Applications/Sunspot_Prediction_with_CNN_and_LSTM.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 641
    },
    "colab_type": "code",
    "id": "GNkzTFfynsmV",
    "outputId": "f636dee8-8cde-4d19-ff5f-d712ef9a51c1"
   },
   "source": [
    "# Sunspot Prediction with CNN and LSTM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-09-24T16:56:27.185453Z",
     "start_time": "2019-09-24T16:56:26.100406Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "56XEQOGknrAk",
    "outputId": "d52053ae-8ab5-4863-97c9-50073f912f46"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.0.0-rc1\n"
     ]
    }
   ],
   "source": [
    "import tensorflow as tf\n",
    "print(tf.__version__)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-09-24T16:56:27.280140Z",
     "start_time": "2019-09-24T16:56:27.187283Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1 Physical GPUs, 1 Logical GPUs\n"
     ]
    }
   ],
   "source": [
    "gpus = tf.config.experimental.list_physical_devices('GPU')\n",
    "if gpus:\n",
    "    try:\n",
    "        # Currently, memory growth needs to be the same across GPUs\n",
    "        for gpu in gpus:\n",
    "            tf.config.experimental.set_memory_growth(gpu, True)\n",
    "        logical_gpus = tf.config.experimental.list_logical_devices('GPU')\n",
    "        print(len(gpus), \"Physical GPUs,\", len(logical_gpus), \"Logical GPUs\")\n",
    "    except RuntimeError as e:\n",
    "        # Memory growth must be set before GPUs have been initialized\n",
    "        print(e)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Getting the dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-09-24T16:56:27.824785Z",
     "start_time": "2019-09-24T16:56:27.281511Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 208
    },
    "colab_type": "code",
    "id": "tP7oqUdkk0gY",
    "outputId": "549237ab-672c-49d5-b36e-bd196eae772f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--2019-09-24 19:56:27--  https://storage.googleapis.com/laurencemoroney-blog.appspot.com/Sunspots.csv\n",
      "Resolving storage.googleapis.com (storage.googleapis.com)... 172.217.169.176, 2a00:1450:4017:807::2010\n",
      "Connecting to storage.googleapis.com (storage.googleapis.com)|172.217.169.176|:443... connected.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 70827 (69K) [application/octet-stream]\n",
      "Saving to: ‘/tmp/sunspots.csv’\n",
      "\n",
      "/tmp/sunspots.csv   100%[===================>]  69.17K  --.-KB/s    in 0.04s   \n",
      "\n",
      "2019-09-24 19:56:27 (1.68 MB/s) - ‘/tmp/sunspots.csv’ saved [70827/70827]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "!wget --no-check-certificate \\\n",
    "    https://storage.googleapis.com/laurencemoroney-blog.appspot.com/Sunspots.csv \\\n",
    "    -O /tmp/sunspots.csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-09-24T16:56:28.030053Z",
     "start_time": "2019-09-24T16:56:27.828670Z"
    },
    "colab": {},
    "colab_type": "code",
    "id": "sLl52leVp5wU"
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "def plot_series(time, series, format=\"-\", start=0, end=None):\n",
    "    plt.plot(time[start:end], series[start:end], format)\n",
    "    plt.xlabel(\"Time\")\n",
    "    plt.ylabel(\"Value\")\n",
    "    plt.grid(True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-09-24T16:56:28.277300Z",
     "start_time": "2019-09-24T16:56:28.031496Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 392
    },
    "colab_type": "code",
    "id": "NcG9r1eClbTh",
    "outputId": "f3545940-e807-4109-f8be-a95a3a16d4cf"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import csv\n",
    "time_step = []\n",
    "sunspots = []\n",
    "\n",
    "with open('/tmp/sunspots.csv') as csvfile:\n",
    "    reader = csv.reader(csvfile, delimiter=',')\n",
    "    next(reader)\n",
    "    for row in reader:\n",
    "        sunspots.append(float(row[2]))\n",
    "        time_step.append(int(row[0]))\n",
    "\n",
    "series = np.array(sunspots)\n",
    "time = np.array(time_step)\n",
    "plt.figure(figsize=(10, 6))\n",
    "plot_series(time, series)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-09-24T16:56:28.525671Z",
     "start_time": "2019-09-24T16:56:28.278707Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 392
    },
    "colab_type": "code",
    "id": "VinZVwUa8WO7",
    "outputId": "e95494c5-13c5-4b60-85f0-f9021720cff3",
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "series = np.array(sunspots)\n",
    "time = np.array(time_step)\n",
    "plt.figure(figsize=(10, 6))\n",
    "plot_series(time, series)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-09-24T16:56:28.531941Z",
     "start_time": "2019-09-24T16:56:28.526993Z"
    },
    "colab": {},
    "colab_type": "code",
    "id": "L92YRw_IpCFG"
   },
   "outputs": [],
   "source": [
    "split_time = 3000\n",
    "time_train = time[:split_time]\n",
    "x_train = series[:split_time]\n",
    "time_valid = time[split_time:]\n",
    "x_valid = series[split_time:]\n",
    "\n",
    "window_size = 30\n",
    "batch_size = 32\n",
    "shuffle_buffer_size = 1000"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-09-24T16:56:28.608649Z",
     "start_time": "2019-09-24T16:56:28.534973Z"
    },
    "colab": {},
    "colab_type": "code",
    "id": "lJwUUZscnG38"
   },
   "outputs": [],
   "source": [
    "def windowed_dataset(series, window_size, batch_size, shuffle_buffer):\n",
    "    series = tf.expand_dims(series, axis=-1)\n",
    "    ds = tf.data.Dataset.from_tensor_slices(series)\n",
    "    ds = ds.window(window_size + 1, shift=1, drop_remainder=True)\n",
    "    ds = ds.flat_map(lambda w: w.batch(window_size + 1))\n",
    "    ds = ds.shuffle(shuffle_buffer)\n",
    "    ds = ds.map(lambda w: (w[:-1], w[1:]))\n",
    "    return ds.batch(batch_size).prefetch(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-09-24T16:56:28.719762Z",
     "start_time": "2019-09-24T16:56:28.610544Z"
    },
    "colab": {},
    "colab_type": "code",
    "id": "4XwGrf-A_wF0"
   },
   "outputs": [],
   "source": [
    "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"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Training with dynamic learning rate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-09-24T16:57:27.648596Z",
     "start_time": "2019-09-24T16:56:28.721531Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "colab_type": "code",
    "id": "AclfYY3Mn6Ph",
    "outputId": "02a3801f-9033-459f-c1c6-f930e65fc01f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:Entity <function windowed_dataset.<locals>.<lambda> at 0x7f709704a7a0> could not be transformed and will be executed as-is. Please report this to the AutoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: module 'gast' has no attribute 'Str'\n",
      "WARNING: Entity <function windowed_dataset.<locals>.<lambda> at 0x7f709704a7a0> could not be transformed and will be executed as-is. Please report this to the AutoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: module 'gast' has no attribute 'Str'\n",
      "WARNING:tensorflow:From /home/einsteinekeine/anaconda3/envs/tensorflow2/lib/python3.7/site-packages/tensorflow_core/python/data/util/random_seed.py:58: where (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.where in 2.0, which has the same broadcast rule as np.where\n",
      "WARNING:tensorflow:Entity <function windowed_dataset.<locals>.<lambda> at 0x7f709704a9e0> could not be transformed and will be executed as-is. Please report this to the AutoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: module 'gast' has no attribute 'Str'\n",
      "WARNING: Entity <function windowed_dataset.<locals>.<lambda> at 0x7f709704a9e0> could not be transformed and will be executed as-is. Please report this to the AutoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: module 'gast' has no attribute 'Str'\n",
      "<PrefetchDataset shapes: ((None, None, 1), (None, None, 1)), types: (tf.float64, tf.float64)>\n",
      "(3000,)\n",
      "Epoch 1/100\n",
      "WARNING:tensorflow:Entity <function Function._initialize_uninitialized_variables.<locals>.initialize_variables at 0x7f70970c90e0> could not be transformed and will be executed as-is. Please report this to the AutoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: module 'gast' has no attribute 'Num'\n",
      "WARNING: Entity <function Function._initialize_uninitialized_variables.<locals>.initialize_variables at 0x7f70970c90e0> could not be transformed and will be executed as-is. Please report this to the AutoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: module 'gast' has no attribute 'Num'\n",
      "12/12 [==============================] - 6s 500ms/step - loss: 80.3072 - mae: 80.3315\n",
      "Epoch 2/100\n",
      "12/12 [==============================] - 1s 47ms/step - loss: 77.9405 - mae: 78.5896\n",
      "Epoch 3/100\n",
      "12/12 [==============================] - 1s 48ms/step - loss: 75.2415 - mae: 75.9445\n",
      "Epoch 4/100\n",
      "12/12 [==============================] - 1s 47ms/step - loss: 72.2421 - mae: 72.7693\n",
      "Epoch 5/100\n",
      "12/12 [==============================] - 1s 49ms/step - loss: 68.5303 - mae: 69.2661\n",
      "Epoch 6/100\n",
      "12/12 [==============================] - 1s 47ms/step - loss: 64.8655 - mae: 65.6140\n",
      "Epoch 7/100\n",
      "12/12 [==============================] - 0s 41ms/step - loss: 61.1551 - mae: 62.0189\n",
      "Epoch 8/100\n",
      "12/12 [==============================] - 0s 38ms/step - loss: 58.2311 - mae: 58.6465\n",
      "Epoch 9/100\n",
      "12/12 [==============================] - 0s 38ms/step - loss: 54.7059 - mae: 55.5682\n",
      "Epoch 10/100\n",
      "12/12 [==============================] - 0s 38ms/step - loss: 52.1101 - mae: 52.8302\n",
      "Epoch 11/100\n",
      "12/12 [==============================] - 0s 39ms/step - loss: 49.8018 - mae: 50.4361\n",
      "Epoch 12/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 47.4305 - mae: 48.3348\n",
      "Epoch 13/100\n",
      "12/12 [==============================] - 0s 38ms/step - loss: 45.7015 - mae: 46.5603\n",
      "Epoch 14/100\n",
