8135 - mae: 5.3153\n",
            "Epoch 71/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.2627 - mae: 5.7625\n",
            "Epoch 72/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.2635 - mae: 5.7684\n",
            "Epoch 73/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.3695 - mae: 5.8514\n",
            "Epoch 74/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 5.0305 - mae: 5.5360\n",
            "Epoch 75/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.1604 - mae: 5.6421\n",
            "Epoch 76/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.0414 - mae: 5.5559\n",
            "Epoch 77/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 5.0557 - mae: 5.5474\n",
            "Epoch 78/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.2191 - mae: 5.6977\n",
            "Epoch 79/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.1960 - mae: 5.6555\n",
            "Epoch 80/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.3428 - mae: 5.8662\n",
            "Epoch 81/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.9491 - mae: 5.4538\n",
            "Epoch 82/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 5.4129 - mae: 5.9502\n",
            "Epoch 83/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.2652 - mae: 5.7838\n",
            "Epoch 84/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.9850 - mae: 5.4801\n",
            "Epoch 85/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 5.1550 - mae: 5.6401\n",
            "Epoch 86/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 5.0942 - mae: 5.5467\n",
            "Epoch 87/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.9103 - mae: 5.3896\n",
            "Epoch 88/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 5.2962 - mae: 5.8259\n",
            "Epoch 89/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.9364 - mae: 6.4256\n",
            "Epoch 90/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.4298 - mae: 5.9235\n",
            "Epoch 91/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 5.0135 - mae: 5.4787\n",
            "Epoch 92/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.9118 - mae: 5.4146\n",
            "Epoch 93/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.0287 - mae: 5.5195\n",
            "Epoch 94/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 5.0464 - mae: 5.5281\n",
            "Epoch 95/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.2497 - mae: 5.7304\n",
            "Epoch 96/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.6079 - mae: 6.0260\n",
            "Epoch 97/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.3895 - mae: 5.9045\n",
            "Epoch 98/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.8659 - mae: 5.3592\n",
            "Epoch 99/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.5658 - mae: 6.1080\n",
            "Epoch 100/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.2967 - mae: 5.8154\n",
            "Epoch 101/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.2265 - mae: 5.7157\n",
            "Epoch 102/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 5.0396 - mae: 5.5104\n",
            "Epoch 103/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.9380 - mae: 5.4349\n",
            "Epoch 104/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.9759 - mae: 5.4768\n",
            "Epoch 105/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 5.0993 - mae: 5.5844\n",
            "Epoch 106/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 5.2538 - mae: 5.7379\n",
            "Epoch 107/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.1860 - mae: 5.6306\n",
            "Epoch 108/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.1509 - mae: 5.6691\n",
            "Epoch 109/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.4329 - mae: 5.9368\n",
            "Epoch 110/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.5408 - mae: 6.0208\n",
            "Epoch 111/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 5.5050 - mae: 5.9540\n",
            "Epoch 112/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.8102 - mae: 5.3154\n",
            "Epoch 113/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.0814 - mae: 5.5953\n",
            "Epoch 114/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.4728 - mae: 5.9711\n",
            "Epoch 115/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 5.3848 - mae: 5.8737\n",
            "Epoch 116/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.4931 - mae: 5.9528\n",
            "Epoch 117/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.7523 - mae: 5.2565\n",
            "Epoch 118/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.9902 - mae: 5.5124\n",
            "Epoch 119/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.3468 - mae: 5.8774\n",
            "Epoch 120/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.2904 - mae: 5.7851\n",
            "Epoch 121/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.2240 - mae: 5.6981\n",
            "Epoch 122/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.8457 - mae: 5.3401\n",
            "Epoch 123/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.8039 - mae: 5.3028\n",
            "Epoch 124/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.8495 - mae: 5.3455\n",
            "Epoch 125/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.9306 - mae: 5.4296\n",
            "Epoch 126/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.1673 - mae: 5.6428\n",
            "Epoch 127/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.0140 - mae: 5.4867\n",
            "Epoch 128/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.1420 - mae: 5.6648\n",
