FFT_Variance_ANN_Keras_model.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "view-in-github"
   },
   "source": [
    "<a href=\"https://colab.research.google.com/github/fbasatemur/FFT-Variance_ANN_with_CUDA_CPP/blob/master/FFT_Variance_ANN_Tasarim.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "X2vArminORbm"
   },
   "outputs": [],
   "source": [
    "# -*- coding: utf-8 -*-\n",
    "\"\"\"\n",
    "Created on Sat Oct 24 15:46:07 2020\n",
    "\n",
    "@author: fbasatemur\n",
    "\"\"\"\n",
    "\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout, BatchNormalization\n",
    "from keras.callbacks import ModelCheckpoint\n",
    "from keras.losses import SparseCategoricalCrossentropy\n",
    "from keras.models import model_from_json\n",
    "from keras.models import load_model\n",
    "from keras import layers, activations\n",
    "from matplotlib import pyplot as plt\n",
    "\n",
    "\n",
    " \n",
    "#%% ANN model\n",
    "def createModel():\n",
    "      \n",
    "      model = Sequential()\n",
    "      \n",
    "      model.add(Dense(100, activation=\"relu\", input_shape=(625,)))  # 25x25 FFT\n",
    "      model.add(BatchNormalization()) \n",
    "      \n",
    "      model.add(Dense(100, activation=\"relu\"))\n",
    "      model.add(BatchNormalization())\n",
    "      \n",
    "      model.add(Dense(100, activation=\"relu\"))\n",
    "      model.add(BatchNormalization())\n",
    "      \n",
    "      model.add(Dense(100, activation=\"relu\"))\n",
    "      model.add(BatchNormalization())\n",
    "      \n",
    "      model.add(Dense(100, activation=\"relu\"))\n",
    "      model.add(BatchNormalization())\n",
    "      \n",
    "      model.add(Dense(100, activation=\"relu\"))\n",
    "      model.add(BatchNormalization())\n",
    "\n",
    "      model.add(Dense(100, activation=\"relu\"))\n",
    "      model.add(BatchNormalization())\n",
    "      \n",
    "      model.add(Dense(20, activation=\"relu\"))\n",
    "      model.add(BatchNormalization())\n",
    "      \n",
    "      model.add(Dense(1, activation=\"sigmoid\"))   # unipolar sigmoid (0-1)\n",
    "      \n",
    "      model.compile(optimizer='Adam', loss='mean_squared_error', metrics=['mae'])\n",
    "      \n",
    "      return model\n",
    "\n",
    "def showAccuracy(history):\n",
    "      \n",
    "      for key in history.history.keys():  \n",
    "            plt.plot(history.history[key], label = key)\n",
    "    \n",
    "      plt.xlabel(\"epochs\")\n",
    "      plt.ylabel(\"values\")\n",
    "      \n",
    "      plt.legend()\n",
    "      plt.show() \n",
    "      \n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "_vw05SUkMPNQ",
    "outputId": "82590c3e-824c-418c-bf74-48fdaf421c60"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mounted at /content/drive\n"
     ]
    }
   ],
   "source": [
    "# drive access\n",
    "\n",
    "from google.colab import drive\n",
    "drive.mount('/content/drive')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "Vs9JMTpTMrtI",
    "outputId": "de318c04-83d7-4ae8-8d20-ecc5b5a4253f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "drive  sample_data\n"
     ]
    }
   ],
   "source": [
    "!ls"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "MiBV4LD0EyQC"
   },
   "outputs": [],
   "source": [
    "#%% read file \n",
    "\n",
    "import numpy as np\n",
    "\n",
    "fftValues = []\n",
    "varianceValues = []\n",
    "imagePaths = []\n",
    "\n",
    "# drive_path_FFT = \"drive/My Drive/Colab Notebooks/FFT_Variance_ANN_Tasarim_Donemi/25_NORM_FFT.txt\" \n",
    "# drive_path_VAR = \"drive/My Drive/Colab Notebooks/FFT_Variance_ANN_Tasarim_Donemi/100_Variance.txt\" \n",
    "# drive_path_IMGPATHS = \"drive/My Drive/Colab Notebooks/FFT_Variance_ANN_Tasarim_Donemi/Image_Paths.txt\" \n",
    "\n",
    "drive_path_FFT = \"drive/My Drive/Colab Notebooks/FFT_Variance_ANN_Tasarim_Donemi/25_NORM_FFT.txt\" \n",
    "drive_path_VAR = \"drive/My Drive/Colab Notebooks/FFT_Variance_ANN_Tasarim_Donemi/Variance.txt\" \n",
    "drive_path_IMGPATHS = \"drive/My Drive/Colab Notebooks/FFT_Variance_ANN_Tasarim_Donemi/Image_Paths.txt\" \n",
    "\n",
    "with open(drive_path_FFT, \"r\") as fileFFT:\n",
    "    fftValues = [line.split() for line in fileFFT]\n",
    "\n",
    "with open(drive_path_VAR, \"r\") as fileVar:\n",
    "    for line in fileVar:\n",
    "        varianceValues.append(line.split()[0])\n",
    "\n",
    "with open(drive_path_IMGPATHS, \"r\") as filePaths:\n",
    "    for line in filePaths:\n",
    "        imagePaths.append(line.split('\\\\')[-1])\n",
    "\n",
    "fileFFT.close()\n",
    "fileVar.close()\n",
    "filePaths.close()\n",
    "\n",
    "fftValues = np.array(fftValues)                          \n",
    "fftValues = fftValues.astype('float')\n",
    "\n",
    "varianceValues = np.array(varianceValues)                          \n",
    "varianceValues = varianceValues.astype('float')\n",
    "\n",
    "filePaths = np.array(filePaths)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "xnsalRwDqV8t",
    "outputId": "fe3cef20-8c06-48e5-d0ae-61d971839e1e"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'numpy.float64'>\n"
     ]
    }
   ],
   "source": [
    "print(type(fftValues[0][0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "n0zBaN7Hzqdw"
   },
   "outputs": [],
   "source": [
    "#%% varianceValues tags\n",
    "# varianceValues labeled with images index\n",
    "\n",
    "labeledVarianceData = np.arange(len(varianceValues))\n",
    "labeledVarianceData = np.vstack((varianceValues, labeledVarianceData)).T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "LyhZKuLwE8T3",
    "outputId": "1d0f29a4-d0c0-4c3e-84e3-50b4a81ccd25"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total images: 771\n",
      "Total train images: 693\n",
      "Total test images: 78\n"
     ]
    }
   ],
   "source": [
    "#%% \n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "x_train, x_test, y_train, y_test = train_test_split(fftValues, labeledVarianceData, test_size=0.1, random_state=0)\n",
    "\n",
    "print(\"Total images: \" + str(len(varianceValues)))\n",
    "print(\"Total train images: \" + str(y_train.T[0].shape[0]))\n",
    "print(\"Total test images: \" + str(y_test.T[0].shape[0]))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "8i6diZxxFEug",
    "outputId": "3602fbd9-8b3b-47e1-b2f5-78b6a149694b"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0797 - mae: 0.2193\n",
      "Epoch 2/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0364 - mae: 0.1475\n",
      "Epoch 3/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0229 - mae: 0.1171\n",
      "Epoch 4/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0179 - mae: 0.1031\n",
      "Epoch 5/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0133 - mae: 0.0885\n",
      "Epoch 6/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0132 - mae: 0.0873\n",
      "Epoch 7/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0106 - mae: 0.0803\n",
      "Epoch 8/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0096 - mae: 0.0753\n",
      "Epoch 9/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0091 - mae: 0.0733\n",
      "Epoch 10/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0089 - mae: 0.0742\n",
      "Epoch 11/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0071 - mae: 0.0648\n",
      "Epoch 12/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0077 - mae: 0.0681\n",
      "Epoch 13/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0082 - mae: 0.0694\n",
      "Epoch 14/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0086 - mae: 0.0725\n",
      "Epoch 15/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0069 - mae: 0.0648\n",
      "Epoch 16/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0069 - mae: 0.0632\n",
      "Epoch 17/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0060 - mae: 0.0612\n",
      "Epoch 18/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0070 - mae: 0.0662\n",
      "Epoch 19/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0064 - mae: 0.0635\n",
      "Epoch 20/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0065 - mae: 0.0618\n",
      "Epoch 21/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0056 - mae: 0.0578\n",
      "Epoch 22/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0072 - mae: 0.0649\n",
      "Epoch 23/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0059 - mae: 0.0601\n",
      "Epoch 24/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0059 - mae: 0.0605\n",
      "Epoch 25/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0057 - mae: 0.0573\n",
      "Epoch 26/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0057 - mae: 0.0584\n",
      "Epoch 27/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0054 - mae: 0.0567\n",
      "Epoch 28/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0043 - mae: 0.0506\n",
      "Epoch 29/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0052 - mae: 0.0554\n",
      "Epoch 30/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0040 - mae: 0.0489\n",
      "Epoch 31/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0049 - mae: 0.0547\n",
      "Epoch 32/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0051 - mae: 0.0555\n",
      "Epoch 33/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0050 - mae: 0.0563\n",
      "Epoch 34/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0034 - mae: 0.0467\n",
      "Epoch 35/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0039 - mae: 0.0486\n",
      "Epoch 36/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0037 - mae: 0.0480\n",
      "Epoch 37/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0041 - mae: 0.0509\n",
      "Epoch 38/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0038 - mae: 0.0485\n",
      "Epoch 39/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0045 - mae: 0.0534\n",
      "Epoch 40/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0033 - mae: 0.0436\n",
      "Epoch 41/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0046 - mae: 0.0524\n",
      "Epoch 42/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0036 - mae: 0.0467\n",
      "Epoch 43/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0040 - mae: 0.0477\n",
      "Epoch 44/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0036 - mae: 0.0471\n",
      "Epoch 45/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0043 - mae: 0.0516\n",
      "Epoch 46/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0044 - mae: 0.0501\n",
      "Epoch 47/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0036 - mae: 0.0469\n",
      "Epoch 48/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0040 - mae: 0.0500\n",
      "Epoch 49/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0042 - mae: 0.0514\n",
      "Epoch 50/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0037 - mae: 0.0454\n",
      "Epoch 51/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0030 - mae: 0.0427\n",
      "Epoch 52/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0031 - mae: 0.0443\n",
      "Epoch 53/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0036 - mae: 0.0461\n",
      "Epoch 54/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0023 - mae: 0.0369\n",
      "Epoch 55/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0030 - mae: 0.0431\n",
      "Epoch 56/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0033 - mae: 0.0436\n",
      "Epoch 57/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0030 - mae: 0.0424\n",
      "Epoch 58/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0037 - mae: 0.0490\n",
      "Epoch 59/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0032 - mae: 0.0443\n",
      "Epoch 60/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0030 - mae: 0.0422\n",
      "Epoch 61/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0026 - mae: 0.0388\n",
      "Epoch 62/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0027 - mae: 0.0412\n",
      "Epoch 63/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0028 - mae: 0.0409\n",
      "Epoch 64/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0034 - mae: 0.0450\n",
      "Epoch 65/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0033 - mae: 0.0436\n",
      "Epoch 66/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0024 - mae: 0.0383\n",
      "Epoch 67/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0033 - mae: 0.0452\n",
      "Epoch 68/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0033 - mae: 0.0452\n",
      "Epoch 69/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0024 - mae: 0.0382\n",
      "Epoch 70/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0028 - mae: 0.0424\n",
      "Epoch 71/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0032 - mae: 0.0443\n",
      "Epoch 72/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0025 - mae: 0.0396\n",
      "Epoch 73/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0028 - mae: 0.0414\n",
      "Epoch 74/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0026 - mae: 0.0390\n",
      "Epoch 75/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0032 - mae: 0.0444\n",
      "Epoch 76/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0022 - mae: 0.0376\n",
      "Epoch 77/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0021 - mae: 0.0351\n",
      "Epoch 78/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0028 - mae: 0.0398\n",
      "Epoch 79/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0029 - mae: 0.0422\n",
      "Epoch 80/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0036 - mae: 0.0455\n",
      "Epoch 81/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0034 - mae: 0.0465\n",
      "Epoch 82/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0028 - mae: 0.0410\n",
      "Epoch 83/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0031 - mae: 0.0423\n",
      "Epoch 84/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0034 - mae: 0.0456\n",
      "Epoch 85/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0021 - mae: 0.0356\n",
      "Epoch 86/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0026 - mae: 0.0391\n",
      "Epoch 87/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0027 - mae: 0.0389\n",
      "Epoch 88/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0026 - mae: 0.0402\n",