      "12/12 [==============================] - 0s 39ms/step - loss: 44.4097 - mae: 45.0471\n",
      "Epoch 15/100\n",
      "12/12 [==============================] - 0s 39ms/step - loss: 43.0279 - mae: 43.8073\n",
      "Epoch 16/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 42.0098 - mae: 42.7752\n",
      "Epoch 17/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 41.3051 - mae: 41.8717\n",
      "Epoch 18/100\n",
      "12/12 [==============================] - 0s 41ms/step - loss: 40.4317 - mae: 41.0402\n",
      "Epoch 19/100\n",
      "12/12 [==============================] - 0s 41ms/step - loss: 39.6792 - mae: 40.2585\n",
      "Epoch 20/100\n",
      "12/12 [==============================] - 0s 41ms/step - loss: 38.8637 - mae: 39.5122\n",
      "Epoch 21/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 38.1928 - mae: 38.7666\n",
      "Epoch 22/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 37.3514 - mae: 37.9312\n",
      "Epoch 23/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 36.2445 - mae: 36.8968\n",
      "Epoch 24/100\n",
      "12/12 [==============================] - 1s 47ms/step - loss: 35.0458 - mae: 35.7751\n",
      "Epoch 25/100\n",
      "12/12 [==============================] - 1s 51ms/step - loss: 34.0178 - mae: 34.6578\n",
      "Epoch 26/100\n",
      "12/12 [==============================] - 1s 53ms/step - loss: 32.8340 - mae: 33.4873\n",
      "Epoch 27/100\n",
      "12/12 [==============================] - 1s 50ms/step - loss: 31.7496 - mae: 32.3715\n",
      "Epoch 28/100\n",
      "12/12 [==============================] - 1s 52ms/step - loss: 30.8722 - mae: 31.4446\n",
      "Epoch 29/100\n",
      "12/12 [==============================] - 1s 50ms/step - loss: 30.2910 - mae: 30.7658\n",
      "Epoch 30/100\n",
      "12/12 [==============================] - 1s 49ms/step - loss: 29.4990 - mae: 30.0809\n",
      "Epoch 31/100\n",
      "12/12 [==============================] - 0s 41ms/step - loss: 28.9362 - mae: 29.5020\n",
      "Epoch 32/100\n",
      "12/12 [==============================] - 0s 39ms/step - loss: 28.5240 - mae: 29.1636\n",
      "Epoch 33/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 28.3124 - mae: 28.8301\n",
      "Epoch 34/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 27.8299 - mae: 28.3203\n",
      "Epoch 35/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 27.6437 - mae: 28.1922\n",
      "Epoch 36/100\n",
      "12/12 [==============================] - 0s 41ms/step - loss: 27.3185 - mae: 27.8401\n",
      "Epoch 37/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 26.9777 - mae: 27.4991\n",
      "Epoch 38/100\n",
      "12/12 [==============================] - 0s 42ms/step - loss: 26.6700 - mae: 27.2407\n",
      "Epoch 39/100\n",
      "12/12 [==============================] - 1s 48ms/step - loss: 26.1427 - mae: 26.6721\n",
      "Epoch 40/100\n",
      "12/12 [==============================] - 1s 47ms/step - loss: 26.0148 - mae: 26.5405\n",
      "Epoch 41/100\n",
      "12/12 [==============================] - 1s 47ms/step - loss: 25.8366 - mae: 26.4308\n",
      "Epoch 42/100\n",
      "12/12 [==============================] - 1s 47ms/step - loss: 25.0225 - mae: 25.6152\n",
      "Epoch 43/100\n",
      "12/12 [==============================] - 1s 46ms/step - loss: 24.6221 - mae: 25.2052\n",
      "Epoch 44/100\n",
      "12/12 [==============================] - 1s 46ms/step - loss: 24.8549 - mae: 25.3525\n",
      "Epoch 45/100\n",
      "12/12 [==============================] - 1s 47ms/step - loss: 24.2028 - mae: 24.7768\n",
      "Epoch 46/100\n",
      "12/12 [==============================] - 0s 38ms/step - loss: 23.6379 - mae: 24.1720\n",
      "Epoch 47/100\n",
      "12/12 [==============================] - 0s 39ms/step - loss: 23.3610 - mae: 23.9052\n",
      "Epoch 48/100\n",
      "12/12 [==============================] - 0s 38ms/step - loss: 22.8374 - mae: 23.3667\n",
      "Epoch 49/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 22.6191 - mae: 23.1726\n",
      "Epoch 50/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 22.3840 - mae: 22.9322\n",
      "Epoch 51/100\n",
      "12/12 [==============================] - 0s 39ms/step - loss: 21.8587 - mae: 22.3651\n",
      "Epoch 52/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 22.1095 - mae: 22.6738\n",
      "Epoch 53/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 22.6315 - mae: 23.2308\n",
      "Epoch 54/100\n",
      "12/12 [==============================] - 0s 38ms/step - loss: 25.1602 - mae: 25.5495\n",
      "Epoch 55/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 22.7928 - mae: 23.3167\n",
      "Epoch 56/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 21.7767 - mae: 22.2807\n",
      "Epoch 57/100\n",
      "12/12 [==============================] - 0s 39ms/step - loss: 21.2909 - mae: 21.7618\n",
      "Epoch 58/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "12/12 [==============================] - 0s 41ms/step - loss: 20.6146 - mae: 21.1392\n",
      "Epoch 59/100\n",
      "12/12 [==============================] - 0s 41ms/step - loss: 20.2411 - mae: 20.7571\n",
      "Epoch 60/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 19.8599 - mae: 20.3675\n",
      "Epoch 61/100\n",
      "12/12 [==============================] - 0s 39ms/step - loss: 19.5709 - mae: 20.0732\n",
      "Epoch 62/100\n",
      "12/12 [==============================] - 0s 41ms/step - loss: 19.2822 - mae: 19.7898\n",
      "Epoch 63/100\n",
      "12/12 [==============================] - 1s 51ms/step - loss: 20.3256 - mae: 20.7618\n",
      "Epoch 64/100\n",
      "12/12 [==============================] - 1s 51ms/step - loss: 22.9920 - mae: 23.8479\n",
      "Epoch 65/100\n",
      "12/12 [==============================] - 1s 49ms/step - loss: 21.4272 - mae: 21.8940\n",
      "Epoch 66/100\n",
      "12/12 [==============================] - 1s 51ms/step - loss: 20.8018 - mae: 21.3008\n",
      "Epoch 67/100\n",
      "12/12 [==============================] - 1s 51ms/step - loss: 19.2876 - mae: 19.7751\n",
      "Epoch 68/100\n",
      "12/12 [==============================] - 1s 49ms/step - loss: 22.4707 - mae: 22.8980\n",
      "Epoch 69/100\n",
      "12/12 [==============================] - 1s 45ms/step - loss: 20.9526 - mae: 21.6596\n",
      "Epoch 70/100\n",
      "12/12 [==============================] - 0s 39ms/step - loss: 21.4790 - mae: 21.9572\n",
      "Epoch 71/100\n",
      "12/12 [==============================] - 0s 39ms/step - loss: 20.9505 - mae: 21.4694\n",
      "Epoch 72/100\n",
      "12/12 [==============================] - 0s 39ms/step - loss: 20.4367 - mae: 20.8523\n",
      "Epoch 73/100\n",
      "12/12 [==============================] - 0s 41ms/step - loss: 22.3824 - mae: 22.8633\n",
      "Epoch 74/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 20.6406 - mae: 21.1487\n",
      "Epoch 75/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 21.2641 - mae: 21.7262\n",
      "Epoch 76/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 22.8114 - mae: 24.3127\n",
      "Epoch 77/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 23.1512 - mae: 24.3892\n",
      "Epoch 78/100\n",
      "12/12 [==============================] - 1s 46ms/step - loss: 35.6037 - mae: 35.8474\n",
      "Epoch 79/100\n",
      "12/12 [==============================] - 1s 46ms/step - loss: 36.2451 - mae: 36.4884\n",
      "Epoch 80/100\n",
      "12/12 [==============================] - 1s 49ms/step - loss: 34.4041 - mae: 34.8155\n",
      "Epoch 81/100\n",
      "12/12 [==============================] - 1s 46ms/step - loss: 36.6789 - mae: 37.1994\n",
      "Epoch 82/100\n",
      "12/12 [==============================] - 1s 49ms/step - loss: 33.6312 - mae: 34.0763\n",
      "Epoch 83/100\n",
      "12/12 [==============================] - 1s 49ms/step - loss: 30.7690 - mae: 31.0514\n",
      "Epoch 84/100\n",
      "12/12 [==============================] - 1s 49ms/step - loss: 26.7410 - mae: 27.4653\n",
      "Epoch 85/100\n",
      "12/12 [==============================] - 0s 38ms/step - loss: 43.1456 - mae: 43.7976\n",
      "Epoch 86/100\n",
      "12/12 [==============================] - 0s 38ms/step - loss: 52.8235 - mae: 53.4180\n",
      "Epoch 87/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 44.5953 - mae: 46.3208\n",
      "Epoch 88/100\n",
      "12/12 [==============================] - 0s 39ms/step - loss: 54.2563 - mae: 54.6236\n",
      "Epoch 89/100\n",
      "12/12 [==============================] - 0s 41ms/step - loss: 50.3584 - mae: 51.3489\n",
      "Epoch 90/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 57.7713 - mae: 58.1819\n",
      "Epoch 91/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 46.7166 - mae: 47.1262\n",
      "Epoch 92/100\n",
      "12/12 [==============================] - 0s 39ms/step - loss: 63.0956 - mae: 63.5087\n",
      "Epoch 93/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 55.8484 - mae: 56.4094\n",
      "Epoch 94/100\n",
      "12/12 [==============================] - 0s 41ms/step - loss: 54.4635 - mae: 54.9577\n",
      "Epoch 95/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 54.2534 - mae: 55.1202\n",
      "Epoch 96/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 54.4320 - mae: 54.9542\n",
      "Epoch 97/100\n",
      "12/12 [==============================] - 0s 39ms/step - loss: 54.1298 - mae: 54.7822\n",
      "Epoch 98/100\n",
      "12/12 [==============================] - 0s 41ms/step - loss: 56.0556 - mae: 56.8690\n",
      "Epoch 99/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 54.2974 - mae: 54.9411\n",
      "Epoch 100/100\n",
      "12/12 [==============================] - 0s 40ms/step - loss: 55.1858 - mae: 55.7529\n"
     ]
    }
   ],
   "source": [
    "tf.keras.backend.clear_session()\n",
    "tf.random.set_seed(51)\n",
    "np.random.seed(51)\n",
    "window_size = 64\n",
    "batch_size = 256\n",
    "train_set = windowed_dataset(x_train, window_size, batch_size, shuffle_buffer_size)\n",
    "print(train_set)\n",
    "print(x_train.shape)\n",
    "\n",
    "model = tf.keras.models.Sequential([\n",