            "Epoch 129/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 5.6854 - mae: 6.1185\n",
            "Epoch 130/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 5.0331 - mae: 5.5507\n",
            "Epoch 131/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.6675 - mae: 5.1400\n",
            "Epoch 132/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.9086 - mae: 5.4107\n",
            "Epoch 133/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.1411 - mae: 5.6239\n",
            "Epoch 134/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.7834 - mae: 5.2794\n",
            "Epoch 135/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.9193 - mae: 5.4100\n",
            "Epoch 136/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.0235 - mae: 5.5105\n",
            "Epoch 137/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.8723 - mae: 5.3488\n",
            "Epoch 138/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.8258 - mae: 5.3051\n",
            "Epoch 139/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.8737 - mae: 5.3586\n",
            "Epoch 140/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.6831 - mae: 5.1873\n",
            "Epoch 141/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.6953 - mae: 5.1859\n",
            "Epoch 142/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 5.0588 - mae: 5.5390\n",
            "Epoch 143/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.8371 - mae: 5.3231\n",
            "Epoch 144/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.9601 - mae: 5.4611\n",
            "Epoch 145/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.6306 - mae: 5.1283\n",
            "Epoch 146/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 5.0026 - mae: 5.4933\n",
            "Epoch 147/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.7422 - mae: 5.2391\n",
            "Epoch 148/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.0207 - mae: 5.5037\n",
            "Epoch 149/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.7012 - mae: 5.1979\n",
            "Epoch 150/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.0062 - mae: 5.5013\n",
            "Epoch 151/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.7464 - mae: 5.2364\n",
            "Epoch 152/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.9430 - mae: 5.4424\n",
            "Epoch 153/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.8751 - mae: 5.3534\n",
            "Epoch 154/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.8408 - mae: 5.3314\n",
            "Epoch 155/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.7779 - mae: 5.2683\n",
            "Epoch 156/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.7920 - mae: 5.2825\n",
            "Epoch 157/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.9634 - mae: 5.4494\n",
            "Epoch 158/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.7899 - mae: 5.2661\n",
            "Epoch 159/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.8686 - mae: 5.3703\n",
            "Epoch 160/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.7271 - mae: 5.2043\n",
            "Epoch 161/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.6255 - mae: 5.1258\n",
            "Epoch 162/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.6798 - mae: 5.1814\n",
            "Epoch 163/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.9250 - mae: 5.4057\n",
            "Epoch 164/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.7610 - mae: 5.2473\n",
            "Epoch 165/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.9237 - mae: 5.4185\n",
            "Epoch 166/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.6562 - mae: 5.1449\n",
            "Epoch 167/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.9317 - mae: 5.4347\n",
            "Epoch 168/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.6016 - mae: 5.0989\n",
            "Epoch 169/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.9349 - mae: 5.4275\n",
            "Epoch 170/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.6942 - mae: 5.1878\n",
            "Epoch 171/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.9674 - mae: 5.4536\n",
            "Epoch 172/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.6758 - mae: 5.1602\n",
            "Epoch 173/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.9629 - mae: 5.4694\n",
            "Epoch 174/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.9367 - mae: 5.4304\n",
            "Epoch 175/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 5.5482 - mae: 5.9986\n",
            "Epoch 176/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.0435 - mae: 5.5053\n",
            "Epoch 177/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.6853 - mae: 5.1809\n",
            "Epoch 178/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.5191 - mae: 5.0124\n",
            "Epoch 179/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.7444 - mae: 5.2547\n",
            "Epoch 180/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.7129 - mae: 5.2107\n",
            "Epoch 181/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.9140 - mae: 5.4000\n",
            "Epoch 182/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.8016 - mae: 5.2799\n",
            "Epoch 183/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.7154 - mae: 5.1988\n",
            "Epoch 184/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.7776 - mae: 5.2714\n",
            "Epoch 185/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.5820 - mae: 5.0807\n",
            "Epoch 186/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.7130 - mae: 5.2057\n",