      "Epoch 89/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0027 - mae: 0.0392\n",
      "Epoch 90/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0025 - mae: 0.0396\n",
      "Epoch 91/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0026 - mae: 0.0404\n",
      "Epoch 92/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0023 - mae: 0.0378\n",
      "Epoch 93/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0021 - mae: 0.0362\n",
      "Epoch 94/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0021 - mae: 0.0355\n",
      "Epoch 95/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0026 - mae: 0.0392\n",
      "Epoch 96/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0024 - mae: 0.0375\n",
      "Epoch 97/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0020 - mae: 0.0331\n",
      "Epoch 98/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0021 - mae: 0.0352\n",
      "Epoch 99/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0030 - mae: 0.0426\n",
      "Epoch 100/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0024 - mae: 0.0377\n",
      "Epoch 101/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0024 - mae: 0.0379\n",
      "Epoch 102/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0025 - mae: 0.0394\n",
      "Epoch 103/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0022 - mae: 0.0359\n",
      "Epoch 104/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0020 - mae: 0.0355\n",
      "Epoch 105/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0017 - mae: 0.0319\n",
      "Epoch 106/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0015 - mae: 0.0310\n",
      "Epoch 107/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0024 - mae: 0.0384\n",
      "Epoch 108/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0022 - mae: 0.0359\n",
      "Epoch 109/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0024 - mae: 0.0380\n",
      "Epoch 110/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0020 - mae: 0.0346\n",
      "Epoch 111/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0025 - mae: 0.0398\n",
      "Epoch 112/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0034 - mae: 0.0442\n",
      "Epoch 113/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0024 - mae: 0.0371\n",
      "Epoch 114/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0022 - mae: 0.0361\n",
      "Epoch 115/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0025 - mae: 0.0379\n",
      "Epoch 116/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0033 - mae: 0.0439\n",
      "Epoch 117/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0031 - mae: 0.0440\n",
      "Epoch 118/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0029 - mae: 0.0413\n",
      "Epoch 119/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0029 - mae: 0.0420\n",
      "Epoch 120/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0032 - mae: 0.0406\n",
      "Epoch 121/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0033 - mae: 0.0434\n",
      "Epoch 122/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0018 - mae: 0.0330\n",
      "Epoch 123/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0022 - mae: 0.0360\n",
      "Epoch 124/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0030 - mae: 0.0416\n",
      "Epoch 125/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0018 - mae: 0.0334\n",
      "Epoch 126/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0020 - mae: 0.0351\n",
      "Epoch 127/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0019 - mae: 0.0348\n",
      "Epoch 128/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0022 - mae: 0.0353\n",
      "Epoch 129/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0020 - mae: 0.0344\n",
      "Epoch 130/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0018 - mae: 0.0330\n",
      "Epoch 131/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0018 - mae: 0.0327\n",
      "Epoch 132/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0289\n",
      "Epoch 133/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0016 - mae: 0.0303\n",
      "Epoch 134/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0019 - mae: 0.0339\n",
      "Epoch 135/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0013 - mae: 0.0284\n",
      "Epoch 136/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0298\n",
      "Epoch 137/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0022 - mae: 0.0351\n",
      "Epoch 138/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0017 - mae: 0.0320\n",
      "Epoch 139/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0028 - mae: 0.0420\n",
      "Epoch 140/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0026 - mae: 0.0392\n",
      "Epoch 141/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0022 - mae: 0.0362\n",
      "Epoch 142/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0018 - mae: 0.0333\n",
      "Epoch 143/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0016 - mae: 0.0311\n",
      "Epoch 144/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0018 - mae: 0.0330\n",
      "Epoch 145/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0018 - mae: 0.0332\n",
      "Epoch 146/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0019 - mae: 0.0332\n",
      "Epoch 147/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0022 - mae: 0.0366\n",
      "Epoch 148/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0024 - mae: 0.0378\n",
      "Epoch 149/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0020 - mae: 0.0359\n",
      "Epoch 150/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0028 - mae: 0.0411\n",
      "Epoch 151/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0021 - mae: 0.0355\n",
      "Epoch 152/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0020 - mae: 0.0349\n",
      "Epoch 153/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0016 - mae: 0.0304\n",
      "Epoch 154/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0026 - mae: 0.0395\n",
      "Epoch 155/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0020 - mae: 0.0344\n",
      "Epoch 156/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0020 - mae: 0.0340\n",
      "Epoch 157/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0024 - mae: 0.0372\n",
      "Epoch 158/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0016 - mae: 0.0312\n",
      "Epoch 159/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0020 - mae: 0.0349\n",
      "Epoch 160/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0022 - mae: 0.0357\n",
      "Epoch 161/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0281\n",
      "Epoch 162/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0022 - mae: 0.0374\n",
      "Epoch 163/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0018 - mae: 0.0326\n",
      "Epoch 164/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0015 - mae: 0.0299\n",
      "Epoch 165/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0018 - mae: 0.0319\n",
      "Epoch 166/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0017 - mae: 0.0320\n",
      "Epoch 167/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0295\n",
      "Epoch 168/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0279\n",
      "Epoch 169/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0317\n",
      "Epoch 170/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0017 - mae: 0.0329\n",
      "Epoch 171/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0036 - mae: 0.0440\n",
      "Epoch 172/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0020 - mae: 0.0348\n",
      "Epoch 173/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0018 - mae: 0.0317\n",
      "Epoch 174/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0021 - mae: 0.0348\n",
      "Epoch 175/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0020 - mae: 0.0350\n",
      "Epoch 176/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0022 - mae: 0.0365\n",
      "Epoch 177/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0020 - mae: 0.0337\n",
      "Epoch 178/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0013 - mae: 0.0293\n",
      "Epoch 179/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0018 - mae: 0.0320\n",
      "Epoch 180/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0296\n",
      "Epoch 181/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0016 - mae: 0.0308\n",
      "Epoch 182/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0292\n",
      "Epoch 183/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0296\n",
      "Epoch 184/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0017 - mae: 0.0322\n",
      "Epoch 185/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0286\n",
      "Epoch 186/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0018 - mae: 0.0329\n",
      "Epoch 187/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0016 - mae: 0.0309\n",
      "Epoch 188/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0020 - mae: 0.0334\n",
      "Epoch 189/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0267\n",
      "Epoch 190/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0294\n",
      "Epoch 191/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0285\n",
      "Epoch 192/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0016 - mae: 0.0316\n",
      "Epoch 193/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0017 - mae: 0.0325\n",
      "Epoch 194/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0272\n",
      "Epoch 195/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0296\n",
      "Epoch 196/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0275\n",
      "Epoch 197/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0257\n",
      "Epoch 198/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0283\n",
      "Epoch 199/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0290\n",
      "Epoch 200/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0290\n",
      "Epoch 201/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0304\n",
      "Epoch 202/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0018 - mae: 0.0316\n",
      "Epoch 203/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0312\n",
      "Epoch 204/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0016 - mae: 0.0313\n",
      "Epoch 205/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0013 - mae: 0.0281\n",
      "Epoch 206/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0290\n",
      "Epoch 207/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0302\n",
      "Epoch 208/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0298\n",
      "Epoch 209/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0287\n",
      "Epoch 210/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0282\n",
      "Epoch 211/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0258\n",
      "Epoch 212/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0284\n",
      "Epoch 213/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0276\n",
      "Epoch 214/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0245\n",
      "Epoch 215/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0016 - mae: 0.0305\n",
      "Epoch 216/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0297\n",
      "Epoch 217/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0017 - mae: 0.0327\n",
      "Epoch 218/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0018 - mae: 0.0322\n",
      "Epoch 219/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0259\n",
      "Epoch 220/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0293\n",
      "Epoch 221/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0267\n",
      "Epoch 222/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0295\n",
      "Epoch 223/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0017 - mae: 0.0317\n",
      "Epoch 224/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0018 - mae: 0.0329\n",
      "Epoch 225/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0019 - mae: 0.0336\n",
      "Epoch 226/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0274\n",
      "Epoch 227/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0011 - mae: 0.0269\n",
      "Epoch 228/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0283\n",
      "Epoch 229/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0017 - mae: 0.0314\n",
      "Epoch 230/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0016 - mae: 0.0298\n",
      "Epoch 231/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0298\n",
      "Epoch 232/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0259\n",
      "Epoch 233/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0277\n",
      "Epoch 234/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0294\n",
      "Epoch 235/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0311\n",
      "Epoch 236/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0016 - mae: 0.0319\n",
      "Epoch 237/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0022 - mae: 0.0368\n",
      "Epoch 238/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0019 - mae: 0.0339\n",
      "Epoch 239/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0021 - mae: 0.0356\n",
      "Epoch 240/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0016 - mae: 0.0314\n",
      "Epoch 241/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0017 - mae: 0.0311\n",
      "Epoch 242/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0020 - mae: 0.0345\n",
      "Epoch 243/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0017 - mae: 0.0321\n",
      "Epoch 244/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0018 - mae: 0.0320\n",
      "Epoch 245/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0020 - mae: 0.0340\n",
      "Epoch 246/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0017 - mae: 0.0319\n",
      "Epoch 247/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0020 - mae: 0.0340\n",
      "Epoch 248/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0020 - mae: 0.0355\n",
      "Epoch 249/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0302\n",
      "Epoch 250/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0280\n",