    "  tf.keras.layers.Conv1D(filters=32, kernel_size=5,\n",
    "                      strides=1, padding=\"causal\",\n",
    "                      activation=\"relu\",\n",
    "                      input_shape=[None, 1]),\n",
    "  tf.keras.layers.LSTM(64, return_sequences=True),\n",
    "  tf.keras.layers.LSTM(64, return_sequences=True),\n",
    "  tf.keras.layers.Dense(30, activation=\"relu\"),\n",
    "  tf.keras.layers.Dense(10, activation=\"relu\"),\n",
    "  tf.keras.layers.Dense(1),\n",
    "  tf.keras.layers.Lambda(lambda x: x * 400)\n",
    "])\n",
    "\n",
    "lr_schedule = tf.keras.callbacks.LearningRateScheduler(\n",
    "    lambda epoch: 1e-8 * 10**(epoch / 20))\n",
    "optimizer = tf.keras.optimizers.SGD(lr=1e-8, momentum=0.9)\n",
    "model.compile(loss=tf.keras.losses.Huber(),\n",
    "              optimizer=optimizer,\n",
    "              metrics=[\"mae\"])\n",
    "\n",
    "with tf.device('/device:GPU:0'):\n",
    "    history = model.fit(train_set, epochs=100, callbacks=[lr_schedule])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-09-24T16:57:28.329477Z",
     "start_time": "2019-09-24T16:57:27.650368Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 290
    },
    "colab_type": "code",
    "id": "vVcKmg7Q_7rD",
    "outputId": "14773dd5-ab46-469a-dc02-96f4b7607b07"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1e-08, 0.0001, 0, 60]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.semilogx(history.history[\"lr\"], history.history[\"loss\"])\n",
    "plt.axis([1e-8, 1e-4, 0, 60])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Training with fixed learning rate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-09-24T17:02:42.715303Z",
     "start_time": "2019-09-24T16:57:28.331536Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "colab_type": "code",
    "id": "QsksvkcXAAgq",
    "outputId": "be930f4c-30b1-4e16-92c0-4778fbca4c27"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:Entity <function windowed_dataset.<locals>.<lambda> at 0x7f6f60211050> could not be transformed and will be executed as-is. Please report this to the AutoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: module 'gast' has no attribute 'Str'\n",
      "WARNING: Entity <function windowed_dataset.<locals>.<lambda> at 0x7f6f60211050> could not be transformed and will be executed as-is. Please report this to the AutoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: module 'gast' has no attribute 'Str'\n",
      "WARNING:tensorflow:Entity <function windowed_dataset.<locals>.<lambda> at 0x7f6f602114d0> could not be transformed and will be executed as-is. Please report this to the AutoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: module 'gast' has no attribute 'Str'\n",
      "WARNING: Entity <function windowed_dataset.<locals>.<lambda> at 0x7f6f602114d0> could not be transformed and will be executed as-is. Please report this to the AutoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: module 'gast' has no attribute 'Str'\n",
      "Epoch 1/500\n",
      "WARNING:tensorflow:Entity <function Function._initialize_uninitialized_variables.<locals>.initialize_variables at 0x7f6f609fecb0> could not be transformed and will be executed as-is. Please report this to the AutoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: module 'gast' has no attribute 'Num'\n",
      "WARNING: Entity <function Function._initialize_uninitialized_variables.<locals>.initialize_variables at 0x7f6f609fecb0> could not be transformed and will be executed as-is. Please report this to the AutoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: module 'gast' has no attribute 'Num'\n",
      "30/30 [==============================] - 4s 145ms/step - loss: 37.8895 - mae: 38.6059\n",
      "Epoch 2/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 24.1465 - mae: 24.6112\n",
      "Epoch 3/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 22.0205 - mae: 22.5112\n",
      "Epoch 4/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 19.8473 - mae: 20.3490\n",
      "Epoch 5/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 19.5129 - mae: 20.0170\n",
      "Epoch 6/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 19.1338 - mae: 19.6244\n",
      "Epoch 7/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 18.7991 - mae: 19.2972\n",
      "Epoch 8/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 18.5390 - mae: 19.0127\n",
      "Epoch 9/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 18.3587 - mae: 18.8708\n",
      "Epoch 10/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 18.4545 - mae: 18.9693\n",
      "Epoch 11/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 17.8308 - mae: 18.3274\n",
      "Epoch 12/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 17.8224 - mae: 18.3112\n",
      "Epoch 13/500\n",
      "30/30 [==============================] - 1s 18ms/step - loss: 18.0649 - mae: 18.5609\n",
      "Epoch 14/500\n",
      "30/30 [==============================] - 1s 18ms/step - loss: 17.7592 - mae: 18.2537\n",
      "Epoch 15/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 17.6222 - mae: 18.1213\n",
      "Epoch 16/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 17.8158 - mae: 18.3041\n",
      "Epoch 17/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 17.7636 - mae: 18.2681\n",
      "Epoch 18/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 17.7889 - mae: 18.2854\n",
      "Epoch 19/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 17.6817 - mae: 18.1676\n",
      "Epoch 20/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 17.3496 - mae: 17.8040\n",
      "Epoch 21/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 17.3972 - mae: 17.8939\n",
      "Epoch 22/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 17.4339 - mae: 17.9057\n",
      "Epoch 23/500\n",
      "30/30 [==============================] - 1s 18ms/step - loss: 17.2551 - mae: 17.7287\n",
      "Epoch 24/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 17.1937 - mae: 17.7033\n",
      "Epoch 25/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 17.1147 - mae: 17.6133\n",
      "Epoch 26/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 17.1008 - mae: 17.5880\n",
      "Epoch 27/500\n",
      "30/30 [==============================] - 1s 24ms/step - loss: 17.1257 - mae: 17.6343\n",
      "Epoch 28/500\n",
      "30/30 [==============================] - 1s 24ms/step - loss: 17.4102 - mae: 17.9088\n",
      "Epoch 29/500\n",
      "30/30 [==============================] - 1s 24ms/step - loss: 17.3233 - mae: 17.7848\n",
      "Epoch 30/500\n",
      "30/30 [==============================] - 1s 24ms/step - loss: 17.0633 - mae: 17.5579\n",
      "Epoch 31/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 17.0145 - mae: 17.5132\n",
      "Epoch 32/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 17.0823 - mae: 17.5841\n",
      "Epoch 33/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.9408 - mae: 17.4139\n",
      "Epoch 34/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 17.0733 - mae: 17.5511\n",
      "Epoch 35/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.8832 - mae: 17.3495\n",
      "Epoch 36/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 17.0080 - mae: 17.4944\n",
      "Epoch 37/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.9199 - mae: 17.4286\n",
      "Epoch 38/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 17.2416 - mae: 17.7362\n",
      "Epoch 39/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 16.9880 - mae: 17.4786\n",
      "Epoch 40/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 17.1945 - mae: 17.6687\n",
      "Epoch 41/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 17.1514 - mae: 17.6371\n",
      "Epoch 42/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 16.8132 - mae: 17.3058\n",
      "Epoch 43/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 16.8227 - mae: 17.3137\n",
      "Epoch 44/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 16.8580 - mae: 17.3533\n",
      "Epoch 45/500\n",
      "30/30 [==============================] - 1s 18ms/step - loss: 16.8932 - mae: 17.3789\n",
      "Epoch 46/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.8880 - mae: 17.3876\n",
      "Epoch 47/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.6823 - mae: 17.1998\n",
      "Epoch 48/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 16.7945 - mae: 17.3072\n",
      "Epoch 49/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.9685 - mae: 17.4662\n",
      "Epoch 50/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 17.0529 - mae: 17.5435\n",
      "Epoch 51/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.6805 - mae: 17.1755\n",
      "Epoch 52/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.9099 - mae: 17.4054\n",
      "Epoch 53/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.8124 - mae: 17.2966\n",
      "Epoch 54/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.6992 - mae: 17.1961\n",
      "Epoch 55/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.6855 - mae: 17.1760\n",
      "Epoch 56/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.6232 - mae: 17.1081\n",
      "Epoch 57/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.6667 - mae: 17.1540\n",
      "Epoch 58/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.6247 - mae: 17.0966\n",
      "Epoch 59/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 16.5710 - mae: 17.0572\n",
      "Epoch 60/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 16.6398 - mae: 17.1284\n",
      "Epoch 61/500\n",
      "30/30 [==============================] - 1s 24ms/step - loss: 16.6891 - mae: 17.1652\n",
      "Epoch 62/500\n",
      "30/30 [==============================] - 1s 25ms/step - loss: 16.7798 - mae: 17.2644\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 63/500\n",
      "30/30 [==============================] - 1s 25ms/step - loss: 16.5472 - mae: 17.0265\n",
      "Epoch 64/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 16.6066 - mae: 17.0912\n",
      "Epoch 65/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.6806 - mae: 17.1575\n",
