Notebooks/2. Linear_Model_Forecast.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Linear Model Forecast.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "OURPn8p6fYFv",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5uobQzuWfYZK",
        "colab_type": "text"
      },
      "source": [
        "# This is Keras/TensorFlow right? I thought this was Deep Learning?\n",
        "\n",
        "- Output = activation(dot(input, kernel) + bias\n",
        "\n",
        "That looks familiar doesn't it? It looks almost identical to y = mx+b. The dot product is sum of the products in two sequences. Well, if there is only two sequences with a length of 1 then it is just the product of those two numbers. This simplifies down to the all to familiar y = mx + b.\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "vidayERjaO5q"
      },
      "source": [
        "## Setup"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_Tg2O-2wgfob",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 33
        },
        "outputId": "43ea3693-d88d-49ed-e7f9-f7f2d2bc719b"
      },
      "source": [
        "# Example of dot product of two sequences of length 1\n",
        "np.dot(2,1) + 5"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "7"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "gqWabzlJ63nL",
        "colab": {}
      },
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "import tensorflow as tf\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "import seaborn as sns\n",
        "\n",
        "keras = tf.keras\n",
        "\n",
        "# set style of charts\n",
        "sns.set(style=\"darkgrid\")\n",
        "plt.rcParams['figure.figsize'] = [10, 10]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "cg1hfKCPldZG",
        "colab": {}
      },
      "source": [
        "def plot_series(time, series, format=\"-\", start=0, end=None, label=None):\n",
        "    plt.plot(time[start:end], series[start:end], format, label=label)\n",
        "    plt.xlabel(\"Time\")\n",
        "    plt.ylabel(\"Value\")\n",
        "    if label:\n",
        "        plt.legend(fontsize=14)\n",
        "    plt.grid(True)\n",
        "\n",
        "def model_forecast(model, series, window_size):\n",
        "    ds = tf.data.Dataset.from_tensor_slices(series)\n",
        "    ds = ds.window(window_size, shift=1, drop_remainder=True)\n",
        "    ds = ds.flat_map(lambda w: w.batch(window_size))\n",
        "    ds = ds.batch(32).prefetch(1)\n",
        "    forecast = model.predict(ds)\n",
        "    return forecast\n",
        "\n",
        "def window_dataset(series, window_size, batch_size=128,\n",
        "                   shuffle_buffer=1000):\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)\n",
        "    dataset = dataset.map(lambda window: (window[:-1], window[-1]))\n",
        "    dataset = dataset.batch(batch_size).prefetch(1)\n",
        "    return dataset"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "ViWVB9qd8OIR"
      },
      "source": [
        "## Forecasting with Machine Learning\n",
        "\n",
        "First, we will train a model to forecast the next step given the previous 30 steps, therefore, we need to create a dataset of 20-step windows for training. Every 20 steps is 1 trading month since the markets are only open Monday - Friday."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "Zmp1JXKxk9Vb",
        "colab": {}
      },
      "source": [
        "# Read in data\n",
        "spy = pd.read_csv('SPY.csv')\n",
        "\n",
        "# Convert series into datetime type\n",
        "spy['Date'] = pd.to_datetime(spy['Date'])\n",
        "\n",
        "# Save target series\n",
        "series = spy['Close']\n",
        "\n",
        "# Create train data set\n",
        "train_split_date = '2014-12-31'\n",
        "train_split_index = np.where(spy.Date == train_split_date)[0][0]\n",
        "x_train = spy.loc[spy['Date'] <= train_split_date]['Close']\n",
        "\n",
        "# Create test data set\n",
        "test_split_date = '2019-01-02'\n",
        "test_split_index = np.where(spy.Date == test_split_date)[0][0]\n",
        "x_test = spy.loc[spy['Date'] >= test_split_date]['Close']\n",
        "\n",
        "# Create valid data set\n",
        "valid_split_index = (train_split_index.max(),test_split_index.min())\n",
        "x_valid = spy.loc[(spy['Date'] < test_split_date) & (spy['Date'] > train_split_date)]['Close']"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xxpJz3EkErei",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 610
        },
        "outputId": "15f0ab85-9b16-4bf8-8b2e-c9c0928b13da"
      },
      "source": [
        "# Plot all lines on one chart to see where one segment starts and another ends\n",
        "plt.plot(x_train, label = 'Train')\n",
        "plt.plot(x_valid, label = 'Validate')\n",
        "plt.plot(x_test, label = 'Test')\n",
        "plt.legend()\n",
        "print(x_train.index.max(),x_valid.index.min(),x_valid.index.max(),x_test.index.min(),x_test.index.max())"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "5521 5522 6527 6528 6949\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x720 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xTOs5lgVEtuG",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Reshape values\n",
        "x_train_values = x_train.values.reshape(-1, 1)\n",
        "x_valid_values = x_valid.values.reshape(-1, 1)\n",
        "x_test_values = x_test.values.reshape(-1, 1)\n",
        "\n",
        "#  Create Scaler Object\n",
        "x_train_scaler = MinMaxScaler(feature_range=(0, 1))\n",
        "\n",
        "# Fit x_train values\n",
        "normalized_x_train = x_train_scaler.fit_transform(x_train_values)\n",
        "\n",
        "# Fit x_valid values\n",
        "normalized_x_valid = x_train_scaler.transform(x_valid_values)\n",
        "\n",
        "# Fit x_test values\n",
        "normalized_x_test = x_train_scaler.transform(x_test_values)\n",
        "\n",
        "# All values normalized to training data\n",
        "spy_normalized_to_traindata = x_train_scaler.transform(series.values.reshape(-1, 1))\n",
        "\n",
        "# Example of how to iverse\n",
        "# inversed = scaler.inverse_transform(normalized_x_train).flatten()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "T1IvwAFn8OIc"
      },
      "source": [
        "### Linear Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lT-WEKKbUIaj",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "4eeb51aa-8eab-491a-83b2-609d410305a1"
      },
      "source": [
        "# Clears any background saved info useful in notebooks\n",
        "keras.backend.clear_session()\n",
        "\n",
        "# Make reproducible \n",
        "tf.random.set_seed(42)\n",
        "np.random.seed(42)\n",
        "\n",
        "# set window size \n",
        "window_size = 20\n",
        "\n",
        "# define training data (20 day windows shifted by 1 every time)\n",
        "train_set = window_dataset(normalized_x_train.flatten(), window_size)\n",
        "\n",
        "# Build Linear Model of a single dense layer\n",
        "model = keras.models.Sequential([\n",
        "  keras.layers.Dense(1, input_shape=[window_size])\n",
        "])\n",
        "\n",
        "# Find optimal learning rate\n",
        "lr_schedule = keras.callbacks.LearningRateScheduler(\n",
        "    lambda epoch: 1e-6 * 10**(epoch / 30))\n",
        "optimizer = keras.optimizers.Nadam(lr=1e-6)\n",
        "model.compile(loss=keras.losses.Huber(),\n",
        "              optimizer=optimizer,\n",
        "              metrics=[\"mae\"])\n",
        "\n",
        "# Fit the model\n",
        "history = model.fit(train_set, epochs=100, callbacks=[lr_schedule])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0482 - mae: 0.2671\n",
            "Epoch 2/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 0.0481 - mae: 0.2668\n",
            "Epoch 3/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0480 - mae: 0.2664\n",
            "Epoch 4/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0479 - mae: 0.2660\n",
            "Epoch 5/100\n",
            "43/43 [==============================] - 0s 12ms/step - loss: 0.0477 - mae: 0.2656\n",
            "Epoch 6/100\n",
            "43/43 [==============================] - 1s 14ms/step - loss: 0.0476 - mae: 0.2652\n",
            "Epoch 7/100\n",
            "43/43 [==============================] - 0s 12ms/step - loss: 0.0474 - mae: 0.2647\n",
            "Epoch 8/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0473 - mae: 0.2643\n",
            "Epoch 9/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0471 - mae: 0.2637\n",
            "Epoch 10/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0469 - mae: 0.2632\n",
            "Epoch 11/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0467 - mae: 0.2625\n",
            "Epoch 12/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0464 - mae: 0.2619\n",
            "Epoch 13/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 0.0462 - mae: 0.2612\n",
            "Epoch 14/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0459 - mae: 0.2604\n",
            "Epoch 15/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 0.0457 - mae: 0.2596\n",
            "Epoch 16/100\n",
            "43/43 [==============================] - 1s 14ms/step - loss: 0.0454 - mae: 0.2587\n",
            "Epoch 17/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 0.0450 - mae: 0.2577\n",
            "Epoch 18/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0447 - mae: 0.2567\n",
            "Epoch 19/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 0.0443 - mae: 0.2556\n",
            "Epoch 20/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0439 - mae: 0.2544\n",
            "Epoch 21/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 0.0435 - mae: 0.2531\n",
            "Epoch 22/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 0.0430 - mae: 0.2517\n",
            "Epoch 23/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0425 - mae: 0.2502\n",
            "Epoch 24/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0420 - mae: 0.2487\n",
            "Epoch 25/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0415 - mae: 0.2469\n",
            "Epoch 26/100\n",
            "43/43 [==============================] - 1s 14ms/step - loss: 0.0409 - mae: 0.2451\n",
            "Epoch 27/100\n",
            "43/43 [==============================] - 1s 14ms/step - loss: 0.0402 - mae: 0.2431\n",
            "Epoch 28/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 0.0395 - mae: 0.2410\n",
            "Epoch 29/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0388 - mae: 0.2387\n",
            "Epoch 30/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0381 - mae: 0.2362\n",
            "Epoch 31/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0372 - mae: 0.2336\n",
            "Epoch 32/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 0.0364 - mae: 0.2308\n",
            "Epoch 33/100\n",
            "43/43 [==============================] - 1s 14ms/step - loss: 0.0355 - mae: 0.2277\n",