      "Epoch 251/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0019 - mae: 0.0332\n",
      "Epoch 252/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0017 - mae: 0.0311\n",
      "Epoch 253/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0018 - mae: 0.0332\n",
      "Epoch 254/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0024 - mae: 0.0368\n",
      "Epoch 255/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0018 - mae: 0.0322\n",
      "Epoch 256/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0019 - mae: 0.0330\n",
      "Epoch 257/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0021 - mae: 0.0372\n",
      "Epoch 258/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0012 - mae: 0.0276\n",
      "Epoch 259/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0015 - mae: 0.0301\n",
      "Epoch 260/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0014 - mae: 0.0287\n",
      "Epoch 261/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0287\n",
      "Epoch 262/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0300\n",
      "Epoch 263/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0011 - mae: 0.0255\n",
      "Epoch 264/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0012 - mae: 0.0273\n",
      "Epoch 265/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0011 - mae: 0.0252\n",
      "Epoch 266/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 8.3434e-04 - mae: 0.0228\n",
      "Epoch 267/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0266\n",
      "Epoch 268/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.9984e-04 - mae: 0.0229\n",
      "Epoch 269/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0016 - mae: 0.0315\n",
      "Epoch 270/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 9.0067e-04 - mae: 0.0231\n",
      "Epoch 271/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0252\n",
      "Epoch 272/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0275\n",
      "Epoch 273/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0041 - mae: 0.0435\n",
      "Epoch 274/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0027 - mae: 0.0403\n",
      "Epoch 275/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0028 - mae: 0.0394\n",
      "Epoch 276/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0022 - mae: 0.0367\n",
      "Epoch 277/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0017 - mae: 0.0316\n",
      "Epoch 278/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0016 - mae: 0.0298\n",
      "Epoch 279/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0018 - mae: 0.0331\n",
      "Epoch 280/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0020 - mae: 0.0359\n",
      "Epoch 281/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0018 - mae: 0.0328\n",
      "Epoch 282/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0021 - mae: 0.0357\n",
      "Epoch 283/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0018 - mae: 0.0327\n",
      "Epoch 284/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0277\n",
      "Epoch 285/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0282\n",
      "Epoch 286/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0283\n",
      "Epoch 287/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0012 - mae: 0.0266\n",
      "Epoch 288/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0017 - mae: 0.0315\n",
      "Epoch 289/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0295\n",
      "Epoch 290/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0015 - mae: 0.0297\n",
      "Epoch 291/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0267\n",
      "Epoch 292/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0275\n",
      "Epoch 293/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0284\n",
      "Epoch 294/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0018 - mae: 0.0332\n",
      "Epoch 295/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0019 - mae: 0.0336\n",
      "Epoch 296/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0291\n",
      "Epoch 297/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0296\n",
      "Epoch 298/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 9.9940e-04 - mae: 0.0243\n",
      "Epoch 299/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0244\n",
      "Epoch 300/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0020 - mae: 0.0298\n",
      "Epoch 301/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0047 - mae: 0.0432\n",
      "Epoch 302/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0018 - mae: 0.0340\n",
      "Epoch 303/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0267\n",
      "Epoch 304/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0274\n",
      "Epoch 305/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0012 - mae: 0.0270\n",
      "Epoch 306/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0249\n",
      "Epoch 307/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0264\n",
      "Epoch 308/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0012 - mae: 0.0268\n",
      "Epoch 309/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0015 - mae: 0.0302\n",
      "Epoch 310/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0011 - mae: 0.0262\n",
      "Epoch 311/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0013 - mae: 0.0277\n",
      "Epoch 312/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0262\n",
      "Epoch 313/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0016 - mae: 0.0299\n",
      "Epoch 314/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0279\n",
      "Epoch 315/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0287\n",
      "Epoch 316/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0272\n",
      "Epoch 317/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0254\n",
      "Epoch 318/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0263\n",
      "Epoch 319/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0019 - mae: 0.0335\n",
      "Epoch 320/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0279\n",
      "Epoch 321/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0260\n",
      "Epoch 322/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0249\n",
      "Epoch 323/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0012 - mae: 0.0261\n",
      "Epoch 324/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0293\n",
      "Epoch 325/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0016 - mae: 0.0317\n",
      "Epoch 326/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0255\n",
      "Epoch 327/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.5569e-04 - mae: 0.0229\n",
      "Epoch 328/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.6780e-04 - mae: 0.0234\n",
      "Epoch 329/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0270\n",
      "Epoch 330/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0014 - mae: 0.0271\n",
      "Epoch 331/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0275\n",
      "Epoch 332/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.9076e-04 - mae: 0.0242\n",
      "Epoch 333/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.8668e-04 - mae: 0.0245\n",
      "Epoch 334/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0242\n",
      "Epoch 335/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0252\n",
      "Epoch 336/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.8525e-04 - mae: 0.0236\n",
      "Epoch 337/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 9.0288e-04 - mae: 0.0226\n",
      "Epoch 338/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.6185e-04 - mae: 0.0209\n",
      "Epoch 339/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0245\n",
      "Epoch 340/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0249\n",
      "Epoch 341/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.3325e-04 - mae: 0.0209\n",
      "Epoch 342/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.2470e-04 - mae: 0.0222\n",
      "Epoch 343/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.7029e-04 - mae: 0.0230\n",
      "Epoch 344/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.2327e-04 - mae: 0.0230\n",
      "Epoch 345/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0255\n",
      "Epoch 346/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0306\n",
      "Epoch 347/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0277\n",
      "Epoch 348/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0298\n",
      "Epoch 349/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0287\n",
      "Epoch 350/1000\n",
      "28/28 [==============================] - 0s 7ms/step - loss: 0.0010 - mae: 0.0245\n",
      "Epoch 351/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0012 - mae: 0.0271\n",
      "Epoch 352/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0294\n",
      "Epoch 353/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0027 - mae: 0.0388\n",
      "Epoch 354/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0016 - mae: 0.0309\n",
      "Epoch 355/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0278\n",
      "Epoch 356/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0292\n",
      "Epoch 357/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0265\n",
      "Epoch 358/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0278\n",
      "Epoch 359/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0260\n",
      "Epoch 360/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0255\n",
      "Epoch 361/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0258\n",
      "Epoch 362/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0262\n",
      "Epoch 363/1000\n",
      "28/28 [==============================] - 0s 4ms/step - loss: 0.0014 - mae: 0.0290\n",
      "Epoch 364/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0258\n",
      "Epoch 365/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.3317e-04 - mae: 0.0224\n",
      "Epoch 366/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0018 - mae: 0.0276\n",
      "Epoch 367/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0281\n",
      "Epoch 368/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0256\n",
      "Epoch 369/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.5016e-04 - mae: 0.0234\n",
      "Epoch 370/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0269\n",
      "Epoch 371/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 9.6141e-04 - mae: 0.0236\n",
      "Epoch 372/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 8.0283e-04 - mae: 0.0209\n",
      "Epoch 373/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.7365e-04 - mae: 0.0227\n",
      "Epoch 374/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0285\n",
      "Epoch 375/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0266\n",
      "Epoch 376/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.8094e-04 - mae: 0.0229\n",
      "Epoch 377/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.5141e-04 - mae: 0.0242\n",
      "Epoch 378/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.7859e-04 - mae: 0.0234\n",
      "Epoch 379/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.4847e-04 - mae: 0.0217\n",
      "Epoch 380/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.4786e-04 - mae: 0.0224\n",
      "Epoch 381/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0267\n",
      "Epoch 382/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0246\n",
      "Epoch 383/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0279\n",
      "Epoch 384/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0254\n",
      "Epoch 385/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0255\n",
      "Epoch 386/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.4445e-04 - mae: 0.0220\n",
      "Epoch 387/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.7987e-04 - mae: 0.0233\n",
      "Epoch 388/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.7692e-04 - mae: 0.0218\n",
      "Epoch 389/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.6916e-04 - mae: 0.0234\n",
      "Epoch 390/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0239\n",
      "Epoch 391/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.7432e-04 - mae: 0.0197\n",
      "Epoch 392/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0253\n",
      "Epoch 393/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0240\n",
      "Epoch 394/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0246\n",
      "Epoch 395/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0271\n",
      "Epoch 396/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0016 - mae: 0.0289\n",
      "Epoch 397/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0261\n",
      "Epoch 398/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.4080e-04 - mae: 0.0240\n",
      "Epoch 399/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.4577e-04 - mae: 0.0209\n",
      "Epoch 400/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.4783e-04 - mae: 0.0236\n",
      "Epoch 401/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.5110e-04 - mae: 0.0221\n",
      "Epoch 402/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.9493e-04 - mae: 0.0215\n",
      "Epoch 403/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.7089e-04 - mae: 0.0210\n",
      "Epoch 404/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.4859e-04 - mae: 0.0214\n",
      "Epoch 405/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.7358e-04 - mae: 0.0222\n",
      "Epoch 406/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.9286e-04 - mae: 0.0229\n",
      "Epoch 407/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.6984e-04 - mae: 0.0235\n",
      "Epoch 408/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.3323e-04 - mae: 0.0206\n",
      "Epoch 409/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.9202e-04 - mae: 0.0198\n",
      "Epoch 410/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.0042e-04 - mae: 0.0194\n",
      "Epoch 411/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 9.2230e-04 - mae: 0.0223\n",
      "Epoch 412/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0257\n",
      "Epoch 413/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.5785e-04 - mae: 0.0218\n",