            "Epoch 187/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.7671 - mae: 5.2526\n",
            "Epoch 188/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.9581 - mae: 5.4419\n",
            "Epoch 189/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.6106 - mae: 5.0957\n",
            "Epoch 190/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.9638 - mae: 5.4868\n",
            "Epoch 191/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.3804 - mae: 5.8469\n",
            "Epoch 192/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.9060 - mae: 5.4214\n",
            "Epoch 193/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.8050 - mae: 5.2904\n",
            "Epoch 194/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 5.1318 - mae: 5.6192\n",
            "Epoch 195/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.0511 - mae: 5.5138\n",
            "Epoch 196/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.6943 - mae: 5.1924\n",
            "Epoch 197/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.6690 - mae: 5.1563\n",
            "Epoch 198/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.5897 - mae: 5.0853\n",
            "Epoch 199/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.5857 - mae: 5.0805\n",
            "Epoch 200/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.6612 - mae: 5.1489\n",
            "Epoch 201/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.9118 - mae: 5.3922\n",
            "Epoch 202/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.7955 - mae: 5.2722\n",
            "Epoch 203/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.8390 - mae: 5.3535\n",
            "Epoch 204/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.1157 - mae: 5.5920\n",
            "Epoch 205/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.8592 - mae: 5.3695\n",
            "Epoch 206/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 5.0194 - mae: 5.4862\n",
            "Epoch 207/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.7341 - mae: 5.2433\n",
            "Epoch 208/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.4870 - mae: 4.9780\n",
            "Epoch 209/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.6883 - mae: 5.1894\n",
            "Epoch 210/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.6620 - mae: 5.1508\n",
            "Epoch 211/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.8434 - mae: 5.3353\n",
            "Epoch 212/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.6575 - mae: 5.1351\n",
            "Epoch 213/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.9131 - mae: 5.4084\n",
            "Epoch 214/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.6060 - mae: 5.0891\n",
            "Epoch 215/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.7411 - mae: 5.2467\n",
            "Epoch 216/400\n",
            "8/8 [==============================] - 0s 28ms/step - loss: 4.4663 - mae: 4.9576\n",
            "Epoch 217/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.7969 - mae: 5.3011\n",
            "Epoch 218/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.6844 - mae: 5.1700\n",
            "Epoch 219/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.9174 - mae: 5.4127\n",
            "Epoch 220/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.6381 - mae: 5.1230\n",
            "Epoch 221/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.9398 - mae: 5.4490\n",
            "Epoch 222/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.7563 - mae: 5.2452\n",
            "Epoch 223/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.9777 - mae: 5.4657\n",
            "Epoch 224/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.7903 - mae: 5.2528\n",
            "Epoch 225/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.6749 - mae: 5.1820\n",
            "Epoch 226/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.4998 - mae: 4.9866\n",
            "Epoch 227/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.7929 - mae: 5.2884\n",
            "Epoch 228/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.7412 - mae: 5.2161\n",
            "Epoch 229/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.6796 - mae: 5.1812\n",
            "Epoch 230/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.6938 - mae: 5.1836\n",
            "Epoch 231/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.7242 - mae: 5.2138\n",
            "Epoch 232/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.5868 - mae: 5.0736\n",
            "Epoch 233/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.7289 - mae: 5.2286\n",
            "Epoch 234/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.4758 - mae: 4.9586\n",
            "Epoch 235/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.7398 - mae: 5.2417\n",
            "Epoch 236/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.5819 - mae: 5.0698\n",
            "Epoch 237/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.9254 - mae: 5.4226\n",
            "Epoch 238/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.6201 - mae: 5.0974\n",
            "Epoch 239/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.8500 - mae: 5.3488\n",
            "Epoch 240/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.5053 - mae: 4.9894\n",
            "Epoch 241/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.8310 - mae: 5.3332\n",
            "Epoch 242/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.5567 - mae: 5.0492\n",
            "Epoch 243/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.8431 - mae: 5.3428\n",
            "Epoch 244/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.5626 - mae: 5.0478\n",