            "Epoch 34/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0345 - mae: 0.2245\n",
            "Epoch 35/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 0.0335 - mae: 0.2211\n",
            "Epoch 36/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0324 - mae: 0.2174\n",
            "Epoch 37/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 0.0313 - mae: 0.2135\n",
            "Epoch 38/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0301 - mae: 0.2093\n",
            "Epoch 39/100\n",
            "43/43 [==============================] - 1s 14ms/step - loss: 0.0289 - mae: 0.2049\n",
            "Epoch 40/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0276 - mae: 0.2002\n",
            "Epoch 41/100\n",
            "43/43 [==============================] - 1s 14ms/step - loss: 0.0263 - mae: 0.1952\n",
            "Epoch 42/100\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 0.0249 - mae: 0.1899\n",
            "Epoch 43/100\n",
            "43/43 [==============================] - 1s 14ms/step - loss: 0.0235 - mae: 0.1843\n",
            "Epoch 44/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0221 - mae: 0.1784\n",
            "Epoch 45/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0206 - mae: 0.1721\n",
            "Epoch 46/100\n",
            "43/43 [==============================] - 1s 14ms/step - loss: 0.0191 - mae: 0.1656\n",
            "Epoch 47/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0176 - mae: 0.1588\n",
            "Epoch 48/100\n",
            "43/43 [==============================] - 1s 14ms/step - loss: 0.0161 - mae: 0.1516\n",
            "Epoch 49/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0146 - mae: 0.1441\n",
            "Epoch 50/100\n",
            "43/43 [==============================] - 1s 14ms/step - loss: 0.0131 - mae: 0.1364\n",
            "Epoch 51/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0116 - mae: 0.1284\n",
            "Epoch 52/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0103 - mae: 0.1202\n",
            "Epoch 53/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0089 - mae: 0.1119\n",
            "Epoch 54/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0077 - mae: 0.1034\n",
            "Epoch 55/100\n",
            "43/43 [==============================] - 1s 14ms/step - loss: 0.0065 - mae: 0.0949\n",
            "Epoch 56/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 0.0054 - mae: 0.0864\n",
            "Epoch 57/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 0.0045 - mae: 0.0780\n",
            "Epoch 58/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0036 - mae: 0.0698\n",
            "Epoch 59/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 0.0029 - mae: 0.0620\n",
            "Epoch 60/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 0.0023 - mae: 0.0546\n",
            "Epoch 61/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0018 - mae: 0.0478\n",
            "Epoch 62/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 0.0014 - mae: 0.0418\n",
            "Epoch 63/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 0.0011 - mae: 0.0365\n",
            "Epoch 64/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 8.2983e-04 - mae: 0.0322\n",
            "Epoch 65/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 6.7028e-04 - mae: 0.0287\n",
            "Epoch 66/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 5.6246e-04 - mae: 0.0262\n",
            "Epoch 67/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 4.9305e-04 - mae: 0.0246\n",
            "Epoch 68/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 4.5081e-04 - mae: 0.0236\n",
            "Epoch 69/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 4.2603e-04 - mae: 0.0231\n",
            "Epoch 70/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 4.1175e-04 - mae: 0.0229\n",
            "Epoch 71/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 4.0334e-04 - mae: 0.0228\n",
            "Epoch 72/100\n",
            "43/43 [==============================] - 0s 11ms/step - loss: 3.9748e-04 - mae: 0.0227\n",
            "Epoch 73/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 3.9270e-04 - mae: 0.0226\n",
            "Epoch 74/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 3.8798e-04 - mae: 0.0225\n",
            "Epoch 75/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 3.8325e-04 - mae: 0.0224\n",
            "Epoch 76/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 3.7776e-04 - mae: 0.0222\n",
            "Epoch 77/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 3.7227e-04 - mae: 0.0221\n",
            "Epoch 78/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 3.6637e-04 - mae: 0.0219\n",
            "Epoch 79/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 3.6011e-04 - mae: 0.0216\n",
            "Epoch 80/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 3.5348e-04 - mae: 0.0214\n",
            "Epoch 81/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 3.4678e-04 - mae: 0.0212\n",
            "Epoch 82/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 3.3978e-04 - mae: 0.0209\n",
            "Epoch 83/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 3.3256e-04 - mae: 0.0206\n",
            "Epoch 84/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 3.2544e-04 - mae: 0.0204\n",
            "Epoch 85/100\n",
            "43/43 [==============================] - 1s 14ms/step - loss: 3.1808e-04 - mae: 0.0201\n",
            "Epoch 86/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 3.1101e-04 - mae: 0.0197\n",
            "Epoch 87/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 3.0390e-04 - mae: 0.0194\n",
            "Epoch 88/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 2.9702e-04 - mae: 0.0191\n",
            "Epoch 89/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 2.9046e-04 - mae: 0.0188\n",
            "Epoch 90/100\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 2.8438e-04 - mae: 0.0185\n",
            "Epoch 91/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 2.7861e-04 - mae: 0.0182\n",
            "Epoch 92/100\n",
            "43/43 [==============================] - 0s 11ms/step - loss: 2.7321e-04 - mae: 0.0178\n",
            "Epoch 93/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 2.6875e-04 - mae: 0.0176\n",
            "Epoch 94/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 2.6432e-04 - mae: 0.0173\n",
            "Epoch 95/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 2.6077e-04 - mae: 0.0171\n",
            "Epoch 96/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 2.5764e-04 - mae: 0.0168\n",
            "Epoch 97/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 2.5507e-04 - mae: 0.0166\n",
            "Epoch 98/100\n",
            "43/43 [==============================] - 1s 12ms/step - loss: 2.5321e-04 - mae: 0.0164\n",
            "Epoch 99/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 2.5107e-04 - mae: 0.0163\n",
            "Epoch 100/100\n",
            "43/43 [==============================] - 1s 13ms/step - loss: 2.4980e-04 - mae: 0.0161\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XtG2-JTwUwHn",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 621
        },
        "outputId": "8b049381-9523-4bc9-91f5-fa4f146af1cd"
      },
      "source": [
        "plt.semilogx(history.history[\"lr\"], history.history[\"loss\"])\n",
        "plt.axis([1e-6, 1, 0, .01])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(1e-06, 1.0, 0.0, 0.01)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x720 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "ieOKdcEQ0A6k",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "4d2d023a-f0db-4309-b968-fb5caf68733c"
      },
      "source": [
        "# Useful to clear everything when rerunning cells\n",
        "keras.backend.clear_session()\n",
        "\n",
        "# Make this reproducible\n",
        "tf.random.set_seed(42)\n",
        "np.random.seed(42)\n",
        "\n",
        "# Create train and validate windows\n",
        "window_size = 20\n",
        "train_set = window_dataset(normalized_x_train.flatten(), window_size)\n",
        "valid_set = window_dataset(normalized_x_valid.flatten(), window_size)\n",
        "\n",
        "# 1 layer producing linear output for 1 output from each window of 20 days\n",
        "model = keras.models.Sequential([\n",
        "  keras.layers.Dense(1, input_shape=[window_size]) # o\n",
        "])\n",
        "\n",
        "# Huber works well with \"mae\"\n",
        "optimizer = keras.optimizers.Nadam(lr=1e-3)\n",
        "model.compile(loss=keras.losses.Huber(),\n",
        "              optimizer=optimizer,\n",
        "              metrics=[\"mae\"])\n",
        "\n",
        "# create save points for best model\n",
        "model_checkpoint = keras.callbacks.ModelCheckpoint(\n",
        "    \"my_checkpoint\", save_best_only=True)\n",
        "\n",
        "# Set up early stop\n",
        "early_stopping = keras.callbacks.EarlyStopping(patience=10)\n",
        "\n",
        "# fit model to data\n",
        "model.fit(train_set, epochs=500,\n",
        "          validation_data=valid_set,\n",
        "          callbacks=[early_stopping, model_checkpoint])\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/500\n",
            "     41/Unknown - 0s 12ms/step - loss: 0.0980 - mae: 0.3945WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow/python/training/tracking/tracking.py:111: Model.state_updates (from tensorflow.python.keras.engine.training) is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "This property should not be used in TensorFlow 2.0, as updates are applied automatically.\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow/python/training/tracking/tracking.py:111: Layer.updates (from tensorflow.python.keras.engine.base_layer) is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "This property should not be used in TensorFlow 2.0, as updates are applied automatically.\n",
            "INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 25ms/step - loss: 0.0977 - mae: 0.3946 - val_loss: 0.3202 - val_mae: 0.7877\n",
            "Epoch 2/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 0.0101 - mae: 0.1240INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 21ms/step - loss: 0.0108 - mae: 0.1285 - val_loss: 0.0474 - val_mae: 0.3004\n",
            "Epoch 3/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 0.0011 - mae: 0.0384    INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 0.0011 - mae: 0.0395 - val_loss: 0.0097 - val_mae: 0.1326\n",
            "Epoch 4/500\n",
            "37/43 [========================>.....] - ETA: 0s - loss: 6.6360e-04 - mae: 0.0288INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 25ms/step - loss: 6.7882e-04 - mae: 0.0291 - val_loss: 0.0050 - val_mae: 0.0921\n",
            "Epoch 5/500\n",
            "43/43 [==============================] - ETA: 0s - loss: 6.5952e-04 - mae: 0.0292INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 6.5952e-04 - mae: 0.0292 - val_loss: 0.0040 - val_mae: 0.0817\n",