      "Epoch 66/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.5957 - mae: 17.0941\n",
      "Epoch 67/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.5785 - mae: 17.0658\n",
      "Epoch 68/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.5345 - mae: 17.0155\n",
      "Epoch 69/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.5238 - mae: 17.0163\n",
      "Epoch 70/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.8414 - mae: 17.2932\n",
      "Epoch 71/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 16.7555 - mae: 17.2621\n",
      "Epoch 72/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 16.4439 - mae: 16.9315\n",
      "Epoch 73/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 16.4488 - mae: 16.9533\n",
      "Epoch 74/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 16.6407 - mae: 17.1208\n",
      "Epoch 75/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 16.6663 - mae: 17.1708\n",
      "Epoch 76/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 16.5764 - mae: 17.0598\n",
      "Epoch 77/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 16.4585 - mae: 16.9630\n",
      "Epoch 78/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.3886 - mae: 16.8939\n",
      "Epoch 79/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.5387 - mae: 17.0229\n",
      "Epoch 80/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.4229 - mae: 16.9118\n",
      "Epoch 81/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.4475 - mae: 16.9347\n",
      "Epoch 82/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.4367 - mae: 16.9123\n",
      "Epoch 83/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.3471 - mae: 16.8529\n",
      "Epoch 84/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 16.3909 - mae: 16.8873\n",
      "Epoch 85/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 16.3639 - mae: 16.8566\n",
      "Epoch 86/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 16.3705 - mae: 16.8680\n",
      "Epoch 87/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 16.4939 - mae: 16.9929\n",
      "Epoch 88/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 16.3682 - mae: 16.8813\n",
      "Epoch 89/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 16.3099 - mae: 16.7970\n",
      "Epoch 90/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.2999 - mae: 16.7856\n",
      "Epoch 91/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.4520 - mae: 16.9443\n",
      "Epoch 92/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.3443 - mae: 16.8449\n",
      "Epoch 93/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.3254 - mae: 16.8229\n",
      "Epoch 94/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.2546 - mae: 16.7462\n",
      "Epoch 95/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.5067 - mae: 17.0138\n",
      "Epoch 96/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.6584 - mae: 17.1381\n",
      "Epoch 97/500\n",
      "30/30 [==============================] - 1s 18ms/step - loss: 16.3716 - mae: 16.8753\n",
      "Epoch 98/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.2372 - mae: 16.7352\n",
      "Epoch 99/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.2931 - mae: 16.7703\n",
      "Epoch 100/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.2433 - mae: 16.7279\n",
      "Epoch 101/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.2163 - mae: 16.7254\n",
      "Epoch 102/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 16.4371 - mae: 16.9468\n",
      "Epoch 103/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.4653 - mae: 16.9816\n",
      "Epoch 104/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 16.3376 - mae: 16.8326\n",
      "Epoch 105/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 16.2293 - mae: 16.7191\n",
      "Epoch 106/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 16.2136 - mae: 16.7002\n",
      "Epoch 107/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 16.2833 - mae: 16.7475\n",
      "Epoch 108/500\n",
      "30/30 [==============================] - 1s 24ms/step - loss: 16.2394 - mae: 16.7317\n",
      "Epoch 109/500\n",
      "30/30 [==============================] - 1s 24ms/step - loss: 16.1574 - mae: 16.6513\n",
      "Epoch 110/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.1916 - mae: 16.6753\n",
      "Epoch 111/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.2748 - mae: 16.7707\n",
      "Epoch 112/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.1217 - mae: 16.6252\n",
      "Epoch 113/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.1437 - mae: 16.6395\n",
      "Epoch 114/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 16.1940 - mae: 16.6929\n",
      "Epoch 115/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.1114 - mae: 16.6131\n",
      "Epoch 116/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 16.1745 - mae: 16.6636\n",
      "Epoch 117/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 16.2849 - mae: 16.7706\n",
      "Epoch 118/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 16.2201 - mae: 16.7038\n",
      "Epoch 119/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 16.1017 - mae: 16.5989\n",
      "Epoch 120/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 16.1414 - mae: 16.6471\n",
      "Epoch 121/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 16.2675 - mae: 16.7423\n",
      "Epoch 122/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 16.1710 - mae: 16.6739\n",
      "Epoch 123/500\n",
      "30/30 [==============================] - 1s 18ms/step - loss: 16.1323 - mae: 16.6090\n",
      "Epoch 124/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.2924 - mae: 16.7824\n",
      "Epoch 125/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.1046 - mae: 16.5925\n",
      "Epoch 126/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.1029 - mae: 16.5937\n",
      "Epoch 127/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.0738 - mae: 16.5624\n",
      "Epoch 128/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.0845 - mae: 16.5695\n",
      "Epoch 129/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.1148 - mae: 16.5715\n",
      "Epoch 130/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.0430 - mae: 16.5342\n",
      "Epoch 131/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.0814 - mae: 16.5473\n",
      "Epoch 132/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 16.0524 - mae: 16.5465\n",
      "Epoch 133/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.0728 - mae: 16.5518\n",
      "Epoch 134/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 16.0060 - mae: 16.5167\n",
      "Epoch 135/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.0418 - mae: 16.5598\n",
      "Epoch 136/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 16.2157 - mae: 16.7142\n",
      "Epoch 137/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 16.1297 - mae: 16.6281\n",
      "Epoch 138/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 16.0508 - mae: 16.5265\n",
      "Epoch 139/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.9455 - mae: 16.4299\n",
      "Epoch 140/500\n",
      "30/30 [==============================] - 1s 24ms/step - loss: 16.0137 - mae: 16.5282\n",
      "Epoch 141/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.9835 - mae: 16.4718\n",
      "Epoch 142/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.9748 - mae: 16.4632\n",
      "Epoch 143/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.9670 - mae: 16.4716\n",
      "Epoch 144/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.0421 - mae: 16.5302\n",
      "Epoch 145/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.0371 - mae: 16.5226\n",
      "Epoch 146/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30/30 [==============================] - 1s 19ms/step - loss: 15.9609 - mae: 16.4611\n",
      "Epoch 147/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.9355 - mae: 16.4373\n",
      "Epoch 148/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.9020 - mae: 16.3938\n",
      "Epoch 149/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 15.9491 - mae: 16.4274\n",
      "Epoch 150/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.9719 - mae: 16.4698\n",
      "Epoch 151/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.8492 - mae: 16.3437\n",
      "Epoch 152/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 16.0506 - mae: 16.5408\n",
      "Epoch 153/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.8931 - mae: 16.3797\n",
      "Epoch 154/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.8901 - mae: 16.3873\n",
      "Epoch 155/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 15.9054 - mae: 16.3911\n",
      "Epoch 156/500\n",
      "30/30 [==============================] - 1s 18ms/step - loss: 15.8725 - mae: 16.3479\n",
      "Epoch 157/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.0934 - mae: 16.5878\n",
      "Epoch 158/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 16.1727 - mae: 16.6786\n",
      "Epoch 159/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.9568 - mae: 16.4381\n",
      "Epoch 160/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.8214 - mae: 16.3142\n",
      "Epoch 161/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.8612 - mae: 16.3497\n",
      "Epoch 162/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.8492 - mae: 16.3650\n",
      "Epoch 163/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.8867 - mae: 16.3759\n",
      "Epoch 164/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 15.9689 - mae: 16.4687\n",
      "Epoch 165/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.8576 - mae: 16.3652\n",
      "Epoch 166/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.9849 - mae: 16.4649\n",
      "Epoch 167/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.8505 - mae: 16.3413\n",
      "Epoch 168/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 15.8088 - mae: 16.3094\n",
      "Epoch 169/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 15.8450 - mae: 16.3411\n",
      "Epoch 170/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.8042 - mae: 16.2852\n",