            "Epoch 245/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.9319 - mae: 5.4176\n",
            "Epoch 246/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.8185 - mae: 5.2778\n",
            "Epoch 247/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.4764 - mae: 4.9770\n",
            "Epoch 248/400\n",
            "8/8 [==============================] - 0s 27ms/step - loss: 4.6276 - mae: 5.1355\n",
            "Epoch 249/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.7669 - mae: 5.2793\n",
            "Epoch 250/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.8876 - mae: 5.3765\n",
            "Epoch 251/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.7669 - mae: 5.2354\n",
            "Epoch 252/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.6129 - mae: 5.1190\n",
            "Epoch 253/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.5414 - mae: 5.0251\n",
            "Epoch 254/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.5789 - mae: 5.0796\n",
            "Epoch 255/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.4604 - mae: 4.9552\n",
            "Epoch 256/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.6839 - mae: 5.1786\n",
            "Epoch 257/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.4355 - mae: 4.9225\n",
            "Epoch 258/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.7339 - mae: 5.2326\n",
            "Epoch 259/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.5283 - mae: 5.0135\n",
            "Epoch 260/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.8893 - mae: 5.3655\n",
            "Epoch 261/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.7590 - mae: 5.2170\n",
            "Epoch 262/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.5762 - mae: 5.0762\n",
            "Epoch 263/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.4482 - mae: 4.9347\n",
            "Epoch 264/400\n",
            "8/8 [==============================] - 0s 28ms/step - loss: 4.7061 - mae: 5.2039\n",
            "Epoch 265/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.5936 - mae: 5.0754\n",
            "Epoch 266/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.6464 - mae: 5.1485\n",
            "Epoch 267/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.5817 - mae: 5.0625\n",
            "Epoch 268/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.6119 - mae: 5.1081\n",
            "Epoch 269/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.4469 - mae: 4.9380\n",
            "Epoch 270/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.6820 - mae: 5.1849\n",
            "Epoch 271/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.5896 - mae: 5.0874\n",
            "Epoch 272/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.7745 - mae: 5.2725\n",
            "Epoch 273/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.5272 - mae: 5.0048\n",
            "Epoch 274/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.7117 - mae: 5.2125\n",
            "Epoch 275/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.5031 - mae: 4.9800\n",
            "Epoch 276/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.7407 - mae: 5.2321\n",
            "Epoch 277/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.7271 - mae: 5.1942\n",
            "Epoch 278/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.5663 - mae: 5.0648\n",
            "Epoch 279/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.3552 - mae: 4.8486\n",
            "Epoch 280/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.5490 - mae: 5.0508\n",
            "Epoch 281/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.4629 - mae: 4.9568\n",
            "Epoch 282/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.6156 - mae: 5.1146\n",
            "Epoch 283/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.4958 - mae: 4.9770\n",
            "Epoch 284/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.6949 - mae: 5.1886\n",
            "Epoch 285/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.5979 - mae: 5.0898\n",
            "Epoch 286/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.6139 - mae: 5.1052\n",
            "Epoch 287/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.3893 - mae: 4.8766\n",
            "Epoch 288/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.6727 - mae: 5.1753\n",
            "Epoch 289/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.4902 - mae: 4.9820\n",
            "Epoch 290/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.7245 - mae: 5.2262\n",
            "Epoch 291/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.4759 - mae: 4.9607\n",
            "Epoch 292/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.6720 - mae: 5.1676\n",
            "Epoch 293/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.6047 - mae: 5.0993\n",
            "Epoch 294/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.9604 - mae: 5.4303\n",
            "Epoch 295/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 5.0377 - mae: 5.4999\n",
            "Epoch 296/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.3708 - mae: 4.8551\n",
            "Epoch 297/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.6241 - mae: 5.1378\n",
            "Epoch 298/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.7608 - mae: 5.2527\n",
            "Epoch 299/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.7810 - mae: 5.2699\n",
            "Epoch 300/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.5515 - mae: 5.0151\n",
            "Epoch 301/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.5024 - mae: 5.0036\n",
            "Epoch 302/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.3483 - mae: 4.8427\n",