      "Epoch 414/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0011 - mae: 0.0229\n",
      "Epoch 415/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.9164e-04 - mae: 0.0238\n",
      "Epoch 416/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.7679e-04 - mae: 0.0228\n",
      "Epoch 417/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0261\n",
      "Epoch 418/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0270\n",
      "Epoch 419/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0257\n",
      "Epoch 420/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0278\n",
      "Epoch 421/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.0177e-04 - mae: 0.0220\n",
      "Epoch 422/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0254\n",
      "Epoch 423/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0256\n",
      "Epoch 424/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0263\n",
      "Epoch 425/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0242\n",
      "Epoch 426/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.6318e-04 - mae: 0.0221\n",
      "Epoch 427/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.2948e-04 - mae: 0.0206\n",
      "Epoch 428/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0252\n",
      "Epoch 429/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0247\n",
      "Epoch 430/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0292\n",
      "Epoch 431/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0302\n",
      "Epoch 432/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.8032e-04 - mae: 0.0234\n",
      "Epoch 433/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.2297e-04 - mae: 0.0220\n",
      "Epoch 434/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.3160e-04 - mae: 0.0231\n",
      "Epoch 435/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 8.1529e-04 - mae: 0.0219\n",
      "Epoch 436/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0012 - mae: 0.0262\n",
      "Epoch 437/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.8330e-04 - mae: 0.0216\n",
      "Epoch 438/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.7976e-04 - mae: 0.0240\n",
      "Epoch 439/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.1380e-04 - mae: 0.0211\n",
      "Epoch 440/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0252\n",
      "Epoch 441/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0248\n",
      "Epoch 442/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0015 - mae: 0.0292\n",
      "Epoch 443/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0246\n",
      "Epoch 444/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.0497e-04 - mae: 0.0221\n",
      "Epoch 445/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0016 - mae: 0.0289\n",
      "Epoch 446/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0241\n",
      "Epoch 447/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.0643e-04 - mae: 0.0232\n",
      "Epoch 448/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.7295e-04 - mae: 0.0237\n",
      "Epoch 449/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.5647e-04 - mae: 0.0195\n",
      "Epoch 450/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0239\n",
      "Epoch 451/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0255\n",
      "Epoch 452/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0251\n",
      "Epoch 453/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0254\n",
      "Epoch 454/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.1715e-04 - mae: 0.0224\n",
      "Epoch 455/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0258\n",
      "Epoch 456/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.4669e-04 - mae: 0.0233\n",
      "Epoch 457/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.1039e-04 - mae: 0.0216\n",
      "Epoch 458/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.3652e-04 - mae: 0.0200\n",
      "Epoch 459/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0228\n",
      "Epoch 460/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.6002e-04 - mae: 0.0226\n",
      "Epoch 461/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.2398e-04 - mae: 0.0204\n",
      "Epoch 462/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.1175e-04 - mae: 0.0205\n",
      "Epoch 463/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.4139e-04 - mae: 0.0178\n",
      "Epoch 464/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.8668e-04 - mae: 0.0197\n",
      "Epoch 465/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.5894e-04 - mae: 0.0223\n",
      "Epoch 466/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.8945e-04 - mae: 0.0243\n",
      "Epoch 467/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.7169e-04 - mae: 0.0223\n",
      "Epoch 468/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0261\n",
      "Epoch 469/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.1661e-04 - mae: 0.0216\n",
      "Epoch 470/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0244\n",
      "Epoch 471/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0260\n",
      "Epoch 472/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.0582e-04 - mae: 0.0221\n",
      "Epoch 473/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.7500e-04 - mae: 0.0198\n",
      "Epoch 474/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.0223e-04 - mae: 0.0236\n",
      "Epoch 475/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.6729e-04 - mae: 0.0213\n",
      "Epoch 476/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.6563e-04 - mae: 0.0195\n",
      "Epoch 477/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.4027e-04 - mae: 0.0194\n",
      "Epoch 478/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.5126e-04 - mae: 0.0201\n",
      "Epoch 479/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.2618e-04 - mae: 0.0234\n",
      "Epoch 480/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0269\n",
      "Epoch 481/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.0964e-04 - mae: 0.0202\n",
      "Epoch 482/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 8.6339e-04 - mae: 0.0230\n",
      "Epoch 483/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.9920e-04 - mae: 0.0217\n",
      "Epoch 484/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.3726e-04 - mae: 0.0225\n",
      "Epoch 485/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.1661e-04 - mae: 0.0223\n",
      "Epoch 486/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.0626e-04 - mae: 0.0222\n",
      "Epoch 487/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.7185e-04 - mae: 0.0206\n",
      "Epoch 488/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.6020e-04 - mae: 0.0213\n",
      "Epoch 489/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.2521e-04 - mae: 0.0209\n",
      "Epoch 490/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.6547e-04 - mae: 0.0230\n",
      "Epoch 491/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 9.3621e-04 - mae: 0.0236\n",
      "Epoch 492/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 8.0161e-04 - mae: 0.0215\n",
      "Epoch 493/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 8.7626e-04 - mae: 0.0217\n",
      "Epoch 494/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.3640e-04 - mae: 0.0206\n",
      "Epoch 495/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.7702e-04 - mae: 0.0216\n",
      "Epoch 496/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.0280e-04 - mae: 0.0198\n",
      "Epoch 497/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.9842e-04 - mae: 0.0181\n",
      "Epoch 498/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.2001e-04 - mae: 0.0218\n",
      "Epoch 499/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.3302e-04 - mae: 0.0206\n",
      "Epoch 500/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.5929e-04 - mae: 0.0194\n",
      "Epoch 501/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.6818e-04 - mae: 0.0196\n",
      "Epoch 502/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.5264e-04 - mae: 0.0172\n",
      "Epoch 503/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.6148e-04 - mae: 0.0236\n",
      "Epoch 504/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0271\n",
      "Epoch 505/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0274\n",
      "Epoch 506/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0017 - mae: 0.0302\n",
      "Epoch 507/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0292\n",
      "Epoch 508/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0012 - mae: 0.0269\n",
      "Epoch 509/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.4199e-04 - mae: 0.0219\n",
      "Epoch 510/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.7690e-04 - mae: 0.0196\n",
      "Epoch 511/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.4350e-04 - mae: 0.0191\n",
      "Epoch 512/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0271\n",
      "Epoch 513/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0258\n",
      "Epoch 514/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0020 - mae: 0.0339\n",
      "Epoch 515/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0015 - mae: 0.0297\n",
      "Epoch 516/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0017 - mae: 0.0317\n",
      "Epoch 517/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0269\n",
      "Epoch 518/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0012 - mae: 0.0266\n",
      "Epoch 519/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0253\n",
      "Epoch 520/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0268\n",
      "Epoch 521/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0254\n",
      "Epoch 522/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.4413e-04 - mae: 0.0212\n",
      "Epoch 523/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0271\n",
      "Epoch 524/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0273\n",
      "Epoch 525/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.5640e-04 - mae: 0.0234\n",
      "Epoch 526/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0237\n",
      "Epoch 527/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.3477e-04 - mae: 0.0209\n",
      "Epoch 528/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.6379e-04 - mae: 0.0215\n",
      "Epoch 529/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.9961e-04 - mae: 0.0191\n",
      "Epoch 530/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.8932e-04 - mae: 0.0209\n",
      "Epoch 531/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.6284e-04 - mae: 0.0227\n",
      "Epoch 532/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.0616e-04 - mae: 0.0209\n",
      "Epoch 533/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.8950e-04 - mae: 0.0186\n",
      "Epoch 534/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.9650e-04 - mae: 0.0213\n",
      "Epoch 535/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0243\n",
      "Epoch 536/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.1754e-04 - mae: 0.0233\n",
      "Epoch 537/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.1918e-04 - mae: 0.0218\n",
      "Epoch 538/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.5869e-04 - mae: 0.0228\n",
      "Epoch 539/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.8133e-04 - mae: 0.0225\n",
      "Epoch 540/1000\n",
      "28/28 [==============================] - 0s 7ms/step - loss: 7.9952e-04 - mae: 0.0216\n",
      "Epoch 541/1000\n",
      "28/28 [==============================] - 0s 7ms/step - loss: 7.5884e-04 - mae: 0.0215\n",
      "Epoch 542/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.4446e-04 - mae: 0.0231\n",
      "Epoch 543/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.8858e-04 - mae: 0.0234\n",
      "Epoch 544/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.6253e-04 - mae: 0.0219\n",
      "Epoch 545/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.9109e-04 - mae: 0.0208\n",
      "Epoch 546/1000\n",
      "28/28 [==============================] - 0s 7ms/step - loss: 8.5834e-04 - mae: 0.0220\n",
      "Epoch 547/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0011 - mae: 0.0241\n",
      "Epoch 548/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0263\n",
      "Epoch 549/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0289\n",
      "Epoch 550/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0249\n",
      "Epoch 551/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.6277e-04 - mae: 0.0226\n",
      "Epoch 552/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.8369e-04 - mae: 0.0212\n",
      "Epoch 553/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.6555e-04 - mae: 0.0210\n",
      "Epoch 554/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.5811e-04 - mae: 0.0197\n",
      "Epoch 555/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0010 - mae: 0.0230\n",
      "Epoch 556/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 8.3682e-04 - mae: 0.0218\n",
      "Epoch 557/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.1883e-04 - mae: 0.0190\n",
      "Epoch 558/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.9712e-04 - mae: 0.0200\n",
      "Epoch 559/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0040 - mae: 0.0470\n",
      "Epoch 560/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0028 - mae: 0.0415\n",
      "Epoch 561/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0013 - mae: 0.0277\n",
      "Epoch 562/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0011 - mae: 0.0260\n",
      "Epoch 563/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0275\n",
      "Epoch 564/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0011 - mae: 0.0247\n",
      "Epoch 565/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0011 - mae: 0.0243\n",
      "Epoch 566/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.6169e-04 - mae: 0.0237\n",
      "Epoch 567/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0265\n",
      "Epoch 568/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.8208e-04 - mae: 0.0225\n",
      "Epoch 569/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 9.5898e-04 - mae: 0.0231\n",