            "Epoch 6/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 6.1500e-04 - mae: 0.0283INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 6.0881e-04 - mae: 0.0281 - val_loss: 0.0035 - val_mae: 0.0754\n",
            "Epoch 7/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 5.6273e-04 - mae: 0.0271INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 5.5294e-04 - mae: 0.0268 - val_loss: 0.0030 - val_mae: 0.0697\n",
            "Epoch 8/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 5.0974e-04 - mae: 0.0259INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 4.9784e-04 - mae: 0.0255 - val_loss: 0.0026 - val_mae: 0.0642\n",
            "Epoch 9/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 4.4726e-04 - mae: 0.0244INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 4.4503e-04 - mae: 0.0241 - val_loss: 0.0022 - val_mae: 0.0586\n",
            "Epoch 10/500\n",
            "39/43 [==========================>...] - ETA: 0s - loss: 4.0477e-04 - mae: 0.0231INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 3.9675e-04 - mae: 0.0228 - val_loss: 0.0019 - val_mae: 0.0537\n",
            "Epoch 11/500\n",
            "43/43 [==============================] - ETA: 0s - loss: 3.5191e-04 - mae: 0.0215INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 3.5191e-04 - mae: 0.0215 - val_loss: 0.0016 - val_mae: 0.0490\n",
            "Epoch 12/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 3.1567e-04 - mae: 0.0206INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 3.1136e-04 - mae: 0.0202 - val_loss: 0.0014 - val_mae: 0.0443\n",
            "Epoch 13/500\n",
            "37/43 [========================>.....] - ETA: 0s - loss: 2.8023e-04 - mae: 0.0194INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 2.7667e-04 - mae: 0.0191 - val_loss: 0.0012 - val_mae: 0.0406\n",
            "Epoch 14/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 2.4411e-04 - mae: 0.0181INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 2.4489e-04 - mae: 0.0180 - val_loss: 0.0010 - val_mae: 0.0371\n",
            "Epoch 15/500\n",
            "43/43 [==============================] - ETA: 0s - loss: 2.1743e-04 - mae: 0.0169INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 2.1743e-04 - mae: 0.0169 - val_loss: 8.8185e-04 - val_mae: 0.0335\n",
            "Epoch 16/500\n",
            "42/43 [============================>.] - ETA: 0s - loss: 1.9628e-04 - mae: 0.0161INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 21ms/step - loss: 1.9545e-04 - mae: 0.0160 - val_loss: 7.8419e-04 - val_mae: 0.0310\n",
            "Epoch 17/500\n",
            "43/43 [==============================] - ETA: 0s - loss: 1.7493e-04 - mae: 0.0152INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 1.7493e-04 - mae: 0.0152 - val_loss: 6.9850e-04 - val_mae: 0.0287\n",
            "Epoch 18/500\n",
            "37/43 [========================>.....] - ETA: 0s - loss: 1.5659e-04 - mae: 0.0144INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 1.5889e-04 - mae: 0.0143 - val_loss: 6.2953e-04 - val_mae: 0.0268\n",
            "Epoch 19/500\n",
            "41/43 [===========================>..] - ETA: 0s - loss: 1.4449e-04 - mae: 0.0137INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 1.4489e-04 - mae: 0.0136 - val_loss: 5.6843e-04 - val_mae: 0.0251\n",
            "Epoch 20/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 1.3485e-04 - mae: 0.0131INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 1.3455e-04 - mae: 0.0130 - val_loss: 5.3348e-04 - val_mae: 0.0240\n",
            "Epoch 21/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 1.2077e-04 - mae: 0.0123INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 1.2518e-04 - mae: 0.0125 - val_loss: 5.0010e-04 - val_mae: 0.0230\n",
            "Epoch 22/500\n",
            "39/43 [==========================>...] - ETA: 0s - loss: 1.1480e-04 - mae: 0.0119INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 1.1782e-04 - mae: 0.0120 - val_loss: 4.6347e-04 - val_mae: 0.0218\n",
            "Epoch 23/500\n",
            "42/43 [============================>.] - ETA: 0s - loss: 1.1294e-04 - mae: 0.0116INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 25ms/step - loss: 1.1295e-04 - mae: 0.0116 - val_loss: 4.5285e-04 - val_mae: 0.0215\n",
            "Epoch 24/500\n",
            "39/43 [==========================>...] - ETA: 0s - loss: 1.0346e-04 - mae: 0.0110INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 1.0773e-04 - mae: 0.0112 - val_loss: 4.3113e-04 - val_mae: 0.0208\n",
            "Epoch 25/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 1.0082e-04 - mae: 0.0107INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 1.0478e-04 - mae: 0.0109 - val_loss: 4.2231e-04 - val_mae: 0.0205\n",
            "Epoch 26/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 1.0204e-04 - mae: 0.0106 - val_loss: 4.2283e-04 - val_mae: 0.0206\n",
            "Epoch 27/500\n",
            "37/43 [========================>.....] - ETA: 0s - loss: 9.2812e-05 - mae: 0.0100INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 9.9561e-05 - mae: 0.0104 - val_loss: 4.0803e-04 - val_mae: 0.0201\n",
            "Epoch 28/500\n",
            "41/43 [===========================>..] - ETA: 0s - loss: 9.6658e-05 - mae: 0.0101INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 9.8380e-05 - mae: 0.0102 - val_loss: 4.0414e-04 - val_mae: 0.0200\n",
            "Epoch 29/500\n",
            "42/43 [============================>.] - ETA: 0s - loss: 9.6481e-05 - mae: 0.0100INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 9.7127e-05 - mae: 0.0101 - val_loss: 3.9956e-04 - val_mae: 0.0199\n",
            "Epoch 30/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 9.0746e-05 - mae: 0.0096INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 9.6381e-05 - mae: 0.0100 - val_loss: 3.9822e-04 - val_mae: 0.0198\n",
            "Epoch 31/500\n",
            "37/43 [========================>.....] - ETA: 0s - loss: 9.0412e-05 - mae: 0.0095INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 9.5616e-05 - mae: 0.0099 - val_loss: 3.9581e-04 - val_mae: 0.0198\n",
            "Epoch 32/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 9.1131e-05 - mae: 0.0096INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 9.5084e-05 - mae: 0.0098 - val_loss: 3.9402e-04 - val_mae: 0.0197\n",
            "Epoch 33/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 8.9228e-05 - mae: 0.0094INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 26ms/step - loss: 9.4662e-05 - mae: 0.0097 - val_loss: 3.9190e-04 - val_mae: 0.0197\n",
            "Epoch 34/500\n",
            "39/43 [==========================>...] - ETA: 0s - loss: 9.1858e-05 - mae: 0.0095INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 21ms/step - loss: 9.4527e-05 - mae: 0.0097 - val_loss: 3.9167e-04 - val_mae: 0.0197\n",
            "Epoch 35/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 8.4189e-05 - mae: 0.0091INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 9.4178e-05 - mae: 0.0097 - val_loss: 3.9095e-04 - val_mae: 0.0197\n",
            "Epoch 36/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 9.1108e-05 - mae: 0.0094INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 9.4026e-05 - mae: 0.0096 - val_loss: 3.9048e-04 - val_mae: 0.0197\n",
            "Epoch 37/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 9.3857e-05 - mae: 0.0096 - val_loss: 3.9189e-04 - val_mae: 0.0197\n",
            "Epoch 38/500\n",
            "41/43 [===========================>..] - ETA: 0s - loss: 9.2022e-05 - mae: 0.0095INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 9.3635e-05 - mae: 0.0096 - val_loss: 3.8965e-04 - val_mae: 0.0197\n",
            "Epoch 39/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 9.3609e-05 - mae: 0.0096 - val_loss: 3.8967e-04 - val_mae: 0.0196\n",
            "Epoch 40/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 8.5776e-05 - mae: 0.0091INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 9.3430e-05 - mae: 0.0096 - val_loss: 3.8951e-04 - val_mae: 0.0196\n",
            "Epoch 41/500\n",
            "35/43 [=======================>......] - ETA: 0s - loss: 8.5766e-05 - mae: 0.0090INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 9.3308e-05 - mae: 0.0096 - val_loss: 3.8845e-04 - val_mae: 0.0196\n",
            "Epoch 42/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 9.3200e-05 - mae: 0.0096 - val_loss: 3.8865e-04 - val_mae: 0.0196\n",
            "Epoch 43/500\n",
            "39/43 [==========================>...] - ETA: 0s - loss: 8.9380e-05 - mae: 0.0093INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 9.3113e-05 - mae: 0.0096 - val_loss: 3.8748e-04 - val_mae: 0.0196\n",
            "Epoch 44/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 9.3002e-05 - mae: 0.0095 - val_loss: 3.8749e-04 - val_mae: 0.0197\n",
            "Epoch 45/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 9.0179e-05 - mae: 0.0094INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 9.2881e-05 - mae: 0.0095 - val_loss: 3.8664e-04 - val_mae: 0.0196\n",
            "Epoch 46/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 9.0629e-05 - mae: 0.0094INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 9.2753e-05 - mae: 0.0095 - val_loss: 3.8597e-04 - val_mae: 0.0196\n",
            "Epoch 47/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 9.2672e-05 - mae: 0.0095 - val_loss: 3.8606e-04 - val_mae: 0.0195\n",
            "Epoch 48/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 8.7467e-05 - mae: 0.0092INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 9.2540e-05 - mae: 0.0095 - val_loss: 3.8543e-04 - val_mae: 0.0196\n",
            "Epoch 49/500\n",
            "43/43 [==============================] - ETA: 0s - loss: 9.2482e-05 - mae: 0.0095INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 9.2482e-05 - mae: 0.0095 - val_loss: 3.8443e-04 - val_mae: 0.0195\n",
            "Epoch 50/500\n",
            "43/43 [==============================] - ETA: 0s - loss: 9.2285e-05 - mae: 0.0095INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 9.2285e-05 - mae: 0.0095 - val_loss: 3.8394e-04 - val_mae: 0.0195\n",
            "Epoch 51/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 9.2139e-05 - mae: 0.0095 - val_loss: 3.8423e-04 - val_mae: 0.0196\n",
            "Epoch 52/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 9.2029e-05 - mae: 0.0095 - val_loss: 3.8559e-04 - val_mae: 0.0195\n",
            "Epoch 53/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 9.0543e-05 - mae: 0.0094INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 9.2008e-05 - mae: 0.0095 - val_loss: 3.8254e-04 - val_mae: 0.0195\n",
            "Epoch 54/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 8.5256e-05 - mae: 0.0091INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 9.1773e-05 - mae: 0.0095 - val_loss: 3.8166e-04 - val_mae: 0.0195\n",
            "Epoch 55/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 9.0364e-05 - mae: 0.0094INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 9.1602e-05 - mae: 0.0095 - val_loss: 3.8120e-04 - val_mae: 0.0194\n",