      "Epoch 171/500\n",
      "30/30 [==============================] - 1s 24ms/step - loss: 15.7619 - mae: 16.2558\n",
      "Epoch 172/500\n",
      "30/30 [==============================] - 1s 24ms/step - loss: 15.7191 - mae: 16.2134\n",
      "Epoch 173/500\n",
      "30/30 [==============================] - 1s 24ms/step - loss: 15.7794 - mae: 16.2623\n",
      "Epoch 174/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.8310 - mae: 16.3212\n",
      "Epoch 175/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.7623 - mae: 16.2530\n",
      "Epoch 176/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.7519 - mae: 16.2336\n",
      "Epoch 177/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.7866 - mae: 16.2761\n",
      "Epoch 178/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.7321 - mae: 16.2263\n",
      "Epoch 179/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.8371 - mae: 16.3172\n",
      "Epoch 180/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.8860 - mae: 16.3687\n",
      "Epoch 181/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.8914 - mae: 16.3814\n",
      "Epoch 182/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.6941 - mae: 16.1911\n",
      "Epoch 183/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.7731 - mae: 16.2667\n",
      "Epoch 184/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.8235 - mae: 16.3020\n",
      "Epoch 185/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.7008 - mae: 16.1920\n",
      "Epoch 186/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.6481 - mae: 16.1550\n",
      "Epoch 187/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.7290 - mae: 16.2116\n",
      "Epoch 188/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.7349 - mae: 16.2340\n",
      "Epoch 189/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.7326 - mae: 16.2358\n",
      "Epoch 190/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.7137 - mae: 16.2033\n",
      "Epoch 191/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.7204 - mae: 16.2009\n",
      "Epoch 192/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.6974 - mae: 16.1873\n",
      "Epoch 193/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 15.7416 - mae: 16.2163\n",
      "Epoch 194/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.6878 - mae: 16.1823\n",
      "Epoch 195/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.6121 - mae: 16.1007\n",
      "Epoch 196/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.5604 - mae: 16.0701\n",
      "Epoch 197/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.6734 - mae: 16.1587\n",
      "Epoch 198/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.5673 - mae: 16.0423\n",
      "Epoch 199/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.5949 - mae: 16.0968\n",
      "Epoch 200/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.9019 - mae: 16.3947\n",
      "Epoch 201/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 15.6460 - mae: 16.1401\n",
      "Epoch 202/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.6982 - mae: 16.1902\n",
      "Epoch 203/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.5902 - mae: 16.0676\n",
      "Epoch 204/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 15.5575 - mae: 16.0510\n",
      "Epoch 205/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.7636 - mae: 16.2543\n",
      "Epoch 206/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 15.6660 - mae: 16.1560\n",
      "Epoch 207/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.6877 - mae: 16.1737\n",
      "Epoch 208/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.5350 - mae: 16.0408\n",
      "Epoch 209/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.5797 - mae: 16.0825\n",
      "Epoch 210/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.5327 - mae: 16.0266\n",
      "Epoch 211/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.6913 - mae: 16.1840\n",
      "Epoch 212/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.6494 - mae: 16.1488\n",
      "Epoch 213/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.4931 - mae: 15.9932\n",
      "Epoch 214/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.5378 - mae: 16.0336\n",
      "Epoch 215/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.6424 - mae: 16.1403\n",
      "Epoch 216/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.7640 - mae: 16.2661\n",
      "Epoch 217/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.6316 - mae: 16.1254\n",
      "Epoch 218/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.6147 - mae: 16.1205\n",
      "Epoch 219/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.4534 - mae: 15.9442\n",
      "Epoch 220/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 15.5368 - mae: 16.0303\n",
      "Epoch 221/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.4471 - mae: 15.9630\n",
      "Epoch 222/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.5618 - mae: 16.0326\n",
      "Epoch 223/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.4318 - mae: 15.9311\n",
      "Epoch 224/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.5144 - mae: 16.0046\n",
      "Epoch 225/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.4764 - mae: 15.9636\n",
      "Epoch 226/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.5431 - mae: 16.0503\n",
      "Epoch 227/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.6171 - mae: 16.0953\n",
      "Epoch 228/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30/30 [==============================] - 1s 19ms/step - loss: 15.6354 - mae: 16.1423\n",
      "Epoch 229/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.3901 - mae: 15.8940\n",
      "Epoch 230/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.4471 - mae: 15.9241\n",
      "Epoch 231/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.4026 - mae: 15.8978\n",
      "Epoch 232/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.3789 - mae: 15.8760\n",
      "Epoch 233/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 15.5286 - mae: 16.0276\n",
      "Epoch 234/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.8534 - mae: 16.3528\n",
      "Epoch 235/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.5380 - mae: 16.0203\n",
      "Epoch 236/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.3686 - mae: 15.8514\n",
      "Epoch 237/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.4117 - mae: 15.8956\n",
      "Epoch 238/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.3775 - mae: 15.8802\n",
      "Epoch 239/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.4386 - mae: 15.9243\n",
      "Epoch 240/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.6793 - mae: 16.1530\n",
      "Epoch 241/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.3029 - mae: 15.8104\n",
      "Epoch 242/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.4105 - mae: 15.8997\n",
      "Epoch 243/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.4368 - mae: 15.9248\n",
      "Epoch 244/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.3747 - mae: 15.8521\n",
      "Epoch 245/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.4066 - mae: 15.9023\n",
      "Epoch 246/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 15.3386 - mae: 15.8057\n",
      "Epoch 247/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.3649 - mae: 15.8302\n",
      "Epoch 248/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.4577 - mae: 15.9526\n",
      "Epoch 249/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.4046 - mae: 15.9109\n",
      "Epoch 250/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.4749 - mae: 15.9607\n",
      "Epoch 251/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.2876 - mae: 15.7536\n",
      "Epoch 252/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 15.3791 - mae: 15.8577\n",
      "Epoch 253/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.3434 - mae: 15.8345\n",
      "Epoch 254/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.2984 - mae: 15.7925\n",
      "Epoch 255/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.2934 - mae: 15.7532\n",
      "Epoch 256/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.2508 - mae: 15.7422\n",
      "Epoch 257/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.3995 - mae: 15.9053\n",
      "Epoch 258/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.3922 - mae: 15.8694\n",
      "Epoch 259/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 15.3532 - mae: 15.8477\n",
      "Epoch 260/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.2220 - mae: 15.7146\n",
      "Epoch 261/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.3940 - mae: 15.8832\n",
      "Epoch 262/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.2477 - mae: 15.7434\n",
      "Epoch 263/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.2732 - mae: 15.7612\n",
      "Epoch 264/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.2952 - mae: 15.7943\n",
      "Epoch 265/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.2135 - mae: 15.6938\n",
      "Epoch 266/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.2492 - mae: 15.7378\n",
      "Epoch 267/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.3157 - mae: 15.8104\n",
      "Epoch 268/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.2287 - mae: 15.7163\n",
      "Epoch 269/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.2858 - mae: 15.7754\n",
      "Epoch 270/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.1862 - mae: 15.6785\n",
      "Epoch 271/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.1946 - mae: 15.6670\n",
      "Epoch 272/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 15.1730 - mae: 15.6529\n",
      "Epoch 273/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.2332 - mae: 15.7202\n",
      "Epoch 274/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.2255 - mae: 15.7345\n",
      "Epoch 275/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.2676 - mae: 15.7326\n",
      "Epoch 276/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.2324 - mae: 15.7478\n",