      "Epoch 570/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0011 - mae: 0.0248\n",
      "Epoch 571/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0265\n",
      "Epoch 572/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0239\n",
      "Epoch 573/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.9613e-04 - mae: 0.0184\n",
      "Epoch 574/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0245\n",
      "Epoch 575/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.0637e-04 - mae: 0.0216\n",
      "Epoch 576/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 8.7864e-04 - mae: 0.0224\n",
      "Epoch 577/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 8.2560e-04 - mae: 0.0213\n",
      "Epoch 578/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.0366e-04 - mae: 0.0200\n",
      "Epoch 579/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.8179e-04 - mae: 0.0202\n",
      "Epoch 580/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0011 - mae: 0.0238\n",
      "Epoch 581/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.4547e-04 - mae: 0.0194\n",
      "Epoch 582/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.3794e-04 - mae: 0.0190\n",
      "Epoch 583/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.7511e-04 - mae: 0.0228\n",
      "Epoch 584/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.9548e-04 - mae: 0.0232\n",
      "Epoch 585/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.5780e-04 - mae: 0.0219\n",
      "Epoch 586/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.0931e-04 - mae: 0.0227\n",
      "Epoch 587/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.8198e-04 - mae: 0.0215\n",
      "Epoch 588/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.6449e-04 - mae: 0.0205\n",
      "Epoch 589/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.7611e-04 - mae: 0.0217\n",
      "Epoch 590/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.5688e-04 - mae: 0.0179\n",
      "Epoch 591/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.5249e-04 - mae: 0.0190\n",
      "Epoch 592/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.8925e-04 - mae: 0.0185\n",
      "Epoch 593/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.1387e-04 - mae: 0.0171\n",
      "Epoch 594/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.6563e-04 - mae: 0.0185\n",
      "Epoch 595/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.4936e-04 - mae: 0.0195\n",
      "Epoch 596/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.5337e-04 - mae: 0.0193\n",
      "Epoch 597/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.8568e-04 - mae: 0.0179\n",
      "Epoch 598/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.0867e-04 - mae: 0.0207\n",
      "Epoch 599/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.8948e-04 - mae: 0.0189\n",
      "Epoch 600/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.8931e-04 - mae: 0.0192\n",
      "Epoch 601/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.4939e-04 - mae: 0.0212\n",
      "Epoch 602/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.8218e-04 - mae: 0.0234\n",
      "Epoch 603/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.3569e-04 - mae: 0.0226\n",
      "Epoch 604/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.0275e-04 - mae: 0.0203\n",
      "Epoch 605/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.1916e-04 - mae: 0.0218\n",
      "Epoch 606/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.3891e-04 - mae: 0.0192\n",
      "Epoch 607/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.1898e-04 - mae: 0.0213\n",
      "Epoch 608/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.2272e-04 - mae: 0.0208\n",
      "Epoch 609/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.9969e-04 - mae: 0.0205\n",
      "Epoch 610/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.4507e-04 - mae: 0.0207\n",
      "Epoch 611/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.1496e-04 - mae: 0.0188\n",
      "Epoch 612/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0228\n",
      "Epoch 613/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.8357e-04 - mae: 0.0213\n",
      "Epoch 614/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.4536e-04 - mae: 0.0236\n",
      "Epoch 615/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0011 - mae: 0.0236\n",
      "Epoch 616/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.7608e-04 - mae: 0.0209\n",
      "Epoch 617/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0248\n",
      "Epoch 618/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.8804e-04 - mae: 0.0213\n",
      "Epoch 619/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.5576e-04 - mae: 0.0208\n",
      "Epoch 620/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.9224e-04 - mae: 0.0211\n",
      "Epoch 621/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0016 - mae: 0.0286\n",
      "Epoch 622/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0019 - mae: 0.0333\n",
      "Epoch 623/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0015 - mae: 0.0288\n",
      "Epoch 624/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0245\n",
      "Epoch 625/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.6386e-04 - mae: 0.0237\n",
      "Epoch 626/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.2090e-04 - mae: 0.0202\n",
      "Epoch 627/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.9977e-04 - mae: 0.0205\n",
      "Epoch 628/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.8975e-04 - mae: 0.0241\n",
      "Epoch 629/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.4678e-04 - mae: 0.0199\n",
      "Epoch 630/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.2969e-04 - mae: 0.0180\n",
      "Epoch 631/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.1370e-04 - mae: 0.0169\n",
      "Epoch 632/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.0555e-04 - mae: 0.0201\n",
      "Epoch 633/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.5080e-04 - mae: 0.0208\n",
      "Epoch 634/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.4351e-04 - mae: 0.0203\n",
      "Epoch 635/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.5258e-04 - mae: 0.0189\n",
      "Epoch 636/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.2082e-04 - mae: 0.0193\n",
      "Epoch 637/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.0888e-04 - mae: 0.0202\n",
      "Epoch 638/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.2044e-04 - mae: 0.0203\n",
      "Epoch 639/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.3243e-04 - mae: 0.0204\n",
      "Epoch 640/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.7170e-04 - mae: 0.0163\n",
      "Epoch 641/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.1797e-04 - mae: 0.0171\n",
      "Epoch 642/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.9496e-04 - mae: 0.0220\n",
      "Epoch 643/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.6725e-04 - mae: 0.0199\n",
      "Epoch 644/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.3077e-04 - mae: 0.0208\n",
      "Epoch 645/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0246\n",
      "Epoch 646/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.7939e-04 - mae: 0.0193\n",
      "Epoch 647/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 8.8242e-04 - mae: 0.0223\n",
      "Epoch 648/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.6086e-04 - mae: 0.0229\n",
      "Epoch 649/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.6895e-04 - mae: 0.0181\n",
      "Epoch 650/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.8800e-04 - mae: 0.0185\n",
      "Epoch 651/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0239\n",
      "Epoch 652/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.1371e-04 - mae: 0.0208\n",
      "Epoch 653/1000\n",
      "28/28 [==============================] - 0s 7ms/step - loss: 0.0011 - mae: 0.0243\n",
      "Epoch 654/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.9734e-04 - mae: 0.0205\n",
      "Epoch 655/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.1502e-04 - mae: 0.0221\n",
      "Epoch 656/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.2831e-04 - mae: 0.0207\n",
      "Epoch 657/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.4616e-04 - mae: 0.0209\n",
      "Epoch 658/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.4678e-04 - mae: 0.0185\n",
      "Epoch 659/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.2139e-04 - mae: 0.0183\n",
      "Epoch 660/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.4988e-04 - mae: 0.0183\n",
      "Epoch 661/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.6622e-04 - mae: 0.0205\n",
      "Epoch 662/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.2108e-04 - mae: 0.0198\n",
      "Epoch 663/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.9785e-04 - mae: 0.0197\n",
      "Epoch 664/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.8814e-04 - mae: 0.0164\n",
      "Epoch 665/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0014 - mae: 0.0209\n",
      "Epoch 666/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0045 - mae: 0.0438\n",
      "Epoch 667/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0028 - mae: 0.0418\n",
      "Epoch 668/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0018 - mae: 0.0329\n",
      "Epoch 669/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0015 - mae: 0.0293\n",
      "Epoch 670/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.9611e-04 - mae: 0.0247\n",
      "Epoch 671/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0256\n",
      "Epoch 672/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0012 - mae: 0.0266\n",
      "Epoch 673/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.8931e-04 - mae: 0.0234\n",
      "Epoch 674/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.8354e-04 - mae: 0.0205\n",
      "Epoch 675/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0231\n",
      "Epoch 676/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.1208e-04 - mae: 0.0186\n",
      "Epoch 677/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.1637e-04 - mae: 0.0205\n",
      "Epoch 678/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.1683e-04 - mae: 0.0203\n",
      "Epoch 679/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.8985e-04 - mae: 0.0199\n",
      "Epoch 680/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.5450e-04 - mae: 0.0199\n",
      "Epoch 681/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.0637e-04 - mae: 0.0209\n",
      "Epoch 682/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.4534e-04 - mae: 0.0187\n",
      "Epoch 683/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.7886e-04 - mae: 0.0198\n",
      "Epoch 684/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.9466e-04 - mae: 0.0200\n",
      "Epoch 685/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.9330e-04 - mae: 0.0196\n",
      "Epoch 686/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.7150e-04 - mae: 0.0223\n",
      "Epoch 687/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0011 - mae: 0.0257\n",
      "Epoch 688/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.6889e-04 - mae: 0.0221\n",
      "Epoch 689/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.3423e-04 - mae: 0.0231\n",
      "Epoch 690/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0245\n",
      "Epoch 691/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.5414e-04 - mae: 0.0219\n",
      "Epoch 692/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0209\n",
      "Epoch 693/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.6957e-04 - mae: 0.0191\n",
      "Epoch 694/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.0352e-04 - mae: 0.0182\n",
      "Epoch 695/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.6693e-04 - mae: 0.0195\n",
      "Epoch 696/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 9.6842e-04 - mae: 0.0226\n",
      "Epoch 697/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 8.6449e-04 - mae: 0.0216\n",
      "Epoch 698/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.9696e-04 - mae: 0.0198\n",
      "Epoch 699/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.5822e-04 - mae: 0.0190\n",
      "Epoch 700/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.0532e-04 - mae: 0.0200\n",
      "Epoch 701/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.1964e-04 - mae: 0.0188\n",
      "Epoch 702/1000\n",
      "28/28 [==============================] - 0s 7ms/step - loss: 5.5150e-04 - mae: 0.0174\n",
      "Epoch 703/1000\n",
      "28/28 [==============================] - 0s 7ms/step - loss: 5.9909e-04 - mae: 0.0184\n",
      "Epoch 704/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.5910e-04 - mae: 0.0206\n",
      "Epoch 705/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.4177e-04 - mae: 0.0221\n",
      "Epoch 706/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.0971e-04 - mae: 0.0207\n",
      "Epoch 707/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.3897e-04 - mae: 0.0175\n",
      "Epoch 708/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.3836e-04 - mae: 0.0219\n",
      "Epoch 709/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.8405e-04 - mae: 0.0182\n",
      "Epoch 710/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.7822e-04 - mae: 0.0188\n",
      "Epoch 711/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.2809e-04 - mae: 0.0194\n",
      "Epoch 712/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.7173e-04 - mae: 0.0180\n",
      "Epoch 713/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.0485e-04 - mae: 0.0180\n",
      "Epoch 714/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 9.2368e-04 - mae: 0.0204\n",
      "Epoch 715/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.5930e-04 - mae: 0.0172\n",
      "Epoch 716/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 8.1893e-04 - mae: 0.0211\n",
      "Epoch 717/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.6238e-04 - mae: 0.0187\n",
      "Epoch 718/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.2370e-04 - mae: 0.0168\n",