            "Epoch 56/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 8.7426e-05 - mae: 0.0092INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 9.1489e-05 - mae: 0.0095 - val_loss: 3.8095e-04 - val_mae: 0.0194\n",
            "Epoch 57/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 9.1295e-05 - mae: 0.0095 - val_loss: 3.8148e-04 - val_mae: 0.0194\n",
            "Epoch 58/500\n",
            "42/43 [============================>.] - ETA: 0s - loss: 9.0805e-05 - mae: 0.0094INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 9.1172e-05 - mae: 0.0095 - val_loss: 3.7982e-04 - val_mae: 0.0195\n",
            "Epoch 59/500\n",
            "39/43 [==========================>...] - ETA: 0s - loss: 8.7581e-05 - mae: 0.0092INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 9.1163e-05 - mae: 0.0094 - val_loss: 3.7851e-04 - val_mae: 0.0194\n",
            "Epoch 60/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 9.0920e-05 - mae: 0.0094 - val_loss: 3.8002e-04 - val_mae: 0.0194\n",
            "Epoch 61/500\n",
            "42/43 [============================>.] - ETA: 0s - loss: 8.9704e-05 - mae: 0.0094INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 9.0782e-05 - mae: 0.0094 - val_loss: 3.7833e-04 - val_mae: 0.0193\n",
            "Epoch 62/500\n",
            "43/43 [==============================] - ETA: 0s - loss: 9.0561e-05 - mae: 0.0094INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 9.0561e-05 - mae: 0.0094 - val_loss: 3.7676e-04 - val_mae: 0.0193\n",
            "Epoch 63/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 8.3449e-05 - mae: 0.0089INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 9.0465e-05 - mae: 0.0094 - val_loss: 3.7668e-04 - val_mae: 0.0193\n",
            "Epoch 64/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 8.7583e-05 - mae: 0.0092INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 9.0306e-05 - mae: 0.0094 - val_loss: 3.7515e-04 - val_mae: 0.0193\n",
            "Epoch 65/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 9.0084e-05 - mae: 0.0094 - val_loss: 3.7544e-04 - val_mae: 0.0194\n",
            "Epoch 66/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 8.3413e-05 - mae: 0.0089INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 27ms/step - loss: 9.0045e-05 - mae: 0.0094 - val_loss: 3.7438e-04 - val_mae: 0.0193\n",
            "Epoch 67/500\n",
            "42/43 [============================>.] - ETA: 0s - loss: 8.8880e-05 - mae: 0.0093INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 8.9814e-05 - mae: 0.0094 - val_loss: 3.7314e-04 - val_mae: 0.0193\n",
            "Epoch 68/500\n",
            "39/43 [==========================>...] - ETA: 0s - loss: 8.5442e-05 - mae: 0.0091INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 21ms/step - loss: 8.9670e-05 - mae: 0.0094 - val_loss: 3.7222e-04 - val_mae: 0.0192\n",
            "Epoch 69/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 8.5386e-05 - mae: 0.0091INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 8.9465e-05 - mae: 0.0094 - val_loss: 3.7145e-04 - val_mae: 0.0192\n",
            "Epoch 70/500\n",
            "43/43 [==============================] - ETA: 0s - loss: 8.9263e-05 - mae: 0.0094INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 8.9263e-05 - mae: 0.0094 - val_loss: 3.7077e-04 - val_mae: 0.0192\n",
            "Epoch 71/500\n",
            "39/43 [==========================>...] - ETA: 0s - loss: 8.5926e-05 - mae: 0.0091INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 21ms/step - loss: 8.9051e-05 - mae: 0.0093 - val_loss: 3.7026e-04 - val_mae: 0.0192\n",
            "Epoch 72/500\n",
            "39/43 [==========================>...] - ETA: 0s - loss: 8.6600e-05 - mae: 0.0091INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 8.8899e-05 - mae: 0.0093 - val_loss: 3.6908e-04 - val_mae: 0.0191\n",
            "Epoch 73/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 8.3376e-05 - mae: 0.0090INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 8.8775e-05 - mae: 0.0093 - val_loss: 3.6827e-04 - val_mae: 0.0191\n",
            "Epoch 74/500\n",
            "37/43 [========================>.....] - ETA: 0s - loss: 8.5477e-05 - mae: 0.0090INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 8.8534e-05 - mae: 0.0093 - val_loss: 3.6750e-04 - val_mae: 0.0191\n",
            "Epoch 75/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 8.3429e-05 - mae: 0.0090INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 8.8279e-05 - mae: 0.0093 - val_loss: 3.6694e-04 - val_mae: 0.0191\n",
            "Epoch 76/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 8.4927e-05 - mae: 0.0090INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 21ms/step - loss: 8.8143e-05 - mae: 0.0093 - val_loss: 3.6575e-04 - val_mae: 0.0191\n",
            "Epoch 77/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 8.1961e-05 - mae: 0.0089INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 8.7869e-05 - mae: 0.0093 - val_loss: 3.6558e-04 - val_mae: 0.0191\n",
            "Epoch 78/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 8.7777e-05 - mae: 0.0093 - val_loss: 3.6715e-04 - val_mae: 0.0192\n",
            "Epoch 79/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 8.2550e-05 - mae: 0.0090INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 21ms/step - loss: 8.7625e-05 - mae: 0.0093 - val_loss: 3.6314e-04 - val_mae: 0.0190\n",
            "Epoch 80/500\n",
            "41/43 [===========================>..] - ETA: 0s - loss: 8.5537e-05 - mae: 0.0091INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 21ms/step - loss: 8.7342e-05 - mae: 0.0093 - val_loss: 3.6221e-04 - val_mae: 0.0190\n",
            "Epoch 81/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 8.7076e-05 - mae: 0.0092 - val_loss: 3.6409e-04 - val_mae: 0.0190\n",
            "Epoch 82/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 8.1851e-05 - mae: 0.0089INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 8.6960e-05 - mae: 0.0092 - val_loss: 3.6034e-04 - val_mae: 0.0189\n",
            "Epoch 83/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 8.6747e-05 - mae: 0.0092 - val_loss: 3.6176e-04 - val_mae: 0.0190\n",
            "Epoch 84/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 8.6388e-05 - mae: 0.0092 - val_loss: 3.6298e-04 - val_mae: 0.0191\n",
            "Epoch 85/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 8.6294e-05 - mae: 0.0092 - val_loss: 3.6152e-04 - val_mae: 0.0191\n",
            "Epoch 86/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 8.0329e-05 - mae: 0.0088INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 8.6096e-05 - mae: 0.0092 - val_loss: 3.5655e-04 - val_mae: 0.0188\n",
            "Epoch 87/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 8.3727e-05 - mae: 0.0090INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 8.5895e-05 - mae: 0.0092 - val_loss: 3.5573e-04 - val_mae: 0.0188\n",
            "Epoch 88/500\n",
            "43/43 [==============================] - ETA: 0s - loss: 8.5606e-05 - mae: 0.0092INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 8.5606e-05 - mae: 0.0092 - val_loss: 3.5469e-04 - val_mae: 0.0188\n",
            "Epoch 89/500\n",
            "39/43 [==========================>...] - ETA: 0s - loss: 8.1223e-05 - mae: 0.0089INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 8.5349e-05 - mae: 0.0092 - val_loss: 3.5431e-04 - val_mae: 0.0187\n",
            "Epoch 90/500\n",
            "41/43 [===========================>..] - ETA: 0s - loss: 8.4190e-05 - mae: 0.0090INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 8.5227e-05 - mae: 0.0091 - val_loss: 3.5408e-04 - val_mae: 0.0187\n",
            "Epoch 91/500\n",
            "43/43 [==============================] - ETA: 0s - loss: 8.5113e-05 - mae: 0.0091INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 8.5113e-05 - mae: 0.0091 - val_loss: 3.5194e-04 - val_mae: 0.0187\n",
            "Epoch 92/500\n",
            "42/43 [============================>.] - ETA: 0s - loss: 8.4030e-05 - mae: 0.0091INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 8.4657e-05 - mae: 0.0091 - val_loss: 3.5125e-04 - val_mae: 0.0187\n",
            "Epoch 93/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 8.4404e-05 - mae: 0.0091 - val_loss: 3.5192e-04 - val_mae: 0.0187\n",
            "Epoch 94/500\n",
            "41/43 [===========================>..] - ETA: 0s - loss: 8.3087e-05 - mae: 0.0090INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 26ms/step - loss: 8.4431e-05 - mae: 0.0091 - val_loss: 3.4819e-04 - val_mae: 0.0186\n",
            "Epoch 95/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 8.3890e-05 - mae: 0.0091 - val_loss: 3.5367e-04 - val_mae: 0.0189\n",
            "Epoch 96/500\n",
            "42/43 [============================>.] - ETA: 0s - loss: 8.3984e-05 - mae: 0.0091INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 8.3713e-05 - mae: 0.0091 - val_loss: 3.4606e-04 - val_mae: 0.0185\n",
            "Epoch 97/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 8.3382e-05 - mae: 0.0091 - val_loss: 3.4761e-04 - val_mae: 0.0186\n",
            "Epoch 98/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 7.7595e-05 - mae: 0.0087INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 8.3230e-05 - mae: 0.0090 - val_loss: 3.4386e-04 - val_mae: 0.0184\n",
            "Epoch 99/500\n",
            "42/43 [============================>.] - ETA: 0s - loss: 8.2816e-05 - mae: 0.0090INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 8.2821e-05 - mae: 0.0090 - val_loss: 3.4266e-04 - val_mae: 0.0184\n",
            "Epoch 100/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 8.2555e-05 - mae: 0.0090 - val_loss: 3.4298e-04 - val_mae: 0.0184\n",
            "Epoch 101/500\n",
            "39/43 [==========================>...] - ETA: 0s - loss: 7.9975e-05 - mae: 0.0088INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 8.2365e-05 - mae: 0.0090 - val_loss: 3.4043e-04 - val_mae: 0.0183\n",
            "Epoch 102/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 7.6072e-05 - mae: 0.0085INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 8.2103e-05 - mae: 0.0090 - val_loss: 3.3911e-04 - val_mae: 0.0183\n",
            "Epoch 103/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 7.7725e-05 - mae: 0.0087INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 8.1865e-05 - mae: 0.0090 - val_loss: 3.3838e-04 - val_mae: 0.0183\n",
            "Epoch 104/500\n",
            "43/43 [==============================] - ETA: 0s - loss: 8.1661e-05 - mae: 0.0090INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 8.1661e-05 - mae: 0.0090 - val_loss: 3.3671e-04 - val_mae: 0.0182\n",
            "Epoch 105/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 8.1339e-05 - mae: 0.0089 - val_loss: 3.3864e-04 - val_mae: 0.0184\n",
            "Epoch 106/500\n",
            "41/43 [===========================>..] - ETA: 0s - loss: 7.9305e-05 - mae: 0.0088INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 8.1028e-05 - mae: 0.0089 - val_loss: 3.3635e-04 - val_mae: 0.0183\n",
            "Epoch 107/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 7.6502e-05 - mae: 0.0086INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 8.0825e-05 - mae: 0.0089 - val_loss: 3.3339e-04 - val_mae: 0.0181\n",