      "Epoch 277/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.5420 - mae: 16.0270\n",
      "Epoch 278/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 15.2639 - mae: 15.7611\n",
      "Epoch 279/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.2120 - mae: 15.6910\n",
      "Epoch 280/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.1504 - mae: 15.6528\n",
      "Epoch 281/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.2292 - mae: 15.7201\n",
      "Epoch 282/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.1260 - mae: 15.6125\n",
      "Epoch 283/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 15.1569 - mae: 15.6434\n",
      "Epoch 284/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.1057 - mae: 15.6036\n",
      "Epoch 285/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 15.1686 - mae: 15.6597\n",
      "Epoch 286/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.1167 - mae: 15.6162\n",
      "Epoch 287/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.1729 - mae: 15.6702\n",
      "Epoch 288/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.1847 - mae: 15.7136\n",
      "Epoch 289/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.4974 - mae: 15.9804\n",
      "Epoch 290/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.2679 - mae: 15.7733\n",
      "Epoch 291/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 15.0709 - mae: 15.5614\n",
      "Epoch 292/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.2493 - mae: 15.7256\n",
      "Epoch 293/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.0438 - mae: 15.5466\n",
      "Epoch 294/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.1778 - mae: 15.6574\n",
      "Epoch 295/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 15.0512 - mae: 15.5478\n",
      "Epoch 296/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.0382 - mae: 15.5296\n",
      "Epoch 297/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 15.1076 - mae: 15.6006\n",
      "Epoch 298/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 15.0530 - mae: 15.5351\n",
      "Epoch 299/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.0384 - mae: 15.5321\n",
      "Epoch 300/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.1299 - mae: 15.6386\n",
      "Epoch 301/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.1064 - mae: 15.5970\n",
      "Epoch 302/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.9790 - mae: 15.4605\n",
      "Epoch 303/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.0110 - mae: 15.5066\n",
      "Epoch 304/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 15.1058 - mae: 15.5964\n",
      "Epoch 305/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.2375 - mae: 15.7219\n",
      "Epoch 306/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.2309 - mae: 15.7301\n",
      "Epoch 307/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 15.2672 - mae: 15.7602\n",
      "Epoch 308/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.0008 - mae: 15.4953\n",
      "Epoch 309/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 15.0123 - mae: 15.5081\n",
      "Epoch 310/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30/30 [==============================] - 1s 20ms/step - loss: 15.0611 - mae: 15.5571\n",
      "Epoch 311/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.1852 - mae: 15.6736\n",
      "Epoch 312/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.1233 - mae: 15.6069\n",
      "Epoch 313/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.9947 - mae: 15.4658\n",
      "Epoch 314/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.0943 - mae: 15.5847\n",
      "Epoch 315/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 14.9568 - mae: 15.4439\n",
      "Epoch 316/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.1041 - mae: 15.5739\n",
      "Epoch 317/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 15.0354 - mae: 15.5422\n",
      "Epoch 318/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.1419 - mae: 15.6546\n",
      "Epoch 319/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.2093 - mae: 15.7089\n",
      "Epoch 320/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.0532 - mae: 15.5250\n",
      "Epoch 321/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.9583 - mae: 15.4560\n",
      "Epoch 322/500\n",
      "30/30 [==============================] - 1s 18ms/step - loss: 14.9700 - mae: 15.4912\n",
      "Epoch 323/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.0725 - mae: 15.5531\n",
      "Epoch 324/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.8652 - mae: 15.3667\n",
      "Epoch 325/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.8929 - mae: 15.3822\n",
      "Epoch 326/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.9265 - mae: 15.4172\n",
      "Epoch 327/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.9509 - mae: 15.4333\n",
      "Epoch 328/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.9134 - mae: 15.4005\n",
      "Epoch 329/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.9116 - mae: 15.4034\n",
      "Epoch 330/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.0927 - mae: 15.5808\n",
      "Epoch 331/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.9286 - mae: 15.4107\n",
      "Epoch 332/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.8675 - mae: 15.3472\n",
      "Epoch 333/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.8702 - mae: 15.3542\n",
      "Epoch 334/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.8431 - mae: 15.3447\n",
      "Epoch 335/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.9312 - mae: 15.4079\n",
      "Epoch 336/500\n",
      "30/30 [==============================] - 1s 24ms/step - loss: 14.9550 - mae: 15.4394\n",
      "Epoch 337/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 14.8542 - mae: 15.3460\n",
      "Epoch 338/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 14.8747 - mae: 15.3412\n",
      "Epoch 339/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 14.8881 - mae: 15.3942\n",
      "Epoch 340/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 14.9567 - mae: 15.4468\n",
      "Epoch 341/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.8472 - mae: 15.3563\n",
      "Epoch 342/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.0139 - mae: 15.4986\n",
      "Epoch 343/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.8054 - mae: 15.2724\n",
      "Epoch 344/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.8408 - mae: 15.3236\n",
      "Epoch 345/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.1904 - mae: 15.6899\n",
      "Epoch 346/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.8436 - mae: 16.3201\n",
      "Epoch 347/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.0402 - mae: 15.5439\n",
      "Epoch 348/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 15.0436 - mae: 15.5271\n",
      "Epoch 349/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.9030 - mae: 15.3948\n",
      "Epoch 350/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.8029 - mae: 15.2866\n",
      "Epoch 351/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.7414 - mae: 15.2338\n",
      "Epoch 352/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.7562 - mae: 15.2514\n",
      "Epoch 353/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.7810 - mae: 15.2596\n",
      "Epoch 354/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 14.9590 - mae: 15.4453\n",
      "Epoch 355/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.9221 - mae: 15.4104\n",
      "Epoch 356/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.7066 - mae: 15.2044\n",
      "Epoch 357/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.7547 - mae: 15.2277\n",
      "Epoch 358/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.7871 - mae: 15.2819\n",
      "Epoch 359/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.9925 - mae: 15.4762\n",
      "Epoch 360/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.6944 - mae: 15.1641\n",
      "Epoch 361/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 14.7717 - mae: 15.2543\n",
      "Epoch 362/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.7487 - mae: 15.2455\n",
      "Epoch 363/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.9341 - mae: 15.4244\n",
      "Epoch 364/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 15.3228 - mae: 15.8026\n",
      "Epoch 365/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.9621 - mae: 15.4402\n",
      "Epoch 366/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.8252 - mae: 15.3161\n",
      "Epoch 367/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 14.7025 - mae: 15.2130\n",
      "Epoch 368/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.7984 - mae: 15.2834\n",
      "Epoch 369/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.7647 - mae: 15.2504\n",
      "Epoch 370/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.7472 - mae: 15.2351\n",
      "Epoch 371/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.7403 - mae: 15.2457\n",
      "Epoch 372/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 15.7732 - mae: 16.2697\n",
      "Epoch 373/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.7542 - mae: 15.2402\n",
      "Epoch 374/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 14.7037 - mae: 15.1962\n",
      "Epoch 375/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 14.7204 - mae: 15.2039\n",
      "Epoch 376/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.8089 - mae: 15.3117\n",
      "Epoch 377/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.7244 - mae: 15.2306\n",
      "Epoch 378/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.6735 - mae: 15.1639\n",
      "Epoch 379/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 14.7379 - mae: 15.2276\n",
      "Epoch 380/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.7652 - mae: 15.2348\n",
      "Epoch 381/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.6328 - mae: 15.1251\n",