            "Epoch 303/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.5261 - mae: 5.0303\n",
            "Epoch 304/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.4842 - mae: 4.9746\n",
            "Epoch 305/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.6084 - mae: 5.1103\n",
            "Epoch 306/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.4803 - mae: 4.9608\n",
            "Epoch 307/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.6127 - mae: 5.1124\n",
            "Epoch 308/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.4506 - mae: 4.9350\n",
            "Epoch 309/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.5369 - mae: 5.0396\n",
            "Epoch 310/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.5095 - mae: 4.9951\n",
            "Epoch 311/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.7421 - mae: 5.2246\n",
            "Epoch 312/400\n",
            "8/8 [==============================] - 0s 28ms/step - loss: 4.7025 - mae: 5.1680\n",
            "Epoch 313/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.4525 - mae: 4.9376\n",
            "Epoch 314/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.5080 - mae: 4.9923\n",
            "Epoch 315/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.4057 - mae: 4.9009\n",
            "Epoch 316/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.4714 - mae: 4.9640\n",
            "Epoch 317/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.5143 - mae: 5.0136\n",
            "Epoch 318/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.5409 - mae: 5.0187\n",
            "Epoch 319/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.6765 - mae: 5.1673\n",
            "Epoch 320/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.4604 - mae: 4.9280\n",
            "Epoch 321/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.6581 - mae: 5.1493\n",
            "Epoch 322/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.4798 - mae: 4.9648\n",
            "Epoch 323/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.5273 - mae: 5.0263\n",
            "Epoch 324/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.5360 - mae: 5.0330\n",
            "Epoch 325/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.6882 - mae: 5.1793\n",
            "Epoch 326/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.6191 - mae: 5.0841\n",
            "Epoch 327/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.4858 - mae: 4.9854\n",
            "Epoch 328/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.3960 - mae: 4.8933\n",
            "Epoch 329/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.4074 - mae: 4.9065\n",
            "Epoch 330/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.4575 - mae: 4.9451\n",
            "Epoch 331/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.4808 - mae: 4.9790\n",
            "Epoch 332/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.5151 - mae: 4.9913\n",
            "Epoch 333/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.5577 - mae: 5.0571\n",
            "Epoch 334/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.3881 - mae: 4.8723\n",
            "Epoch 335/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.5354 - mae: 5.0315\n",
            "Epoch 336/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.5479 - mae: 5.0472\n",
            "Epoch 337/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.6615 - mae: 5.1533\n",
            "Epoch 338/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.6139 - mae: 5.0785\n",
            "Epoch 339/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.4597 - mae: 4.9573\n",
            "Epoch 340/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.4044 - mae: 4.9003\n",
            "Epoch 341/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.4026 - mae: 4.9028\n",
            "Epoch 342/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.4961 - mae: 4.9857\n",
            "Epoch 343/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.4865 - mae: 4.9843\n",
            "Epoch 344/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.5129 - mae: 4.9897\n",
            "Epoch 345/400\n",
            "8/8 [==============================] - 0s 27ms/step - loss: 4.5893 - mae: 5.0856\n",
            "Epoch 346/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.4548 - mae: 4.9195\n",
            "Epoch 347/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.3973 - mae: 4.8833\n",
            "Epoch 348/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.3218 - mae: 4.8100\n",
            "Epoch 349/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.3841 - mae: 4.8763\n",
            "Epoch 350/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.4064 - mae: 4.8931\n",
            "Epoch 351/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.4815 - mae: 4.9848\n",
            "Epoch 352/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.5600 - mae: 5.0477\n",
            "Epoch 353/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.8013 - mae: 5.2842\n",
            "Epoch 354/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.7817 - mae: 5.2424\n",
            "Epoch 355/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.3809 - mae: 4.8530\n",
            "Epoch 356/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.5082 - mae: 4.9722\n",
            "Epoch 357/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.3063 - mae: 4.7963\n",
            "Epoch 358/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.2347 - mae: 4.7043\n",
            "Epoch 359/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.4311 - mae: 4.9374\n",
            "Epoch 360/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.3678 - mae: 4.8616\n",