            "Epoch 108/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 7.3979e-05 - mae: 0.0084INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 8.0423e-05 - mae: 0.0089 - val_loss: 3.3280e-04 - val_mae: 0.0181\n",
            "Epoch 109/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 7.7069e-05 - mae: 0.0086INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 8.0279e-05 - mae: 0.0089 - val_loss: 3.3164e-04 - val_mae: 0.0181\n",
            "Epoch 110/500\n",
            "43/43 [==============================] - ETA: 0s - loss: 7.9809e-05 - mae: 0.0089INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 7.9809e-05 - mae: 0.0089 - val_loss: 3.2952e-04 - val_mae: 0.0181\n",
            "Epoch 111/500\n",
            "39/43 [==========================>...] - ETA: 0s - loss: 7.6811e-05 - mae: 0.0086INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 24ms/step - loss: 7.9611e-05 - mae: 0.0088 - val_loss: 3.2801e-04 - val_mae: 0.0180\n",
            "Epoch 112/500\n",
            "43/43 [==============================] - ETA: 0s - loss: 7.9311e-05 - mae: 0.0088INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 7.9311e-05 - mae: 0.0088 - val_loss: 3.2753e-04 - val_mae: 0.0180\n",
            "Epoch 113/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 7.9010e-05 - mae: 0.0088 - val_loss: 3.2913e-04 - val_mae: 0.0181\n",
            "Epoch 114/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 7.8828e-05 - mae: 0.0088 - val_loss: 3.2773e-04 - val_mae: 0.0181\n",
            "Epoch 115/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 7.8595e-05 - mae: 0.0088 - val_loss: 3.2876e-04 - val_mae: 0.0181\n",
            "Epoch 116/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 7.6205e-05 - mae: 0.0086INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 7.8268e-05 - mae: 0.0088 - val_loss: 3.2445e-04 - val_mae: 0.0180\n",
            "Epoch 117/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 7.7895e-05 - mae: 0.0088 - val_loss: 3.2644e-04 - val_mae: 0.0181\n",
            "Epoch 118/500\n",
            "41/43 [===========================>..] - ETA: 0s - loss: 7.7066e-05 - mae: 0.0087INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 7.7762e-05 - mae: 0.0087 - val_loss: 3.2083e-04 - val_mae: 0.0178\n",
            "Epoch 119/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 7.7540e-05 - mae: 0.0087 - val_loss: 3.4483e-04 - val_mae: 0.0189\n",
            "Epoch 120/500\n",
            "39/43 [==========================>...] - ETA: 0s - loss: 7.4769e-05 - mae: 0.0085INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 24ms/step - loss: 7.7466e-05 - mae: 0.0087 - val_loss: 3.1622e-04 - val_mae: 0.0177\n",
            "Epoch 121/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 7.6675e-05 - mae: 0.0087 - val_loss: 3.1946e-04 - val_mae: 0.0178\n",
            "Epoch 122/500\n",
            "39/43 [==========================>...] - ETA: 0s - loss: 7.4435e-05 - mae: 0.0085INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 25ms/step - loss: 7.6775e-05 - mae: 0.0087 - val_loss: 3.1496e-04 - val_mae: 0.0177\n",
            "Epoch 123/500\n",
            "37/43 [========================>.....] - ETA: 0s - loss: 6.9928e-05 - mae: 0.0082INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 7.6076e-05 - mae: 0.0087 - val_loss: 3.1343e-04 - val_mae: 0.0176\n",
            "Epoch 124/500\n",
            "41/43 [===========================>..] - ETA: 0s - loss: 7.4338e-05 - mae: 0.0086INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 7.5927e-05 - mae: 0.0087 - val_loss: 3.1096e-04 - val_mae: 0.0175\n",
            "Epoch 125/500\n",
            "43/43 [==============================] - ETA: 0s - loss: 7.5660e-05 - mae: 0.0086INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 7.5660e-05 - mae: 0.0086 - val_loss: 3.1030e-04 - val_mae: 0.0175\n",
            "Epoch 126/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 7.5365e-05 - mae: 0.0086 - val_loss: 3.1069e-04 - val_mae: 0.0175\n",
            "Epoch 127/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 6.8435e-05 - mae: 0.0081INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 7.4861e-05 - mae: 0.0086 - val_loss: 3.1030e-04 - val_mae: 0.0176\n",
            "Epoch 128/500\n",
            "43/43 [==============================] - ETA: 0s - loss: 7.4566e-05 - mae: 0.0086INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 7.4566e-05 - mae: 0.0086 - val_loss: 3.0837e-04 - val_mae: 0.0175\n",
            "Epoch 129/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 7.0294e-05 - mae: 0.0083INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 7.4564e-05 - mae: 0.0086 - val_loss: 3.0403e-04 - val_mae: 0.0173\n",
            "Epoch 130/500\n",
            "39/43 [==========================>...] - ETA: 0s - loss: 7.1271e-05 - mae: 0.0083INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 7.3965e-05 - mae: 0.0085 - val_loss: 3.0320e-04 - val_mae: 0.0173\n",
            "Epoch 131/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 7.3568e-05 - mae: 0.0085 - val_loss: 3.1054e-04 - val_mae: 0.0177\n",
            "Epoch 132/500\n",
            "42/43 [============================>.] - ETA: 0s - loss: 7.3592e-05 - mae: 0.0085INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 7.3402e-05 - mae: 0.0085 - val_loss: 3.0069e-04 - val_mae: 0.0172\n",
            "Epoch 133/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 7.3162e-05 - mae: 0.0085 - val_loss: 3.0446e-04 - val_mae: 0.0174\n",
            "Epoch 134/500\n",
            "42/43 [============================>.] - ETA: 0s - loss: 7.2309e-05 - mae: 0.0084INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 7.2770e-05 - mae: 0.0085 - val_loss: 2.9738e-04 - val_mae: 0.0172\n",
            "Epoch 135/500\n",
            "41/43 [===========================>..] - ETA: 0s - loss: 7.0795e-05 - mae: 0.0083INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 27ms/step - loss: 7.2600e-05 - mae: 0.0085 - val_loss: 2.9526e-04 - val_mae: 0.0170\n",
            "Epoch 136/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 7.2013e-05 - mae: 0.0084 - val_loss: 2.9744e-04 - val_mae: 0.0172\n",
            "Epoch 137/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 7.1774e-05 - mae: 0.0084 - val_loss: 2.9763e-04 - val_mae: 0.0172\n",
            "Epoch 138/500\n",
            "39/43 [==========================>...] - ETA: 0s - loss: 6.7922e-05 - mae: 0.0082INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 7.1660e-05 - mae: 0.0084 - val_loss: 2.9088e-04 - val_mae: 0.0169\n",
            "Epoch 139/500\n",
            "42/43 [============================>.] - ETA: 0s - loss: 7.0485e-05 - mae: 0.0083INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 7.1005e-05 - mae: 0.0084 - val_loss: 2.8939e-04 - val_mae: 0.0169\n",
            "Epoch 140/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 6.7687e-05 - mae: 0.0081INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 21ms/step - loss: 7.0933e-05 - mae: 0.0084 - val_loss: 2.8883e-04 - val_mae: 0.0169\n",
            "Epoch 141/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 7.0527e-05 - mae: 0.0083 - val_loss: 2.9000e-04 - val_mae: 0.0169\n",
            "Epoch 142/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 6.8368e-05 - mae: 0.0081INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 7.0383e-05 - mae: 0.0083 - val_loss: 2.8555e-04 - val_mae: 0.0168\n",
            "Epoch 143/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 6.9950e-05 - mae: 0.0083 - val_loss: 2.9368e-04 - val_mae: 0.0171\n",
            "Epoch 144/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 6.9798e-05 - mae: 0.0083 - val_loss: 2.8724e-04 - val_mae: 0.0170\n",
            "Epoch 145/500\n",
            "37/43 [========================>.....] - ETA: 0s - loss: 6.4706e-05 - mae: 0.0079INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 6.9399e-05 - mae: 0.0083 - val_loss: 2.8326e-04 - val_mae: 0.0167\n",
            "Epoch 146/500\n",
            "41/43 [===========================>..] - ETA: 0s - loss: 6.8882e-05 - mae: 0.0082INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 6.9608e-05 - mae: 0.0083 - val_loss: 2.7929e-04 - val_mae: 0.0166\n",
            "Epoch 147/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 6.8853e-05 - mae: 0.0082 - val_loss: 2.9156e-04 - val_mae: 0.0171\n",
            "Epoch 148/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 6.8231e-05 - mae: 0.0082 - val_loss: 2.8302e-04 - val_mae: 0.0167\n",
            "Epoch 149/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 6.8151e-05 - mae: 0.0082 - val_loss: 2.8341e-04 - val_mae: 0.0167\n",
            "Epoch 150/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 6.7882e-05 - mae: 0.0082 - val_loss: 2.9096e-04 - val_mae: 0.0174\n",
            "Epoch 151/500\n",
            "43/43 [==============================] - ETA: 0s - loss: 6.8475e-05 - mae: 0.0082INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 6.8475e-05 - mae: 0.0082 - val_loss: 2.7216e-04 - val_mae: 0.0163\n",
            "Epoch 152/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 6.5778e-05 - mae: 0.0080INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 6.7229e-05 - mae: 0.0081 - val_loss: 2.7069e-04 - val_mae: 0.0163\n",
            "Epoch 153/500\n",
            "35/43 [=======================>......] - ETA: 0s - loss: 6.0627e-05 - mae: 0.0076INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 21ms/step - loss: 6.6961e-05 - mae: 0.0081 - val_loss: 2.7025e-04 - val_mae: 0.0164\n",
            "Epoch 154/500\n",
            "37/43 [========================>.....] - ETA: 0s - loss: 6.4650e-05 - mae: 0.0079INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 6.6499e-05 - mae: 0.0081 - val_loss: 2.6776e-04 - val_mae: 0.0162\n",
            "Epoch 155/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 6.6518e-05 - mae: 0.0081 - val_loss: 2.7196e-04 - val_mae: 0.0166\n",
            "Epoch 156/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 6.1882e-05 - mae: 0.0077INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 6.6656e-05 - mae: 0.0081 - val_loss: 2.6533e-04 - val_mae: 0.0161\n",
            "Epoch 157/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 6.5654e-05 - mae: 0.0080 - val_loss: 2.6588e-04 - val_mae: 0.0161\n",
            "Epoch 158/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 6.1697e-05 - mae: 0.0077INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 6.5304e-05 - mae: 0.0080 - val_loss: 2.6519e-04 - val_mae: 0.0161\n",
            "Epoch 159/500\n",
            "43/43 [==============================] - ETA: 0s - loss: 6.5151e-05 - mae: 0.0080INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 6.5151e-05 - mae: 0.0080 - val_loss: 2.6067e-04 - val_mae: 0.0160\n",
            "Epoch 160/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 6.4971e-05 - mae: 0.0080 - val_loss: 2.8013e-04 - val_mae: 0.0172\n",
            "Epoch 161/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 6.3228e-05 - mae: 0.0079INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 6.4746e-05 - mae: 0.0080 - val_loss: 2.5876e-04 - val_mae: 0.0160\n",