      "Epoch 719/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.1668e-04 - mae: 0.0173\n",
      "Epoch 720/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.6654e-04 - mae: 0.0166\n",
      "Epoch 721/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.6598e-04 - mae: 0.0174\n",
      "Epoch 722/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.4705e-04 - mae: 0.0190\n",
      "Epoch 723/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.8066e-04 - mae: 0.0174\n",
      "Epoch 724/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.8293e-04 - mae: 0.0184\n",
      "Epoch 725/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.5669e-04 - mae: 0.0187\n",
      "Epoch 726/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.3471e-04 - mae: 0.0209\n",
      "Epoch 727/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.5134e-04 - mae: 0.0209\n",
      "Epoch 728/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.9231e-04 - mae: 0.0209\n",
      "Epoch 729/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.8001e-04 - mae: 0.0199\n",
      "Epoch 730/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.1523e-04 - mae: 0.0198\n",
      "Epoch 731/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.2930e-04 - mae: 0.0173\n",
      "Epoch 732/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.8705e-04 - mae: 0.0181\n",
      "Epoch 733/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 4.9296e-04 - mae: 0.0167\n",
      "Epoch 734/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.9766e-04 - mae: 0.0187\n",
      "Epoch 735/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.7010e-04 - mae: 0.0177\n",
      "Epoch 736/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 4.6326e-04 - mae: 0.0152\n",
      "Epoch 737/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.1432e-04 - mae: 0.0171\n",
      "Epoch 738/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.7312e-04 - mae: 0.0163\n",
      "Epoch 739/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 4.0907e-04 - mae: 0.0148\n",
      "Epoch 740/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.2117e-04 - mae: 0.0165\n",
      "Epoch 741/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.5930e-04 - mae: 0.0178\n",
      "Epoch 742/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.5997e-04 - mae: 0.0183\n",
      "Epoch 743/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.8389e-04 - mae: 0.0203\n",
      "Epoch 744/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.3580e-04 - mae: 0.0173\n",
      "Epoch 745/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.0570e-04 - mae: 0.0163\n",
      "Epoch 746/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.9722e-04 - mae: 0.0175\n",
      "Epoch 747/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 4.8251e-04 - mae: 0.0162\n",
      "Epoch 748/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.2882e-04 - mae: 0.0187\n",
      "Epoch 749/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.6875e-04 - mae: 0.0179\n",
      "Epoch 750/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.9637e-04 - mae: 0.0177\n",
      "Epoch 751/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.8725e-04 - mae: 0.0194\n",
      "Epoch 752/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0229\n",
      "Epoch 753/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 8.3701e-04 - mae: 0.0222\n",
      "Epoch 754/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.0286e-04 - mae: 0.0187\n",
      "Epoch 755/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.7350e-04 - mae: 0.0194\n",
      "Epoch 756/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.9674e-04 - mae: 0.0194\n",
      "Epoch 757/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.5511e-04 - mae: 0.0177\n",
      "Epoch 758/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.8237e-04 - mae: 0.0159\n",
      "Epoch 759/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.5905e-04 - mae: 0.0172\n",
      "Epoch 760/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.5304e-04 - mae: 0.0215\n",
      "Epoch 761/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.5541e-04 - mae: 0.0205\n",
      "Epoch 762/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.2746e-04 - mae: 0.0188\n",
      "Epoch 763/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.5392e-04 - mae: 0.0186\n",
      "Epoch 764/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.2790e-04 - mae: 0.0182\n",
      "Epoch 765/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.0258e-04 - mae: 0.0181\n",
      "Epoch 766/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.3504e-04 - mae: 0.0161\n",
      "Epoch 767/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.3439e-04 - mae: 0.0163\n",
      "Epoch 768/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.9206e-04 - mae: 0.0159\n",
      "Epoch 769/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 4.8872e-04 - mae: 0.0164\n",
      "Epoch 770/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 4.6324e-04 - mae: 0.0156\n",
      "Epoch 771/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0012 - mae: 0.0254\n",
      "Epoch 772/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0011 - mae: 0.0248\n",
      "Epoch 773/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0011 - mae: 0.0246\n",
      "Epoch 774/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.5195e-04 - mae: 0.0213\n",
      "Epoch 775/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.0357e-04 - mae: 0.0205\n",
      "Epoch 776/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.9545e-04 - mae: 0.0188\n",
      "Epoch 777/1000\n",
      "28/28 [==============================] - 0s 7ms/step - loss: 7.0391e-04 - mae: 0.0192\n",
      "Epoch 778/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 8.0352e-04 - mae: 0.0200\n",
      "Epoch 779/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.5197e-04 - mae: 0.0181\n",
      "Epoch 780/1000\n",
      "28/28 [==============================] - 0s 7ms/step - loss: 6.2158e-04 - mae: 0.0180\n",
      "Epoch 781/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.2446e-04 - mae: 0.0217\n",
      "Epoch 782/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.2528e-04 - mae: 0.0182\n",
      "Epoch 783/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.5737e-04 - mae: 0.0176\n",
      "Epoch 784/1000\n",
      "28/28 [==============================] - 0s 7ms/step - loss: 5.4872e-04 - mae: 0.0174\n",
      "Epoch 785/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.6503e-04 - mae: 0.0164\n",
      "Epoch 786/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.9254e-04 - mae: 0.0182\n",
      "Epoch 787/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.7097e-04 - mae: 0.0181\n",
      "Epoch 788/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.1557e-04 - mae: 0.0192\n",
      "Epoch 789/1000\n",
      "28/28 [==============================] - 0s 7ms/step - loss: 7.5041e-04 - mae: 0.0201\n",
      "Epoch 790/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.6455e-04 - mae: 0.0177\n",
      "Epoch 791/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.6615e-04 - mae: 0.0173\n",
      "Epoch 792/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.4886e-04 - mae: 0.0155\n",
      "Epoch 793/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.7701e-04 - mae: 0.0171\n",
      "Epoch 794/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.0363e-04 - mae: 0.0169\n",
      "Epoch 795/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 8.4437e-04 - mae: 0.0201\n",
      "Epoch 796/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.3592e-04 - mae: 0.0186\n",
      "Epoch 797/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.8073e-04 - mae: 0.0188\n",
      "Epoch 798/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.6796e-04 - mae: 0.0184\n",
      "Epoch 799/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.6098e-04 - mae: 0.0194\n",
      "Epoch 800/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0017 - mae: 0.0235\n",
      "Epoch 801/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 8.7351e-04 - mae: 0.0229\n",
      "Epoch 802/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.3088e-04 - mae: 0.0192\n",
      "Epoch 803/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.4084e-04 - mae: 0.0188\n",
      "Epoch 804/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 8.4008e-04 - mae: 0.0213\n",
      "Epoch 805/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.6794e-04 - mae: 0.0208\n",
      "Epoch 806/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.9461e-04 - mae: 0.0197\n",
      "Epoch 807/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.6168e-04 - mae: 0.0244\n",
      "Epoch 808/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.1228e-04 - mae: 0.0189\n",
      "Epoch 809/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.9045e-04 - mae: 0.0177\n",
      "Epoch 810/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.6992e-04 - mae: 0.0178\n",
      "Epoch 811/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.0108e-04 - mae: 0.0203\n",
      "Epoch 812/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.0247e-04 - mae: 0.0191\n",
      "Epoch 813/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.9559e-04 - mae: 0.0186\n",
      "Epoch 814/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.6775e-04 - mae: 0.0167\n",
      "Epoch 815/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.1255e-04 - mae: 0.0184\n",
      "Epoch 816/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 4.2320e-04 - mae: 0.0151\n",
      "Epoch 817/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.1308e-04 - mae: 0.0194\n",
      "Epoch 818/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.5909e-04 - mae: 0.0158\n",
      "Epoch 819/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.3809e-04 - mae: 0.0153\n",
      "Epoch 820/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 4.6866e-04 - mae: 0.0158\n",
      "Epoch 821/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.8333e-04 - mae: 0.0163\n",
      "Epoch 822/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.0858e-04 - mae: 0.0179\n",
      "Epoch 823/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.8499e-04 - mae: 0.0179\n",
      "Epoch 824/1000\n",
      "28/28 [==============================] - 0s 7ms/step - loss: 5.3270e-04 - mae: 0.0174\n",
      "Epoch 825/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 4.1354e-04 - mae: 0.0145\n",
      "Epoch 826/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.0766e-04 - mae: 0.0145\n",
      "Epoch 827/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.8421e-04 - mae: 0.0162\n",
      "Epoch 828/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.3985e-04 - mae: 0.0170\n",
      "Epoch 829/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.0291e-04 - mae: 0.0168\n",
      "Epoch 830/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.8251e-04 - mae: 0.0160\n",
      "Epoch 831/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.5281e-04 - mae: 0.0162\n",
      "Epoch 832/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 4.2188e-04 - mae: 0.0148\n",
      "Epoch 833/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 3.5326e-04 - mae: 0.0141\n",
      "Epoch 834/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 3.9725e-04 - mae: 0.0150\n",
      "Epoch 835/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.8191e-04 - mae: 0.0161\n",
      "Epoch 836/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.0852e-04 - mae: 0.0172\n",
      "Epoch 837/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.4077e-04 - mae: 0.0184\n",
      "Epoch 838/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.2991e-04 - mae: 0.0150\n",
      "Epoch 839/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.2742e-04 - mae: 0.0189\n",
      "Epoch 840/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0018 - mae: 0.0332\n",
      "Epoch 841/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0014 - mae: 0.0285\n",
      "Epoch 842/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.8231e-04 - mae: 0.0212\n",
      "Epoch 843/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.1917e-04 - mae: 0.0213\n",
      "Epoch 844/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.7042e-04 - mae: 0.0216\n",
      "Epoch 845/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.5952e-04 - mae: 0.0195\n",
      "Epoch 846/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.9307e-04 - mae: 0.0200\n",
      "Epoch 847/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.4483e-04 - mae: 0.0193\n",
      "Epoch 848/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.4931e-04 - mae: 0.0211\n",
      "Epoch 849/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.9613e-04 - mae: 0.0191\n",
      "Epoch 850/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.6043e-04 - mae: 0.0198\n",
      "Epoch 851/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0237\n",
      "Epoch 852/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.3939e-04 - mae: 0.0200\n",
      "Epoch 853/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.6007e-04 - mae: 0.0179\n",
      "Epoch 854/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.9833e-04 - mae: 0.0180\n",
      "Epoch 855/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.2008e-04 - mae: 0.0232\n",
      "Epoch 856/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.4747e-04 - mae: 0.0222\n",
      "Epoch 857/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.8680e-04 - mae: 0.0216\n",
      "Epoch 858/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.4493e-04 - mae: 0.0236\n",
      "Epoch 859/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.2390e-04 - mae: 0.0198\n",
      "Epoch 860/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.5670e-04 - mae: 0.0191\n",
      "Epoch 861/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 8.3094e-04 - mae: 0.0204\n",
      "Epoch 862/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.1244e-04 - mae: 0.0190\n",
      "Epoch 863/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.2028e-04 - mae: 0.0172\n",
      "Epoch 864/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.7278e-04 - mae: 0.0187\n",
      "Epoch 865/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.0119e-04 - mae: 0.0201\n",