      "Epoch 382/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.5653 - mae: 15.0638\n",
      "Epoch 383/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.7895 - mae: 15.2769\n",
      "Epoch 384/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.6107 - mae: 15.0813\n",
      "Epoch 385/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 14.6333 - mae: 15.1318\n",
      "Epoch 386/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.7217 - mae: 15.2036\n",
      "Epoch 387/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.6755 - mae: 15.1613\n",
      "Epoch 388/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.7232 - mae: 15.2257\n",
      "Epoch 389/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.6816 - mae: 15.1675\n",
      "Epoch 390/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 14.6509 - mae: 15.1606\n",
      "Epoch 391/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.7051 - mae: 15.2003\n",
      "Epoch 392/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30/30 [==============================] - 1s 22ms/step - loss: 14.5510 - mae: 15.0429\n",
      "Epoch 393/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.6151 - mae: 15.1203\n",
      "Epoch 394/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.7421 - mae: 15.2329\n",
      "Epoch 395/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 14.5712 - mae: 15.0576\n",
      "Epoch 396/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 14.6044 - mae: 15.0926\n",
      "Epoch 397/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 14.7058 - mae: 15.1888\n",
      "Epoch 398/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 14.6495 - mae: 15.1395\n",
      "Epoch 399/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 14.6612 - mae: 15.1632\n",
      "Epoch 400/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.5517 - mae: 15.0635\n",
      "Epoch 401/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 14.8218 - mae: 15.3183\n",
      "Epoch 402/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.6694 - mae: 15.1747\n",
      "Epoch 403/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 14.8557 - mae: 15.3438\n",
      "Epoch 404/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.5199 - mae: 15.0175\n",
      "Epoch 405/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 14.5631 - mae: 15.0555\n",
      "Epoch 406/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.5306 - mae: 15.0016\n",
      "Epoch 407/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.5798 - mae: 15.0612\n",
      "Epoch 408/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.5897 - mae: 15.0791\n",
      "Epoch 409/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.4423 - mae: 14.9337\n",
      "Epoch 410/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.4861 - mae: 14.9750\n",
      "Epoch 411/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.5891 - mae: 15.0772\n",
      "Epoch 412/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 14.4991 - mae: 14.9859\n",
      "Epoch 413/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.5282 - mae: 15.0174\n",
      "Epoch 414/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.4977 - mae: 14.9871\n",
      "Epoch 415/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.5512 - mae: 15.0299\n",
      "Epoch 416/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 14.6266 - mae: 15.1223\n",
      "Epoch 417/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 14.4645 - mae: 14.9419\n",
      "Epoch 418/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.4594 - mae: 14.9471\n",
      "Epoch 419/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.4586 - mae: 14.9254\n",
      "Epoch 420/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.4575 - mae: 14.9387\n",
      "Epoch 421/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.5301 - mae: 15.0228\n",
      "Epoch 422/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.4954 - mae: 14.9755\n",
      "Epoch 423/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.4116 - mae: 14.8976\n",
      "Epoch 424/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.4655 - mae: 14.9417\n",
      "Epoch 425/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 14.5348 - mae: 15.0348\n",
      "Epoch 426/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.4765 - mae: 14.9818\n",
      "Epoch 427/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.5378 - mae: 15.0150\n",
      "Epoch 428/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.4993 - mae: 14.9812\n",
      "Epoch 429/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.4878 - mae: 14.9773\n",
      "Epoch 430/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.3889 - mae: 14.8805\n",
      "Epoch 431/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.3871 - mae: 14.8778\n",
      "Epoch 432/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.5259 - mae: 14.9988\n",
      "Epoch 433/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.5124 - mae: 14.9844\n",
      "Epoch 434/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.4314 - mae: 14.9190\n",
      "Epoch 435/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.3571 - mae: 14.8428\n",
      "Epoch 436/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.3620 - mae: 14.8567\n",
      "Epoch 437/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.5331 - mae: 15.0314\n",
      "Epoch 438/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.4744 - mae: 14.9540\n",
      "Epoch 439/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.4114 - mae: 14.8984\n",
      "Epoch 440/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.4921 - mae: 14.9801\n",
      "Epoch 441/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.6266 - mae: 15.1135\n",
      "Epoch 442/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.6631 - mae: 15.1493\n",
      "Epoch 443/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.4121 - mae: 14.8911\n",
      "Epoch 444/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.3333 - mae: 14.8432\n",
      "Epoch 445/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.3667 - mae: 14.8502\n",
      "Epoch 446/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.2975 - mae: 14.8071\n",
      "Epoch 447/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.3627 - mae: 14.8531\n",
      "Epoch 448/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.5187 - mae: 15.0182\n",
      "Epoch 449/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.4866 - mae: 14.9718\n",
      "Epoch 450/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.3973 - mae: 14.8882\n",
      "Epoch 451/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 14.4909 - mae: 14.9836\n",
      "Epoch 452/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.3846 - mae: 14.8634\n",
      "Epoch 453/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.3186 - mae: 14.8003\n",
      "Epoch 454/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.2476 - mae: 14.7352\n",
      "Epoch 455/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.5060 - mae: 14.9923\n",
      "Epoch 456/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.3244 - mae: 14.8261\n",
      "Epoch 457/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.7338 - mae: 15.2276\n",
      "Epoch 458/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.4303 - mae: 14.9295\n",
      "Epoch 459/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.4100 - mae: 14.9039\n",
      "Epoch 460/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.3607 - mae: 14.8498\n",
      "Epoch 461/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.2623 - mae: 14.7581\n",
      "Epoch 462/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.2737 - mae: 14.7632\n",
      "Epoch 463/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.2722 - mae: 14.7568\n",
      "Epoch 464/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.2791 - mae: 14.7506\n",
      "Epoch 465/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.2963 - mae: 14.7746\n",
      "Epoch 466/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.3657 - mae: 14.8392\n",
      "Epoch 467/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.5026 - mae: 14.9971\n",
      "Epoch 468/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.2674 - mae: 14.7477\n",
      "Epoch 469/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.2395 - mae: 14.7259\n",
      "Epoch 470/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.2147 - mae: 14.7095\n",
      "Epoch 471/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.2453 - mae: 14.7342\n",
      "Epoch 472/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.3614 - mae: 14.8487\n",
      "Epoch 473/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.2717 - mae: 14.7336\n",
      "Epoch 474/500\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30/30 [==============================] - 1s 19ms/step - loss: 14.2179 - mae: 14.7025\n",
      "Epoch 475/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.1687 - mae: 14.6600\n",
      "Epoch 476/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 14.2302 - mae: 14.7166\n",
      "Epoch 477/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.2482 - mae: 14.7355\n",
      "Epoch 478/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 14.2978 - mae: 14.8003\n",
      "Epoch 479/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.2005 - mae: 14.6886\n",
      "Epoch 480/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.1963 - mae: 14.6868\n",
      "Epoch 481/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 14.2687 - mae: 14.7501\n",
      "Epoch 482/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 14.2422 - mae: 14.7346\n",
      "Epoch 483/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.1654 - mae: 14.6592\n",
      "Epoch 484/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.2320 - mae: 14.7071\n",
      "Epoch 485/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.1879 - mae: 14.6922\n",
      "Epoch 486/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.1477 - mae: 14.6340\n",
      "Epoch 487/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.2663 - mae: 14.7695\n",