            "Epoch 162/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 6.1516e-05 - mae: 0.0077INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 21ms/step - loss: 6.4187e-05 - mae: 0.0080 - val_loss: 2.5756e-04 - val_mae: 0.0159\n",
            "Epoch 163/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 6.3805e-05 - mae: 0.0079 - val_loss: 2.5912e-04 - val_mae: 0.0159\n",
            "Epoch 164/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 6.0229e-05 - mae: 0.0076INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 6.3858e-05 - mae: 0.0079 - val_loss: 2.5408e-04 - val_mae: 0.0158\n",
            "Epoch 165/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 5.7819e-05 - mae: 0.0074INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 6.3178e-05 - mae: 0.0079 - val_loss: 2.5338e-04 - val_mae: 0.0159\n",
            "Epoch 166/500\n",
            "41/43 [===========================>..] - ETA: 0s - loss: 6.2732e-05 - mae: 0.0079INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 6.3623e-05 - mae: 0.0079 - val_loss: 2.5103e-04 - val_mae: 0.0157\n",
            "Epoch 167/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 6.2931e-05 - mae: 0.0079 - val_loss: 2.5884e-04 - val_mae: 0.0160\n",
            "Epoch 168/500\n",
            "43/43 [==============================] - 1s 18ms/step - loss: 6.2587e-05 - mae: 0.0079 - val_loss: 2.5796e-04 - val_mae: 0.0159\n",
            "Epoch 169/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 6.2146e-05 - mae: 0.0078 - val_loss: 2.6965e-04 - val_mae: 0.0164\n",
            "Epoch 170/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 6.2031e-05 - mae: 0.0078 - val_loss: 2.7356e-04 - val_mae: 0.0166\n",
            "Epoch 171/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 5.8092e-05 - mae: 0.0075INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 6.1573e-05 - mae: 0.0078 - val_loss: 2.4710e-04 - val_mae: 0.0158\n",
            "Epoch 172/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 5.8818e-05 - mae: 0.0076INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 24ms/step - loss: 6.0955e-05 - mae: 0.0077 - val_loss: 2.4423e-04 - val_mae: 0.0154\n",
            "Epoch 173/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 6.2539e-05 - mae: 0.0078 - val_loss: 2.6795e-04 - val_mae: 0.0170\n",
            "Epoch 174/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 6.1155e-05 - mae: 0.0078 - val_loss: 2.4625e-04 - val_mae: 0.0158\n",
            "Epoch 175/500\n",
            "41/43 [===========================>..] - ETA: 0s - loss: 6.1881e-05 - mae: 0.0078INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 25ms/step - loss: 6.2858e-05 - mae: 0.0079 - val_loss: 2.4243e-04 - val_mae: 0.0154\n",
            "Epoch 176/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 6.0344e-05 - mae: 0.0077 - val_loss: 2.4676e-04 - val_mae: 0.0159\n",
            "Epoch 177/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 5.3977e-05 - mae: 0.0072INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 6.0162e-05 - mae: 0.0077 - val_loss: 2.3919e-04 - val_mae: 0.0153\n",
            "Epoch 178/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 6.3011e-05 - mae: 0.0079 - val_loss: 2.9291e-04 - val_mae: 0.0185\n",
            "Epoch 179/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 5.9457e-05 - mae: 0.0077 - val_loss: 2.4332e-04 - val_mae: 0.0159\n",
            "Epoch 180/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 5.8875e-05 - mae: 0.0076 - val_loss: 2.5237e-04 - val_mae: 0.0158\n",
            "Epoch 181/500\n",
            "43/43 [==============================] - ETA: 0s - loss: 5.9029e-05 - mae: 0.0076INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 5.9029e-05 - mae: 0.0076 - val_loss: 2.3471e-04 - val_mae: 0.0154\n",
            "Epoch 182/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 5.5980e-05 - mae: 0.0074INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 5.8184e-05 - mae: 0.0076 - val_loss: 2.3322e-04 - val_mae: 0.0151\n",
            "Epoch 183/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 6.0485e-05 - mae: 0.0077 - val_loss: 2.9808e-04 - val_mae: 0.0189\n",
            "Epoch 184/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 5.3600e-05 - mae: 0.0072INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 5.8155e-05 - mae: 0.0076 - val_loss: 2.2838e-04 - val_mae: 0.0149\n",
            "Epoch 185/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 5.9666e-05 - mae: 0.0077 - val_loss: 2.4330e-04 - val_mae: 0.0155\n",
            "Epoch 186/500\n",
            "41/43 [===========================>..] - ETA: 0s - loss: 5.6623e-05 - mae: 0.0074INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 5.7385e-05 - mae: 0.0075 - val_loss: 2.2557e-04 - val_mae: 0.0149\n",
            "Epoch 187/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 5.8340e-05 - mae: 0.0076 - val_loss: 2.6844e-04 - val_mae: 0.0175\n",
            "Epoch 188/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 5.7353e-05 - mae: 0.0075 - val_loss: 2.3129e-04 - val_mae: 0.0154\n",
            "Epoch 189/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 5.6811e-05 - mae: 0.0075 - val_loss: 2.2768e-04 - val_mae: 0.0149\n",
            "Epoch 190/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 5.6029e-05 - mae: 0.0074 - val_loss: 2.3228e-04 - val_mae: 0.0156\n",
            "Epoch 191/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 5.6284e-05 - mae: 0.0075 - val_loss: 2.3720e-04 - val_mae: 0.0153\n",
            "Epoch 192/500\n",
            "43/43 [==============================] - ETA: 0s - loss: 5.5731e-05 - mae: 0.0074INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 27ms/step - loss: 5.5731e-05 - mae: 0.0074 - val_loss: 2.2054e-04 - val_mae: 0.0149\n",
            "Epoch 193/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 5.6154e-05 - mae: 0.0074 - val_loss: 2.4649e-04 - val_mae: 0.0165\n",
            "Epoch 194/500\n",
            "42/43 [============================>.] - ETA: 0s - loss: 5.4952e-05 - mae: 0.0074INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 5.5185e-05 - mae: 0.0074 - val_loss: 2.1657e-04 - val_mae: 0.0146\n",
            "Epoch 195/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 5.5173e-05 - mae: 0.0073 - val_loss: 2.3962e-04 - val_mae: 0.0162\n",
            "Epoch 196/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 5.2067e-05 - mae: 0.0071INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 5.5333e-05 - mae: 0.0074 - val_loss: 2.1618e-04 - val_mae: 0.0147\n",
            "Epoch 197/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 5.4878e-05 - mae: 0.0074 - val_loss: 2.1966e-04 - val_mae: 0.0150\n",
            "Epoch 198/500\n",
            "35/43 [=======================>......] - ETA: 0s - loss: 5.0174e-05 - mae: 0.0069INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 5.5135e-05 - mae: 0.0074 - val_loss: 2.1453e-04 - val_mae: 0.0147\n",
            "Epoch 199/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 5.4115e-05 - mae: 0.0073 - val_loss: 2.4200e-04 - val_mae: 0.0156\n",
            "Epoch 200/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 5.6442e-05 - mae: 0.0075 - val_loss: 2.5479e-04 - val_mae: 0.0162\n",
            "Epoch 201/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 5.7835e-05 - mae: 0.0075 - val_loss: 2.5346e-04 - val_mae: 0.0171\n",
            "Epoch 202/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 5.5962e-05 - mae: 0.0074 - val_loss: 2.2102e-04 - val_mae: 0.0147\n",
            "Epoch 203/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 5.4559e-05 - mae: 0.0073 - val_loss: 2.1696e-04 - val_mae: 0.0145\n",
            "Epoch 204/500\n",
            "43/43 [==============================] - 1s 18ms/step - loss: 5.2963e-05 - mae: 0.0072 - val_loss: 2.2291e-04 - val_mae: 0.0155\n",
            "Epoch 205/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 5.0813e-05 - mae: 0.0070INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 5.2864e-05 - mae: 0.0072 - val_loss: 2.0807e-04 - val_mae: 0.0145\n",
            "Epoch 206/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 5.2336e-05 - mae: 0.0072 - val_loss: 2.3155e-04 - val_mae: 0.0152\n",
            "Epoch 207/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 5.0744e-05 - mae: 0.0070INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 5.4439e-05 - mae: 0.0073 - val_loss: 2.0692e-04 - val_mae: 0.0145\n",
            "Epoch 208/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 5.2436e-05 - mae: 0.0072 - val_loss: 2.1826e-04 - val_mae: 0.0153\n",
            "Epoch 209/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 5.0810e-05 - mae: 0.0070INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 5.1949e-05 - mae: 0.0071 - val_loss: 2.0280e-04 - val_mae: 0.0142\n",
            "Epoch 210/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 5.1746e-05 - mae: 0.0071 - val_loss: 2.1789e-04 - val_mae: 0.0146\n",
            "Epoch 211/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 5.1654e-05 - mae: 0.0071 - val_loss: 2.0736e-04 - val_mae: 0.0142\n",
            "Epoch 212/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 5.0787e-05 - mae: 0.0070 - val_loss: 2.0975e-04 - val_mae: 0.0143\n",
            "Epoch 213/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 4.6345e-05 - mae: 0.0066INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 21ms/step - loss: 5.2460e-05 - mae: 0.0072 - val_loss: 1.9844e-04 - val_mae: 0.0139\n",
            "Epoch 214/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 5.1495e-05 - mae: 0.0071 - val_loss: 2.3525e-04 - val_mae: 0.0155\n",
            "Epoch 215/500\n",
            "42/43 [============================>.] - ETA: 0s - loss: 5.0087e-05 - mae: 0.0070INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 5.0529e-05 - mae: 0.0070 - val_loss: 1.9611e-04 - val_mae: 0.0139\n",
            "Epoch 216/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 5.0034e-05 - mae: 0.0070 - val_loss: 2.0620e-04 - val_mae: 0.0142\n",
            "Epoch 217/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 5.1606e-05 - mae: 0.0071 - val_loss: 1.9775e-04 - val_mae: 0.0142\n",
            "Epoch 218/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 4.9690e-05 - mae: 0.0070 - val_loss: 2.0010e-04 - val_mae: 0.0139\n",
            "Epoch 219/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 4.6576e-05 - mae: 0.0067INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 5.1369e-05 - mae: 0.0071 - val_loss: 1.9407e-04 - val_mae: 0.0139\n",
            "Epoch 220/500\n",
            "41/43 [===========================>..] - ETA: 0s - loss: 4.8645e-05 - mae: 0.0069INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 4.9852e-05 - mae: 0.0070 - val_loss: 1.9206e-04 - val_mae: 0.0137\n",
            "Epoch 221/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 4.9115e-05 - mae: 0.0069 - val_loss: 1.9779e-04 - val_mae: 0.0138\n",
            "Epoch 222/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 4.6076e-05 - mae: 0.0066INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 4.9118e-05 - mae: 0.0069 - val_loss: 1.9028e-04 - val_mae: 0.0136\n",