      "Epoch 866/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.5547e-04 - mae: 0.0209\n",
      "Epoch 867/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.7230e-04 - mae: 0.0207\n",
      "Epoch 868/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 8.7090e-04 - mae: 0.0219\n",
      "Epoch 869/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.9665e-04 - mae: 0.0203\n",
      "Epoch 870/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.2275e-04 - mae: 0.0176\n",
      "Epoch 871/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 4.8316e-04 - mae: 0.0161\n",
      "Epoch 872/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.5471e-04 - mae: 0.0183\n",
      "Epoch 873/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.3075e-04 - mae: 0.0173\n",
      "Epoch 874/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 3.9652e-04 - mae: 0.0151\n",
      "Epoch 875/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.9028e-04 - mae: 0.0161\n",
      "Epoch 876/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.2562e-04 - mae: 0.0174\n",
      "Epoch 877/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.1542e-04 - mae: 0.0170\n",
      "Epoch 878/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.0594e-04 - mae: 0.0218\n",
      "Epoch 879/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.2521e-04 - mae: 0.0197\n",
      "Epoch 880/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.9120e-04 - mae: 0.0205\n",
      "Epoch 881/1000\n",
      "28/28 [==============================] - 0s 7ms/step - loss: 7.4812e-04 - mae: 0.0203\n",
      "Epoch 882/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.4896e-04 - mae: 0.0172\n",
      "Epoch 883/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.4383e-04 - mae: 0.0185\n",
      "Epoch 884/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.8236e-04 - mae: 0.0182\n",
      "Epoch 885/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.0282e-04 - mae: 0.0188\n",
      "Epoch 886/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.9205e-04 - mae: 0.0211\n",
      "Epoch 887/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.7001e-04 - mae: 0.0199\n",
      "Epoch 888/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0225\n",
      "Epoch 889/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0012 - mae: 0.0248\n",
      "Epoch 890/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.9311e-04 - mae: 0.0201\n",
      "Epoch 891/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.0936e-04 - mae: 0.0196\n",
      "Epoch 892/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.6064e-04 - mae: 0.0189\n",
      "Epoch 893/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.9882e-04 - mae: 0.0181\n",
      "Epoch 894/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.5196e-04 - mae: 0.0197\n",
      "Epoch 895/1000\n",
      "28/28 [==============================] - 0s 7ms/step - loss: 0.0014 - mae: 0.0267\n",
      "Epoch 896/1000\n",
      "28/28 [==============================] - 0s 7ms/step - loss: 0.0012 - mae: 0.0265\n",
      "Epoch 897/1000\n",
      "28/28 [==============================] - 0s 7ms/step - loss: 7.4331e-04 - mae: 0.0203\n",
      "Epoch 898/1000\n",
      "28/28 [==============================] - 0s 7ms/step - loss: 5.8632e-04 - mae: 0.0186\n",
      "Epoch 899/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.0234e-04 - mae: 0.0182\n",
      "Epoch 900/1000\n",
      "28/28 [==============================] - 0s 7ms/step - loss: 4.3112e-04 - mae: 0.0154\n",
      "Epoch 901/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.1565e-04 - mae: 0.0169\n",
      "Epoch 902/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.5799e-04 - mae: 0.0158\n",
      "Epoch 903/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.0907e-04 - mae: 0.0182\n",
      "Epoch 904/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.1398e-04 - mae: 0.0170\n",
      "Epoch 905/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.5730e-04 - mae: 0.0174\n",
      "Epoch 906/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.1672e-04 - mae: 0.0160\n",
      "Epoch 907/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.2727e-04 - mae: 0.0177\n",
      "Epoch 908/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.0412e-04 - mae: 0.0159\n",
      "Epoch 909/1000\n",
      "28/28 [==============================] - 0s 7ms/step - loss: 4.2363e-04 - mae: 0.0148\n",
      "Epoch 910/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 4.7616e-04 - mae: 0.0166\n",
      "Epoch 911/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.3728e-04 - mae: 0.0158\n",
      "Epoch 912/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.0668e-04 - mae: 0.0158\n",
      "Epoch 913/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 4.7787e-04 - mae: 0.0154\n",
      "Epoch 914/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.6313e-04 - mae: 0.0198\n",
      "Epoch 915/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.5673e-04 - mae: 0.0159\n",
      "Epoch 916/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0011 - mae: 0.0233\n",
      "Epoch 917/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.1608e-04 - mae: 0.0221\n",
      "Epoch 918/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.3569e-04 - mae: 0.0199\n",
      "Epoch 919/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.2233e-04 - mae: 0.0186\n",
      "Epoch 920/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.9132e-04 - mae: 0.0180\n",
      "Epoch 921/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0235\n",
      "Epoch 922/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0011 - mae: 0.0251\n",
      "Epoch 923/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 8.0588e-04 - mae: 0.0192\n",
      "Epoch 924/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.6515e-04 - mae: 0.0189\n",
      "Epoch 925/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.8552e-04 - mae: 0.0172\n",
      "Epoch 926/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.9800e-04 - mae: 0.0193\n",
      "Epoch 927/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.9966e-04 - mae: 0.0184\n",
      "Epoch 928/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.7798e-04 - mae: 0.0198\n",
      "Epoch 929/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0214\n",
      "Epoch 930/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.8224e-04 - mae: 0.0164\n",
      "Epoch 931/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.7018e-04 - mae: 0.0154\n",
      "Epoch 932/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.9815e-04 - mae: 0.0168\n",
      "Epoch 933/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.3296e-04 - mae: 0.0171\n",
      "Epoch 934/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 4.5498e-04 - mae: 0.0153\n",
      "Epoch 935/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 4.5308e-04 - mae: 0.0144\n",
      "Epoch 936/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.5467e-04 - mae: 0.0169\n",
      "Epoch 937/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.2909e-04 - mae: 0.0175\n",
      "Epoch 938/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.4387e-04 - mae: 0.0179\n",
      "Epoch 939/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.8217e-04 - mae: 0.0179\n",
      "Epoch 940/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.3365e-04 - mae: 0.0146\n",
      "Epoch 941/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.5641e-04 - mae: 0.0158\n",
      "Epoch 942/1000\n",
      "28/28 [==============================] - 0s 7ms/step - loss: 5.7390e-04 - mae: 0.0183\n",
      "Epoch 943/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.3426e-04 - mae: 0.0167\n",
      "Epoch 944/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.0838e-04 - mae: 0.0171\n",
      "Epoch 945/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 3.8889e-04 - mae: 0.0144\n",
      "Epoch 946/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.5343e-04 - mae: 0.0170\n",
      "Epoch 947/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.7257e-04 - mae: 0.0172\n",
      "Epoch 948/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.4941e-04 - mae: 0.0162\n",
      "Epoch 949/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.0133e-04 - mae: 0.0146\n",
      "Epoch 950/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 4.3975e-04 - mae: 0.0147\n",
      "Epoch 951/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.5938e-04 - mae: 0.0160\n",
      "Epoch 952/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 4.3913e-04 - mae: 0.0145\n",
      "Epoch 953/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.1575e-04 - mae: 0.0160\n",
      "Epoch 954/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.7578e-04 - mae: 0.0162\n",
      "Epoch 955/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 3.7456e-04 - mae: 0.0146\n",
      "Epoch 956/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.4033e-04 - mae: 0.0146\n",
      "Epoch 957/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.0145e-04 - mae: 0.0163\n",
      "Epoch 958/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0023 - mae: 0.0298\n",
      "Epoch 959/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0270\n",
      "Epoch 960/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0029 - mae: 0.0382\n",
      "Epoch 961/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0026 - mae: 0.0388\n",
      "Epoch 962/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0013 - mae: 0.0289\n",
      "Epoch 963/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0014 - mae: 0.0282\n",
      "Epoch 964/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0015 - mae: 0.0295\n",
      "Epoch 965/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 0.0010 - mae: 0.0254\n",
      "Epoch 966/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 8.5520e-04 - mae: 0.0225\n",
      "Epoch 967/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 9.6614e-04 - mae: 0.0243\n",
      "Epoch 968/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.1260e-04 - mae: 0.0216\n",
      "Epoch 969/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.2018e-04 - mae: 0.0204\n",
      "Epoch 970/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 7.4037e-04 - mae: 0.0206\n",
      "Epoch 971/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 9.0916e-04 - mae: 0.0219\n",
      "Epoch 972/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.2067e-04 - mae: 0.0202\n",
      "Epoch 973/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.6158e-04 - mae: 0.0202\n",
      "Epoch 974/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.4634e-04 - mae: 0.0183\n",
      "Epoch 975/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.0641e-04 - mae: 0.0191\n",
      "Epoch 976/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.8980e-04 - mae: 0.0177\n",
      "Epoch 977/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.4505e-04 - mae: 0.0204\n",
      "Epoch 978/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.4416e-04 - mae: 0.0175\n",
      "Epoch 979/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.3583e-04 - mae: 0.0172\n",
      "Epoch 980/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 0.0012 - mae: 0.0216\n",
      "Epoch 981/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.4770e-04 - mae: 0.0198\n",
      "Epoch 982/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.4774e-04 - mae: 0.0184\n",
      "Epoch 983/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 8.0095e-04 - mae: 0.0212\n",
      "Epoch 984/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 8.3002e-04 - mae: 0.0213\n",
      "Epoch 985/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.1936e-04 - mae: 0.0193\n",
      "Epoch 986/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.2826e-04 - mae: 0.0191\n",
      "Epoch 987/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.5251e-04 - mae: 0.0175\n",
      "Epoch 988/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.5952e-04 - mae: 0.0169\n",
      "Epoch 989/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.8653e-04 - mae: 0.0184\n",
      "Epoch 990/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 6.8099e-04 - mae: 0.0194\n",
      "Epoch 991/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.9924e-04 - mae: 0.0186\n",
      "Epoch 992/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.4913e-04 - mae: 0.0180\n",
      "Epoch 993/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 5.2001e-04 - mae: 0.0163\n",
      "Epoch 994/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 7.4548e-04 - mae: 0.0188\n",
      "Epoch 995/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 4.9971e-04 - mae: 0.0166\n",
      "Epoch 996/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 5.0425e-04 - mae: 0.0172\n",
      "Epoch 997/1000\n",
      "28/28 [==============================] - 0s 5ms/step - loss: 6.8579e-04 - mae: 0.0181\n",
      "Epoch 998/1000\n",
      "28/28 [==============================] - 0s 7ms/step - loss: 4.5485e-04 - mae: 0.0155\n",
      "Epoch 999/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 4.3175e-04 - mae: 0.0152\n",
      "Epoch 1000/1000\n",
      "28/28 [==============================] - 0s 6ms/step - loss: 4.5100e-04 - mae: 0.0157\n"
     ]
    }
   ],
   "source": [
    "#%%\n",
    "\n",
    "model = createModel()\n",
    "history = model.fit(x_train, y_train.T[0], epochs=1000, batch_size=25)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "id": "jEu2zDT2FJoZ",
    "outputId": "12364cd8-258a-4248-de07-471ee1a4a634"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:5 out of the last 13 calls to <function Model.make_test_function.<locals>.test_function at 0x7f54c2a461e0> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
      "3/3 [==============================] - 0s 3ms/step - loss: 0.0018 - mae: 0.0325\n",
      "[0.0017872523749247193, 0.03252248093485832]\n",
      "WARNING:tensorflow:5 out of the last 13 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7f54c313b840> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
      " ----Predict----      -------Real-------    -----Images Tags - Paths----- \n",
      "0 : [0.61262256]              0.6             681 - test7_205.bmp\n",
      "\n",
      "1 : [0.11936939]              0.1             764 - test7_288.bmp\n",