      "Epoch 488/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.2665 - mae: 14.7457\n",
      "Epoch 489/500\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 14.1536 - mae: 14.6474\n",
      "Epoch 490/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 14.4426 - mae: 14.9262\n",
      "Epoch 491/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 14.1252 - mae: 14.6228\n",
      "Epoch 492/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.1726 - mae: 14.6566\n",
      "Epoch 493/500\n",
      "30/30 [==============================] - 1s 23ms/step - loss: 14.1981 - mae: 14.6979\n",
      "Epoch 494/500\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 14.2621 - mae: 14.7355\n",
      "Epoch 495/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.1924 - mae: 14.6770\n",
      "Epoch 496/500\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 14.1687 - mae: 14.6640\n",
      "Epoch 497/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.1660 - mae: 14.6710\n",
      "Epoch 498/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.2021 - mae: 14.6990\n",
      "Epoch 499/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.4419 - mae: 14.9272\n",
      "Epoch 500/500\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 14.2902 - mae: 14.7806\n"
     ]
    }
   ],
   "source": [
    "tf.keras.backend.clear_session()\n",
    "tf.random.set_seed(51)\n",
    "np.random.seed(51)\n",
    "train_set = windowed_dataset(x_train, window_size=60, batch_size=100, shuffle_buffer=shuffle_buffer_size)\n",
    "model = tf.keras.models.Sequential([\n",
    "  tf.keras.layers.Conv1D(filters=60, kernel_size=5,\n",
    "                      strides=1, padding=\"causal\",\n",
    "                      activation=\"relu\",\n",
    "                      input_shape=[None, 1]),\n",
    "  tf.keras.layers.LSTM(60, return_sequences=True),\n",
    "  tf.keras.layers.LSTM(60, return_sequences=True),\n",
    "  tf.keras.layers.Dense(30, activation=\"relu\"),\n",
    "  tf.keras.layers.Dense(10, activation=\"relu\"),\n",
    "  tf.keras.layers.Dense(1),\n",
    "  tf.keras.layers.Lambda(lambda x: x * 400)\n",
    "])\n",
    "\n",
    "\n",
    "optimizer = tf.keras.optimizers.SGD(lr=1e-5, momentum=0.9)\n",
    "model.compile(loss=tf.keras.losses.Huber(),\n",
    "              optimizer=optimizer,\n",
    "              metrics=[\"mae\"])\n",
    "\n",
    "with tf.device('/device:GPU:0'):\n",
    "    history = model.fit(train_set,epochs=500)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-09-24T17:02:44.108674Z",
     "start_time": "2019-09-24T17:02:42.718179Z"
    },
    "colab": {},
    "colab_type": "code",
    "id": "GaC6NNMRp0lb"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:Entity <function model_forecast.<locals>.<lambda> at 0x7f6f62f35290> could not be transformed and will be executed as-is. Please report this to the AutoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: module 'gast' has no attribute 'Str'\n",
      "WARNING: Entity <function model_forecast.<locals>.<lambda> at 0x7f6f62f35290> could not be transformed and will be executed as-is. Please report this to the AutoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: module 'gast' has no attribute 'Str'\n"
     ]
    }
   ],
   "source": [
    "rnn_forecast = model_forecast(model, series[..., np.newaxis], window_size)\n",
    "rnn_forecast = rnn_forecast[split_time - window_size:-1, -1, 0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-09-24T17:02:44.384864Z",
     "start_time": "2019-09-24T17:02:44.110394Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 392
    },
    "colab_type": "code",
    "id": "PrktQX3hKYex",
    "outputId": "3126e6a2-6f0a-4501-a5ea-360aab623062"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10, 6))\n",
    "plot_series(time_valid, x_valid)\n",
    "plot_series(time_valid, rnn_forecast)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-09-24T17:02:44.391230Z",
     "start_time": "2019-09-24T17:02:44.386376Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "13XrorC5wQoE",
    "outputId": "392b1fbb-cc0c-4c3f-b246-73de03c381dc"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "16.713465"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tf.keras.metrics.mean_absolute_error(x_valid, rnn_forecast).numpy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-09-24T17:02:44.903704Z",
     "start_time": "2019-09-24T17:02:44.392551Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 609
    },
    "colab_type": "code",
    "id": "MD2kyYUVt3O0",
    "outputId": "4d4ce37c-ebd7-43cc-d0f3-761c37774bda"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.image  as mpimg\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "#-----------------------------------------------------------\n",
    "# Retrieve a list of list results on training and test data\n",
    "# sets for each training epoch\n",
    "#-----------------------------------------------------------\n",
    "loss=history.history['loss']\n",
    "\n",
    "epochs=range(len(loss)) # Get number of epochs\n",
    "\n",
    "\n",
    "#------------------------------------------------\n",
    "# Plot training and validation loss per epoch\n",
    "#------------------------------------------------\n",
    "plt.plot(epochs, loss, 'r')\n",
    "plt.title('Training loss')\n",
    "plt.xlabel(\"Epochs\")\n",
    "plt.ylabel(\"Loss\")\n",
    "plt.legend([\"Loss\"])\n",
    "\n",
    "plt.figure()\n",
    "\n",
    "\n",
    "\n",
    "zoomed_loss = loss[200:]\n",
    "zoomed_epochs = range(200,500)\n",
    "\n",
    "\n",
    "#------------------------------------------------\n",
    "# Plot training and validation loss per epoch\n",
    "#------------------------------------------------\n",
    "plt.plot(zoomed_epochs, zoomed_loss, 'r')\n",
    "plt.title('Training loss')\n",
    "plt.xlabel(\"Epochs\")\n",
    "plt.ylabel(\"Loss\")\n",
    "plt.legend([\"Loss\"])\n",
    "\n",
    "plt.figure()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-09-24T17:02:44.914983Z",
     "start_time": "2019-09-24T17:02:44.905897Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "colab_type": "code",
    "id": "AOVzQXxCwkzP",
    "outputId": "a68bf3ce-bb7d-4759-b51e-8eea04dc138d"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[132.39406   112.78627   120.04337   145.41568   130.55598   157.03128\n",
      " 183.12141   173.30356   159.86633   134.42693   176.71802   173.98831\n",
      " 175.61153   173.70627   216.28029   223.81328   219.36989   214.78069\n",
      " 206.83441   186.2329    203.68369   204.7134    214.49535   212.84723\n",
      " 173.75735   158.47885   151.93634   167.68198   169.65515   140.87003\n",
      " 197.43643   137.55527   132.02121   194.01651   159.63382   156.57611\n",
      " 178.5815    146.41219   135.01994   145.65503   178.72993   175.31612\n",
      " 166.38628   163.79926   180.16513   165.72029   187.5343    166.06955\n",
      " 162.34784   141.75885   109.99956   114.48419   103.414986  100.17789\n",
      " 113.49175   110.195145  101.00628    98.453026   91.031334   79.062706\n",
      "  70.09896    67.39182    76.12046    77.33553    67.77344    69.45995\n",
      "  69.97802    78.11617    73.95269    64.46089    65.54299    70.56998\n",
      "  54.999268   52.775105   49.66294    48.025757   47.587234   60.002457\n",
      "  42.20348    41.83481    45.376007   30.029684   23.759619   25.518333\n",
      "  32.45274    17.654694   17.10551    13.627184   11.108663   12.771288\n",
      "  23.494703   25.367397   21.217957   19.93902    14.672168   17.99605\n",
      "  15.038149   16.310284   12.89636    10.283625    9.55708     9.968388\n",
      "  12.315084    8.98063    10.043265    9.078982    8.671172    8.073385\n",
      "  10.44836     4.659347    3.0851974   9.842518    8.56297     7.5472383\n",
      "   8.6813345   8.814947    6.4493284   6.479862    7.3563156   8.318872\n",
      "   8.085336    7.396306    6.30987     4.725218    3.2765894   2.7767448\n",
      "   3.3965707   3.98694     3.1308935   3.730729    5.283299    5.758859\n",
      "   9.13222    13.351272   24.445623   25.338799   28.693966   34.476746\n",
      "  34.57684    25.960827   32.513405   41.273872   46.87179    54.841602\n",
      "  56.517105   56.903946   57.44937    68.07315    81.160614   79.24189\n",
      "  76.69603    79.45963    86.24124   104.02198    96.81924   106.2032\n",
      " 109.22828    99.450905   71.40992   104.35477   114.22244   105.51558\n",
      " 112.782715  109.708694  105.28331   101.37634    90.89383    93.13653\n",
      "  80.69733    98.25609    99.42064    94.50509   110.6673     91.7499\n",
      "  81.94328    79.23122    83.39551    77.00644   108.814      96.03214\n",
      "  94.61139   101.59859   114.27413   115.30251   106.291794  112.00078\n",
      " 106.76189    98.584915  100.99409   111.20226   104.1764     98.58725\n",
      " 123.902275  106.08931   105.120514   83.61785    88.37969    97.84444\n",
      "  86.71745    83.37334    76.47836    82.8193     74.33712    66.68258\n",
      "  64.68637    64.25146    60.172123   57.530594   51.302326   55.546333\n",
      "  44.65448    46.20221    39.168964   25.627098   30.91723    30.32552\n",
      "  31.706451   24.905085   24.429522   17.313898   23.276907   17.75652\n",
      "  14.460105   15.802214   16.57652    20.060698   13.8686     13.980672\n",
      "  10.554935    7.9983654   7.791856    5.069023    7.502687    7.8458457\n",
      "   9.714435 ]\n"
     ]
    }
   ],
   "source": [
    "print(rnn_forecast)"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "S+P Week 4 Lesson 5.ipynb",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  },
  "latex_envs": {
   "LaTeX_envs_menu_present": true,
   "autoclose": false,
   "autocomplete": true,
   "bibliofile": "biblio.bib",
   "cite_by": "apalike",
   "current_citInitial": 1,
   "eqLabelWithNumbers": true,
   "eqNumInitial": 1,
   "hotkeys": {
    "equation": "Ctrl-E",
    "itemize": "Ctrl-I"
   },
   "labels_anchors": false,
   "latex_user_defs": false,
   "report_style_numbering": false,
   "user_envs_cfg": false
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}