            "Epoch 223/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 4.8684e-05 - mae: 0.0069 - val_loss: 2.8986e-04 - val_mae: 0.0194\n",
            "Epoch 224/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 4.6005e-05 - mae: 0.0067INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 4.9184e-05 - mae: 0.0070 - val_loss: 1.8973e-04 - val_mae: 0.0135\n",
            "Epoch 225/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 4.5212e-05 - mae: 0.0065INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 4.8244e-05 - mae: 0.0069 - val_loss: 1.8841e-04 - val_mae: 0.0135\n",
            "Epoch 226/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 4.8445e-05 - mae: 0.0069 - val_loss: 1.9124e-04 - val_mae: 0.0140\n",
            "Epoch 227/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 5.2274e-05 - mae: 0.0072 - val_loss: 2.4963e-04 - val_mae: 0.0175\n",
            "Epoch 228/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 4.6272e-05 - mae: 0.0068INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 4.8266e-05 - mae: 0.0069 - val_loss: 1.8545e-04 - val_mae: 0.0134\n",
            "Epoch 229/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 4.3813e-05 - mae: 0.0065INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 4.7855e-05 - mae: 0.0068 - val_loss: 1.8457e-04 - val_mae: 0.0134\n",
            "Epoch 230/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 4.7975e-05 - mae: 0.0069 - val_loss: 1.9418e-04 - val_mae: 0.0143\n",
            "Epoch 231/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 4.8218e-05 - mae: 0.0069 - val_loss: 1.9631e-04 - val_mae: 0.0138\n",
            "Epoch 232/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 4.8904e-05 - mae: 0.0069 - val_loss: 1.9952e-04 - val_mae: 0.0147\n",
            "Epoch 233/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 5.0943e-05 - mae: 0.0071 - val_loss: 1.8608e-04 - val_mae: 0.0134\n",
            "Epoch 234/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 4.6741e-05 - mae: 0.0067 - val_loss: 1.8568e-04 - val_mae: 0.0138\n",
            "Epoch 235/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 4.7015e-05 - mae: 0.0068 - val_loss: 2.4237e-04 - val_mae: 0.0161\n",
            "Epoch 236/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 4.6588e-05 - mae: 0.0068 - val_loss: 2.0925e-04 - val_mae: 0.0145\n",
            "Epoch 237/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 4.3539e-05 - mae: 0.0065INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 4.6083e-05 - mae: 0.0067 - val_loss: 1.8242e-04 - val_mae: 0.0136\n",
            "Epoch 238/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 4.2540e-05 - mae: 0.0064INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 4.6371e-05 - mae: 0.0067 - val_loss: 1.7785e-04 - val_mae: 0.0132\n",
            "Epoch 239/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 4.5380e-05 - mae: 0.0066 - val_loss: 1.7804e-04 - val_mae: 0.0131\n",
            "Epoch 240/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 4.4972e-05 - mae: 0.0066INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 4.5925e-05 - mae: 0.0067 - val_loss: 1.7653e-04 - val_mae: 0.0132\n",
            "Epoch 241/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 4.7692e-05 - mae: 0.0068 - val_loss: 3.1488e-04 - val_mae: 0.0208\n",
            "Epoch 242/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 4.6353e-05 - mae: 0.0068 - val_loss: 1.7881e-04 - val_mae: 0.0131\n",
            "Epoch 243/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 4.5524e-05 - mae: 0.0067 - val_loss: 2.1724e-04 - val_mae: 0.0150\n",
            "Epoch 244/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 4.6898e-05 - mae: 0.0068 - val_loss: 1.8724e-04 - val_mae: 0.0135\n",
            "Epoch 245/500\n",
            "42/43 [============================>.] - ETA: 0s - loss: 4.3828e-05 - mae: 0.0065INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 4.4717e-05 - mae: 0.0066 - val_loss: 1.7410e-04 - val_mae: 0.0132\n",
            "Epoch 246/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 4.4450e-05 - mae: 0.0065INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 24ms/step - loss: 4.6244e-05 - mae: 0.0067 - val_loss: 1.7233e-04 - val_mae: 0.0129\n",
            "Epoch 247/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 4.4736e-05 - mae: 0.0066 - val_loss: 1.7298e-04 - val_mae: 0.0131\n",
            "Epoch 248/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 4.5378e-05 - mae: 0.0067 - val_loss: 2.2261e-04 - val_mae: 0.0164\n",
            "Epoch 249/500\n",
            "35/43 [=======================>......] - ETA: 0s - loss: 4.2324e-05 - mae: 0.0064INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 4.7760e-05 - mae: 0.0069 - val_loss: 1.7021e-04 - val_mae: 0.0129\n",
            "Epoch 250/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 4.2741e-05 - mae: 0.0064INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 4.4991e-05 - mae: 0.0066 - val_loss: 1.7009e-04 - val_mae: 0.0128\n",
            "Epoch 251/500\n",
            "37/43 [========================>.....] - ETA: 0s - loss: 4.1901e-05 - mae: 0.0063INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 4.4540e-05 - mae: 0.0066 - val_loss: 1.6909e-04 - val_mae: 0.0129\n",
            "Epoch 252/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 4.2471e-05 - mae: 0.0064INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 4.3461e-05 - mae: 0.0065 - val_loss: 1.6847e-04 - val_mae: 0.0128\n",
            "Epoch 253/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 4.1956e-05 - mae: 0.0063INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 4.3836e-05 - mae: 0.0065 - val_loss: 1.6785e-04 - val_mae: 0.0128\n",
            "Epoch 254/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 4.0707e-05 - mae: 0.0062INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 27ms/step - loss: 4.3557e-05 - mae: 0.0065 - val_loss: 1.6771e-04 - val_mae: 0.0127\n",
            "Epoch 255/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 4.3178e-05 - mae: 0.0065 - val_loss: 1.8856e-04 - val_mae: 0.0145\n",
            "Epoch 256/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 4.3389e-05 - mae: 0.0065 - val_loss: 1.8161e-04 - val_mae: 0.0133\n",
            "Epoch 257/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 4.3043e-05 - mae: 0.0065 - val_loss: 1.9741e-04 - val_mae: 0.0141\n",
            "Epoch 258/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 4.2060e-05 - mae: 0.0064INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 24ms/step - loss: 4.2862e-05 - mae: 0.0065 - val_loss: 1.6548e-04 - val_mae: 0.0126\n",
            "Epoch 259/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 4.2761e-05 - mae: 0.0064 - val_loss: 1.7941e-04 - val_mae: 0.0132\n",
            "Epoch 260/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 4.3201e-05 - mae: 0.0065 - val_loss: 2.1698e-04 - val_mae: 0.0152\n",
            "Epoch 261/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 4.2800e-05 - mae: 0.0065 - val_loss: 1.7083e-04 - val_mae: 0.0133\n",
            "Epoch 262/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 4.2621e-05 - mae: 0.0065 - val_loss: 1.7961e-04 - val_mae: 0.0140\n",
            "Epoch 263/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 4.3345e-05 - mae: 0.0065 - val_loss: 1.7090e-04 - val_mae: 0.0134\n",
            "Epoch 264/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 4.1979e-05 - mae: 0.0064 - val_loss: 1.8723e-04 - val_mae: 0.0137\n",
            "Epoch 265/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 4.1734e-05 - mae: 0.0064 - val_loss: 1.7145e-04 - val_mae: 0.0135\n",
            "Epoch 266/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 4.0149e-05 - mae: 0.0062INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 4.1811e-05 - mae: 0.0064 - val_loss: 1.6547e-04 - val_mae: 0.0130\n",
            "Epoch 267/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 4.0005e-05 - mae: 0.0062INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 4.3860e-05 - mae: 0.0065 - val_loss: 1.6489e-04 - val_mae: 0.0126\n",
            "Epoch 268/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 4.0888e-05 - mae: 0.0063INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 4.1414e-05 - mae: 0.0063 - val_loss: 1.6220e-04 - val_mae: 0.0128\n",
            "Epoch 269/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 4.1418e-05 - mae: 0.0064 - val_loss: 1.6740e-04 - val_mae: 0.0127\n",
            "Epoch 270/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 4.1235e-05 - mae: 0.0063 - val_loss: 2.5506e-04 - val_mae: 0.0171\n",
            "Epoch 271/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 4.6701e-05 - mae: 0.0068 - val_loss: 1.7803e-04 - val_mae: 0.0141\n",
            "Epoch 272/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 3.9013e-05 - mae: 0.0061INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 4.1016e-05 - mae: 0.0063 - val_loss: 1.5786e-04 - val_mae: 0.0123\n",
            "Epoch 273/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 4.0490e-05 - mae: 0.0063 - val_loss: 1.5896e-04 - val_mae: 0.0123\n",
            "Epoch 274/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 4.0528e-05 - mae: 0.0063 - val_loss: 1.6885e-04 - val_mae: 0.0135\n",
            "Epoch 275/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 3.7103e-05 - mae: 0.0060INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 4.0654e-05 - mae: 0.0063 - val_loss: 1.5503e-04 - val_mae: 0.0123\n",
            "Epoch 276/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 4.1125e-05 - mae: 0.0063 - val_loss: 1.9318e-04 - val_mae: 0.0141\n",
            "Epoch 277/500\n",
            "43/43 [==============================] - 1s 18ms/step - loss: 4.0960e-05 - mae: 0.0063 - val_loss: 1.6986e-04 - val_mae: 0.0129\n",
            "Epoch 278/500\n",
            "41/43 [===========================>..] - ETA: 0s - loss: 4.1982e-05 - mae: 0.0064INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 4.2215e-05 - mae: 0.0064 - val_loss: 1.5381e-04 - val_mae: 0.0121\n",
            "Epoch 279/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 3.9915e-05 - mae: 0.0062 - val_loss: 1.6034e-04 - val_mae: 0.0129\n",
            "Epoch 280/500\n",
            "41/43 [===========================>..] - ETA: 0s - loss: 4.1654e-05 - mae: 0.0064INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 4.2074e-05 - mae: 0.0064 - val_loss: 1.5377e-04 - val_mae: 0.0121\n",
            "Epoch 281/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 3.9668e-05 - mae: 0.0062 - val_loss: 1.5451e-04 - val_mae: 0.0121\n",
            "Epoch 282/500\n",
            "43/43 [==============================] - ETA: 0s - loss: 3.9567e-05 - mae: 0.0062INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 3.9567e-05 - mae: 0.0062 - val_loss: 1.5135e-04 - val_mae: 0.0121\n",
            "Epoch 283/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 3.9671e-05 - mae: 0.0062 - val_loss: 1.8480e-04 - val_mae: 0.0137\n",