            "Epoch 361/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.4475 - mae: 4.9483\n",
            "Epoch 362/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.3919 - mae: 4.8878\n",
            "Epoch 363/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.4438 - mae: 4.9405\n",
            "Epoch 364/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.4149 - mae: 4.9060\n",
            "Epoch 365/400\n",
            "8/8 [==============================] - 0s 28ms/step - loss: 4.4960 - mae: 4.9932\n",
            "Epoch 366/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.4374 - mae: 4.9241\n",
            "Epoch 367/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.6218 - mae: 5.1215\n",
            "Epoch 368/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.4281 - mae: 4.8960\n",
            "Epoch 369/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.4949 - mae: 4.9893\n",
            "Epoch 370/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.4215 - mae: 4.8994\n",
            "Epoch 371/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.6447 - mae: 5.1249\n",
            "Epoch 372/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.6918 - mae: 5.1497\n",
            "Epoch 373/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.3406 - mae: 4.8351\n",
            "Epoch 374/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.3811 - mae: 4.8690\n",
            "Epoch 375/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.3876 - mae: 4.8901\n",
            "Epoch 376/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.5045 - mae: 5.0080\n",
            "Epoch 377/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.4805 - mae: 4.9839\n",
            "Epoch 378/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.4832 - mae: 4.9693\n",
            "Epoch 379/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.6799 - mae: 5.1661\n",
            "Epoch 380/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.7871 - mae: 5.2483\n",
            "Epoch 381/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.2664 - mae: 4.7465\n",
            "Epoch 382/400\n",
            "8/8 [==============================] - 0s 25ms/step - loss: 4.3735 - mae: 4.8322\n",
            "Epoch 383/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.3233 - mae: 4.8162\n",
            "Epoch 384/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.4963 - mae: 4.9984\n",
            "Epoch 385/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.5007 - mae: 4.9995\n",
            "Epoch 386/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.7072 - mae: 5.1745\n",
            "Epoch 387/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.3690 - mae: 4.8617\n",
            "Epoch 388/400\n",
            "8/8 [==============================] - 0s 26ms/step - loss: 4.3604 - mae: 4.8504\n",
            "Epoch 389/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.3868 - mae: 4.8879\n",
            "Epoch 390/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.4175 - mae: 4.9106\n",
            "Epoch 391/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.5137 - mae: 5.0148\n",
            "Epoch 392/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.3929 - mae: 4.8731\n",
            "Epoch 393/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.5807 - mae: 5.0718\n",
            "Epoch 394/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.5666 - mae: 5.0292\n",
            "Epoch 395/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.2781 - mae: 4.7548\n",
            "Epoch 396/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.3701 - mae: 4.8459\n",
            "Epoch 397/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.2908 - mae: 4.7806\n",
            "Epoch 398/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.4227 - mae: 4.9224\n",
            "Epoch 399/400\n",
            "8/8 [==============================] - 0s 24ms/step - loss: 4.4672 - mae: 4.9714\n",
            "Epoch 400/400\n",
            "8/8 [==============================] - 0s 23ms/step - loss: 4.8405 - mae: 5.3135\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ejBynEKekaKw",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 392
        },
        "outputId": "7707b839-30ff-47a2-88ca-89dd88722c34"
      },
      "source": [
        "forecast=[]\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": 17,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmcAAAF3CAYAAADgjOwXAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nIQrORERERGqIgjMRERGRGvL/AR5UGwU/sOXpAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 720x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "hR2BO0Dai_ZT",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "2662b05f-9692-4632-ee44-ab5128ebd610"
      },
      "source": [
        "tf.keras.metrics.mean_absolute_error(x_valid, results).numpy()"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "6.4141674"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vXaFHOaJqE6M",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 607
        },
        "outputId": "ae60a5f8-334e-4016-8937-1042e2166953"
      },
      "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": 19,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 0 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEWCAYAAABrDZDcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\niIikuP8PFMVJa+uU3wwAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nURSlHRUFRVEUpR0VBUVRFKWd/wfcXzk427xsmgAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 0 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uz-GXTCbsVfT",
        "colab_type": "text"
      },
      "source": [
        "# Seems to work fine!"
      ]
    }
  ]
}