            "Epoch 284/500\n",
            "36/43 [========================>.....] - ETA: 0s - loss: 3.7429e-05 - mae: 0.0059INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 3.9648e-05 - mae: 0.0062 - val_loss: 1.5035e-04 - val_mae: 0.0121\n",
            "Epoch 285/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 3.8714e-05 - mae: 0.0061INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 3.9109e-05 - mae: 0.0062 - val_loss: 1.5016e-04 - val_mae: 0.0121\n",
            "Epoch 286/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 3.8911e-05 - mae: 0.0061 - val_loss: 1.5180e-04 - val_mae: 0.0123\n",
            "Epoch 287/500\n",
            "43/43 [==============================] - 1s 18ms/step - loss: 3.9051e-05 - mae: 0.0061 - val_loss: 1.6461e-04 - val_mae: 0.0134\n",
            "Epoch 288/500\n",
            "43/43 [==============================] - 1s 19ms/step - loss: 3.9788e-05 - mae: 0.0062 - val_loss: 1.6436e-04 - val_mae: 0.0127\n",
            "Epoch 289/500\n",
            "43/43 [==============================] - 1s 19ms/step - loss: 3.9013e-05 - mae: 0.0061 - val_loss: 1.5725e-04 - val_mae: 0.0123\n",
            "Epoch 290/500\n",
            "43/43 [==============================] - 1s 20ms/step - loss: 4.0616e-05 - mae: 0.0063 - val_loss: 1.5058e-04 - val_mae: 0.0120\n",
            "Epoch 291/500\n",
            "43/43 [==============================] - 1s 19ms/step - loss: 3.8796e-05 - mae: 0.0061 - val_loss: 1.6884e-04 - val_mae: 0.0138\n",
            "Epoch 292/500\n",
            "43/43 [==============================] - 1s 18ms/step - loss: 3.8471e-05 - mae: 0.0061 - val_loss: 1.6747e-04 - val_mae: 0.0129\n",
            "Epoch 293/500\n",
            "43/43 [==============================] - 1s 19ms/step - loss: 3.8167e-05 - mae: 0.0061 - val_loss: 1.8956e-04 - val_mae: 0.0141\n",
            "Epoch 294/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 3.6791e-05 - mae: 0.0059INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 25ms/step - loss: 3.8548e-05 - mae: 0.0061 - val_loss: 1.5008e-04 - val_mae: 0.0124\n",
            "Epoch 295/500\n",
            "43/43 [==============================] - 1s 18ms/step - loss: 3.8319e-05 - mae: 0.0061 - val_loss: 1.5023e-04 - val_mae: 0.0124\n",
            "Epoch 296/500\n",
            "43/43 [==============================] - 1s 19ms/step - loss: 3.8271e-05 - mae: 0.0061 - val_loss: 1.6934e-04 - val_mae: 0.0130\n",
            "Epoch 297/500\n",
            "43/43 [==============================] - 1s 19ms/step - loss: 3.8273e-05 - mae: 0.0061 - val_loss: 1.5354e-04 - val_mae: 0.0127\n",
            "Epoch 298/500\n",
            "40/43 [==========================>...] - ETA: 0s - loss: 3.6973e-05 - mae: 0.0060INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 24ms/step - loss: 3.7909e-05 - mae: 0.0061 - val_loss: 1.4640e-04 - val_mae: 0.0121\n",
            "Epoch 299/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 3.8310e-05 - mae: 0.0061 - val_loss: 1.7235e-04 - val_mae: 0.0132\n",
            "Epoch 300/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 3.8895e-05 - mae: 0.0062 - val_loss: 1.9626e-04 - val_mae: 0.0145\n",
            "Epoch 301/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 3.9352e-05 - mae: 0.0062 - val_loss: 1.7727e-04 - val_mae: 0.0145\n",
            "Epoch 302/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 3.7474e-05 - mae: 0.0061 - val_loss: 1.7475e-04 - val_mae: 0.0133\n",
            "Epoch 303/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 3.8758e-05 - mae: 0.0062 - val_loss: 1.4856e-04 - val_mae: 0.0119\n",
            "Epoch 304/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 3.6873e-05 - mae: 0.0060 - val_loss: 1.4798e-04 - val_mae: 0.0124\n",
            "Epoch 305/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 3.5094e-05 - mae: 0.0058INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 22ms/step - loss: 3.7264e-05 - mae: 0.0060 - val_loss: 1.4103e-04 - val_mae: 0.0116\n",
            "Epoch 306/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 3.7065e-05 - mae: 0.0060 - val_loss: 1.4987e-04 - val_mae: 0.0126\n",
            "Epoch 307/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 3.8486e-05 - mae: 0.0061 - val_loss: 1.4259e-04 - val_mae: 0.0116\n",
            "Epoch 308/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 3.7701e-05 - mae: 0.0060 - val_loss: 1.4293e-04 - val_mae: 0.0116\n",
            "Epoch 309/500\n",
            "37/43 [========================>.....] - ETA: 0s - loss: 3.4249e-05 - mae: 0.0057INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 3.6623e-05 - mae: 0.0059 - val_loss: 1.3906e-04 - val_mae: 0.0115\n",
            "Epoch 310/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 3.6465e-05 - mae: 0.0060 - val_loss: 1.4504e-04 - val_mae: 0.0118\n",
            "Epoch 311/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 3.7704e-05 - mae: 0.0060 - val_loss: 1.6691e-04 - val_mae: 0.0130\n",
            "Epoch 312/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 3.6789e-05 - mae: 0.0060 - val_loss: 1.5623e-04 - val_mae: 0.0124\n",
            "Epoch 313/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 3.6781e-05 - mae: 0.0060 - val_loss: 1.6586e-04 - val_mae: 0.0129\n",
            "Epoch 314/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 3.8738e-05 - mae: 0.0061 - val_loss: 1.6767e-04 - val_mae: 0.0140\n",
            "Epoch 315/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 3.6379e-05 - mae: 0.0060 - val_loss: 1.4065e-04 - val_mae: 0.0119\n",
            "Epoch 316/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 3.7097e-05 - mae: 0.0060 - val_loss: 1.4240e-04 - val_mae: 0.0121\n",
            "Epoch 317/500\n",
            "41/43 [===========================>..] - ETA: 0s - loss: 3.5428e-05 - mae: 0.0059INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 3.5844e-05 - mae: 0.0059 - val_loss: 1.3606e-04 - val_mae: 0.0114\n",
            "Epoch 318/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 3.5866e-05 - mae: 0.0059 - val_loss: 1.4084e-04 - val_mae: 0.0116\n",
            "Epoch 319/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 3.7152e-05 - mae: 0.0060 - val_loss: 1.3905e-04 - val_mae: 0.0119\n",
            "Epoch 320/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 3.6133e-05 - mae: 0.0059 - val_loss: 1.8015e-04 - val_mae: 0.0138\n",
            "Epoch 321/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 3.6192e-05 - mae: 0.0059 - val_loss: 1.3919e-04 - val_mae: 0.0119\n",
            "Epoch 322/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 3.8338e-05 - mae: 0.0061 - val_loss: 1.8074e-04 - val_mae: 0.0138\n",
            "Epoch 323/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 3.5831e-05 - mae: 0.0059 - val_loss: 1.4138e-04 - val_mae: 0.0121\n",
            "Epoch 324/500\n",
            "38/43 [=========================>....] - ETA: 0s - loss: 3.3378e-05 - mae: 0.0056INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 3.5383e-05 - mae: 0.0059 - val_loss: 1.3379e-04 - val_mae: 0.0113\n",
            "Epoch 325/500\n",
            "43/43 [==============================] - 1s 15ms/step - loss: 3.5147e-05 - mae: 0.0058 - val_loss: 1.3517e-04 - val_mae: 0.0113\n",
            "Epoch 326/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 3.4972e-05 - mae: 0.0058 - val_loss: 1.3448e-04 - val_mae: 0.0115\n",
            "Epoch 327/500\n",
            "41/43 [===========================>..] - ETA: 0s - loss: 3.5128e-05 - mae: 0.0058INFO:tensorflow:Assets written to: my_checkpoint/assets\n",
            "43/43 [==============================] - 1s 23ms/step - loss: 3.5336e-05 - mae: 0.0058 - val_loss: 1.3239e-04 - val_mae: 0.0113\n",
            "Epoch 328/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 3.5138e-05 - mae: 0.0058 - val_loss: 1.3517e-04 - val_mae: 0.0113\n",
            "Epoch 329/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 3.5030e-05 - mae: 0.0058 - val_loss: 1.3508e-04 - val_mae: 0.0117\n",
            "Epoch 330/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 3.6038e-05 - mae: 0.0059 - val_loss: 1.4901e-04 - val_mae: 0.0121\n",
            "Epoch 331/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 3.5270e-05 - mae: 0.0058 - val_loss: 1.3692e-04 - val_mae: 0.0114\n",
            "Epoch 332/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 3.5237e-05 - mae: 0.0058 - val_loss: 1.3874e-04 - val_mae: 0.0121\n",
            "Epoch 333/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 3.4797e-05 - mae: 0.0058 - val_loss: 1.3375e-04 - val_mae: 0.0116\n",
            "Epoch 334/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 3.5912e-05 - mae: 0.0059 - val_loss: 1.7908e-04 - val_mae: 0.0138\n",
            "Epoch 335/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 3.5534e-05 - mae: 0.0059 - val_loss: 1.5882e-04 - val_mae: 0.0127\n",
            "Epoch 336/500\n",
            "43/43 [==============================] - 1s 16ms/step - loss: 3.4622e-05 - mae: 0.0058 - val_loss: 1.3577e-04 - val_mae: 0.0114\n",
            "Epoch 337/500\n",
            "43/43 [==============================] - 1s 17ms/step - loss: 3.5224e-05 - mae: 0.0058 - val_loss: 1.4928e-04 - val_mae: 0.0121\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tensorflow.python.keras.callbacks.History at 0x7f5281270978>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "FnIWROQ08OIj",
        "colab": {}
      },
      "source": [
        "lin_forecast = model_forecast(model, spy_normalized_to_traindata.flatten()[x_test.index.min() - window_size:-1], window_size)[:, 0]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "xd7Tj_fA8OIk",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 53
        },
        "outputId": "fbae2db7-be54-4ba9-fabd-1a05e77d6000"
      },
      "source": [
        "# Undo the scaling\n",
        "lin_forecast = x_train_scaler.inverse_transform(lin_forecast.reshape(-1,1)).flatten()\n",
        "lin_forecast.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(422,)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "F-nftslfgQJs",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 627
        },
        "outputId": "4454b69d-bd4b-479a-9279-bed6345e0d9c"
      },
      "source": [
        "# Plot results\n",
        "plt.title('Linear Forecast')\n",
        "plt.ylabel('Dollars $')\n",
        "plt.xlabel('Timestep in Days')\n",
        "plot_series(x_test.index, x_test)\n",
        "plot_series(x_test.index, lin_forecast)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x720 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gKyiDSZ9hEXK",
        "colab_type": "text"
      },
      "source": [
        "#### Linear Model Result"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "W4E_jXktf7iv",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c7fa362a-9592-4701-d12f-46d46db60710"
      },
      "source": [
        "keras.metrics.mean_absolute_error(x_test, lin_forecast).numpy()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "3.8283994"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "S9fM69__PvMi",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}