      "\n",
      "2 : [0.87347037]              0.85             214 - 10X80_63.bmp\n",
      "\n",
      "3 : [0.2379711]              0.25             14 - 10X70_14.bmp\n",
      "\n",
      "4 : [0.69830555]              0.7             187 - 10X80_36.bmp\n",
      "\n",
      "5 : [0.11160514]              0.1             761 - test7_285.bmp\n",
      "\n",
      "6 : [0.6892598]              0.7             79 - 10X70_79.bmp\n",
      "\n",
      "7 : [0.586354]              0.6             680 - test7_204.bmp\n",
      "\n",
      "8 : [0.8329718]              0.9             643 - test7_167.bmp\n",
      "\n",
      "9 : [0.52501816]              0.5             531 - test7_55.bmp\n",
      "\n",
      "10 : [0.33242023]              0.35             31 - 10X70_31.bmp\n",
      "\n",
      "11 : [0.8961382]              0.9             235 - 10X80_84.bmp\n",
      "\n",
      "12 : [0.91910183]              0.9             392 - 20X80_51.bmp\n",
      "\n",
      "13 : [0.85558033]              0.85             583 - test7_107.bmp\n",
      "\n",
      "14 : [0.14860913]              0.15             739 - test7_263.bmp\n",
      "\n",
      "15 : [0.81825924]              0.8             204 - 10X80_53.bmp\n",
      "\n",
      "16 : [0.7179094]              0.75             85 - 10X70_85.bmp\n",
      "\n",
      "17 : [0.1243242]              0.1             755 - test7_279.bmp\n",
      "\n",
      "18 : [0.7387301]              0.75             662 - test7_186.bmp\n",
      "\n",
      "19 : [0.20380592]              0.25             350 - 20X80_9.bmp\n",
      "\n",
      "20 : [0.8553375]              0.9             239 - 10X80_88.bmp\n",
      "\n",
      "21 : [0.95927185]              0.9             113 - 10X70_113.bmp\n",
      "\n",
      "22 : [0.11185563]              0.1             751 - test7_275.bmp\n",
      "\n",
      "23 : [0.8124556]              0.8             316 - 20X70_68.bmp\n",
      "\n",
      "24 : [0.37754416]              0.35             27 - 10X70_27.bmp\n",
      "\n",
      "25 : [0.8952559]              0.95             618 - test7_142.bmp\n",
      "\n",
      "26 : [0.6472562]              0.65             75 - 10X70_75.bmp\n",
      "\n",
      "27 : [0.7181574]              0.75             413 - 20X80_72.bmp\n",
      "\n",
      "28 : [0.9994408]              0.95             605 - test7_129.bmp\n",
      "\n",
      "29 : [0.24509782]              0.25             725 - test7_249.bmp\n",
      "\n",
      "30 : [0.13893542]              0.2             454 - 20X80_113.bmp\n",
      "\n",
      "31 : [0.38740057]              0.35             356 - 20X80_15.bmp\n",
      "\n",
      "32 : [0.7232551]              0.75             569 - test7_93.bmp\n",
      "\n",
      "33 : [0.5993525]              0.55             367 - 20X80_26.bmp\n",
      "\n",
      "34 : [0.61125153]              0.6             682 - test7_206.bmp\n",
      "\n",
      "35 : [0.8601288]              0.85             406 - 20X80_65.bmp\n",
      "\n",
      "36 : [0.5543722]              0.55             162 - 10X80_11.bmp\n",
      "\n",
      "37 : [0.7745959]              0.8             575 - test7_99.bmp\n",
      "\n",
      "38 : [0.61568457]              0.65             175 - 10X80_24.bmp\n",
      "\n",
      "39 : [0.4336904]              0.45             40 - 10X70_40.bmp\n",
      "\n",
      "40 : [0.08280453]              0.1             762 - test7_286.bmp\n",
      "\n",
      "41 : [0.8443843]              0.8             145 - 10X70_145.bmp\n",
      "\n",
      "42 : [0.8322641]              0.8             142 - 10X70_142.bmp\n",
      "\n",
      "43 : [0.8178951]              0.8             578 - test7_102.bmp\n",
      "\n",
      "44 : [0.1782813]              0.25             8 - 10X70_8.bmp\n",
      "\n",
      "45 : [0.698311]              0.7             666 - test7_190.bmp\n",
      "\n",
      "46 : [0.81395006]              0.8             577 - test7_101.bmp\n",
      "\n",
      "47 : [0.898363]              0.95             397 - 20X80_56.bmp\n",
      "\n",
      "48 : [0.20303974]              0.3             263 - 20X70_15.bmp\n",
      "\n",
      "49 : [0.94356143]              0.95             628 - test7_152.bmp\n",
      "\n",
      "50 : [0.73894125]              0.75             570 - test7_94.bmp\n",
      "\n",
      "51 : [0.0882071]              0.1             770 - test7_294.bmp\n",
      "\n",
      "52 : [0.5217184]              0.6             424 - 20X80_83.bmp\n",
      "\n",
      "53 : [0.8390467]              0.8             408 - 20X80_67.bmp\n",
      "\n",
      "54 : [0.5152116]              0.55             55 - 10X70_55.bmp\n",
      "\n",
      "55 : [0.09162337]              0.1             482 - test7_6.bmp\n",
      "\n",
      "56 : [0.32682315]              0.45             434 - 20X80_93.bmp\n",
      "\n",
      "57 : [0.1698137]              0.2             1 - 10X70_1.bmp\n",
      "\n",
      "58 : [0.8736111]              0.9             389 - 20X80_48.bmp\n",
      "\n",
      "59 : [0.23681661]              0.25             453 - 20X80_112.bmp\n",
      "\n",
      "60 : [0.902632]              0.95             231 - 10X80_80.bmp\n",
      "\n",
      "61 : [0.719374]              0.75             665 - test7_189.bmp\n",
      "\n",
      "62 : [0.90992135]              0.85             215 - 10X80_64.bmp\n",
      "\n",
      "63 : [0.29572225]              0.3             261 - 20X70_13.bmp\n",
      "\n",
      "64 : [0.4404135]              0.4             268 - 20X70_20.bmp\n",
      "\n",
      "65 : [0.92066467]              0.85             241 - 10X80_90.bmp\n",
      "\n",
      "66 : [0.19065017]              0.3             501 - test7_25.bmp\n",
      "\n",
      "67 : [0.7550626]              0.85             382 - 20X80_41.bmp\n",
      "\n",
      "68 : [0.2545246]              0.3             354 - 20X80_13.bmp\n",
      "\n",
      "69 : [0.74132705]              0.75             412 - 20X80_71.bmp\n",
      "\n",
      "70 : [0.4086233]              0.35             266 - 20X70_18.bmp\n",
      "\n",
      "71 : [1.]              0.95             627 - test7_151.bmp\n",
      "\n",
      "72 : [0.15148348]              0.15             249 - 20X70_1.bmp\n",
      "\n",
      "73 : [0.9436964]              0.95             622 - test7_146.bmp\n",
      "\n",
      "74 : [0.5214487]              0.45             271 - 20X70_23.bmp\n",
      "\n",
      "75 : [0.47157037]              0.55             366 - 20X80_25.bmp\n",
      "\n",
      "76 : [0.8211596]              0.85             103 - 10X70_103.bmp\n",
      "\n",
      "77 : [0.9968215]              0.95             608 - test7_132.bmp\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#%%\n",
    "\n",
    "showAccuracy(history)\n",
    "print(model.evaluate(x_test, y_test.T[0]))\n",
    "\n",
    "#%% y_test_predict print\n",
    "\n",
    "y_test_predict = model.predict(x_test)\n",
    "\n",
    "print(\" ----Predict----      -------Real-------    -----Images Tags - Paths----- \")\n",
    "for i in range(len(y_test_predict)):\n",
    "      print( str(i) + \" : \"+ str(y_test_predict[i]) + \"              \" + str(y_test.T[0][i]) + \"             \" + str(int(y_test.T[1][i])) + \" - \" + imagePaths[int(y_test.T[1][i])])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "qCVL1PK4FQj7"
   },
   "outputs": [],
   "source": [
    "#%% model save .json\n",
    "\n",
    "model_json = model.to_json()\n",
    "\n",
    "with open(\"drive/My Drive/Colab Notebooks/FFT_Variance_ANN_Tasarim_Donemi/model_100_25_FFT.json\", \"w\") as json_file:\n",
    "    json_file.write(model_json)\n",
    "\n",
    "# serialize weights to HDF5\n",
    "model.save_weights(\"drive/My Drive/Colab Notebooks/FFT_Variance_ANN_Tasarim_Donemi/model_100_25_FFT.h5\")\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "CmCaVzvDFTEh"
   },
   "outputs": [],
   "source": [
    "#%% model load\n",
    "\n",
    "json_file = open('drive/My Drive/Colab Notebooks/FFT_Variance_ANN_Tasarim_Donemi/model_100_25_FFT.json', 'r')\n",
    "\n",
    "loaded_model_json = json_file.read()\n",
    "json_file.close()\n",
    "loaded_model = model_from_json(loaded_model_json)\n",
    "\n",
    "loaded_model.load_weights(\"drive/My Drive/Colab Notebooks/FFT_Variance_ANN_Tasarim_Donemi/model_100_25_FFT.h5\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "xYxYq8YkFVQb",
    "outputId": "5529cad1-db1b-491d-fd09-9685a7adff04"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " ----Predict----      -------Real-------    -----Images Tags - Paths----- \n",
      "0 : [0.61262256]      0.6         681 - test7_205.bmp\n",
      "\n",
      "1 : [0.11936939]      0.1         764 - test7_288.bmp\n",
      "\n",
      "2 : [0.8734704]      0.85         214 - 10X80_63.bmp\n",
      "\n",
      "3 : [0.23797113]      0.25         14 - 10X70_14.bmp\n",
      "\n",
      "4 : [0.69830555]      0.7         187 - 10X80_36.bmp\n",
      "\n",
      "5 : [0.11160514]      0.1         761 - test7_285.bmp\n",
      "\n",
      "6 : [0.6892598]      0.7         79 - 10X70_79.bmp\n",
      "\n",
      "7 : [0.586354]      0.6         680 - test7_204.bmp\n",
      "\n",
      "8 : [0.83297193]      0.9         643 - test7_167.bmp\n",
      "\n",
      "9 : [0.52501816]      0.5         531 - test7_55.bmp\n",
      "\n",
      "10 : [0.33242023]      0.35         31 - 10X70_31.bmp\n",
      "\n",
      "11 : [0.8961381]      0.9         235 - 10X80_84.bmp\n",
      "\n",
      "12 : [0.9191019]      0.9         392 - 20X80_51.bmp\n",
      "\n",
      "13 : [0.85558033]      0.85         583 - test7_107.bmp\n",
      "\n",
      "14 : [0.14860907]      0.15         739 - test7_263.bmp\n",
      "\n",
      "15 : [0.81825924]      0.8         204 - 10X80_53.bmp\n",
      "\n",
      "16 : [0.7179094]      0.75         85 - 10X70_85.bmp\n",
      "\n",
      "17 : [0.12432414]      0.1         755 - test7_279.bmp\n",
      "\n",
      "18 : [0.7387301]      0.75         662 - test7_186.bmp\n",
      "\n",
      "19 : [0.20380592]      0.25         350 - 20X80_9.bmp\n",
      "\n",
      "20 : [0.8553374]      0.9         239 - 10X80_88.bmp\n",
      "\n",
      "21 : [0.95927185]      0.9         113 - 10X70_113.bmp\n",
      "\n",
      "22 : [0.11185554]      0.1         751 - test7_275.bmp\n",
      "\n",
      "23 : [0.81245565]      0.8         316 - 20X70_68.bmp\n",
      "\n",
      "24 : [0.37754405]      0.35         27 - 10X70_27.bmp\n",
      "\n",
      "25 : [0.895256]      0.95         618 - test7_142.bmp\n",
      "\n",
      "26 : [0.64725626]      0.65         75 - 10X70_75.bmp\n",
      "\n",
      "27 : [0.7181574]      0.75         413 - 20X80_72.bmp\n",
      "\n",
      "28 : [0.9994409]      0.95         605 - test7_129.bmp\n",
      "\n",
      "29 : [0.24509782]      0.25         725 - test7_249.bmp\n",
      "\n",
      "30 : [0.13893533]      0.2         454 - 20X80_113.bmp\n",
      "\n",
      "31 : [0.38740057]      0.35         356 - 20X80_15.bmp\n",
      "\n",
      "32 : [0.72325504]      0.75         569 - test7_93.bmp\n",
      "\n",
      "33 : [0.5993525]      0.55         367 - 20X80_26.bmp\n",
      "\n",
      "34 : [0.6112516]      0.6         682 - test7_206.bmp\n",
      "\n",
      "35 : [0.8601289]      0.85         406 - 20X80_65.bmp\n",
      "\n",
      "36 : [0.5543722]      0.55         162 - 10X80_11.bmp\n",
      "\n",
      "37 : [0.7745961]      0.8         575 - test7_99.bmp\n",
      "\n",
      "38 : [0.6156846]      0.65         175 - 10X80_24.bmp\n",
      "\n",
      "39 : [0.43369037]      0.45         40 - 10X70_40.bmp\n",
      "\n",
      "40 : [0.08280453]      0.1         762 - test7_286.bmp\n",
      "\n",
      "41 : [0.8443843]      0.8         145 - 10X70_145.bmp\n",
      "\n",
      "42 : [0.8322642]      0.8         142 - 10X70_142.bmp\n",
      "\n",
      "43 : [0.8178952]      0.8         578 - test7_102.bmp\n",
      "\n",
      "44 : [0.17828128]      0.25         8 - 10X70_8.bmp\n",
      "\n",
      "45 : [0.698311]      0.7         666 - test7_190.bmp\n",
      "\n",
      "46 : [0.8139502]      0.8         577 - test7_101.bmp\n",
      "\n",
      "47 : [0.898363]      0.95         397 - 20X80_56.bmp\n",
      "\n",
      "48 : [0.20303965]      0.3         263 - 20X70_15.bmp\n",
      "\n",
      "49 : [0.9435616]      0.95         628 - test7_152.bmp\n",
      "\n",
      "50 : [0.73894125]      0.75         570 - test7_94.bmp\n",
      "\n",
      "51 : [0.0882071]      0.1         770 - test7_294.bmp\n",
      "\n",
      "52 : [0.5217184]      0.6         424 - 20X80_83.bmp\n",
      "\n",
      "53 : [0.8390467]      0.8         408 - 20X80_67.bmp\n",
      "\n",
      "54 : [0.51521164]      0.55         55 - 10X70_55.bmp\n",
      "\n",
      "55 : [0.09162328]      0.1         482 - test7_6.bmp\n",
      "\n",
      "56 : [0.326823]      0.45         434 - 20X80_93.bmp\n",
      "\n",
      "57 : [0.1698136]      0.2         1 - 10X70_1.bmp\n",
      "\n",
      "58 : [0.8736113]      0.9         389 - 20X80_48.bmp\n",
      "\n",
      "59 : [0.23681661]      0.25         453 - 20X80_112.bmp\n",
      "\n",
      "60 : [0.902632]      0.95         231 - 10X80_80.bmp\n",
      "\n",
      "61 : [0.719374]      0.75         665 - test7_189.bmp\n",
      "\n",
      "62 : [0.9099214]      0.85         215 - 10X80_64.bmp\n",
      "\n",
      "63 : [0.29572225]      0.3         261 - 20X70_13.bmp\n",
      "\n",
      "64 : [0.44041353]      0.4         268 - 20X70_20.bmp\n",
      "\n",
      "65 : [0.92066455]      0.85         241 - 10X80_90.bmp\n",
      "\n",
      "66 : [0.1906502]      0.3         501 - test7_25.bmp\n",
      "\n",
      "67 : [0.7550626]      0.85         382 - 20X80_41.bmp\n",
      "\n",
      "68 : [0.25452468]      0.3         354 - 20X80_13.bmp\n",
      "\n",
      "69 : [0.74132705]      0.75         412 - 20X80_71.bmp\n",
      "\n",
      "70 : [0.40862328]      0.35         266 - 20X70_18.bmp\n",
      "\n",
      "71 : [1.]      0.95         627 - test7_151.bmp\n",
      "\n",
      "72 : [0.15148339]      0.15         249 - 20X70_1.bmp\n",
      "\n",
      "73 : [0.9436964]      0.95         622 - test7_146.bmp\n",
      "\n",
      "74 : [0.5214487]      0.45         271 - 20X70_23.bmp\n",
      "\n",
      "75 : [0.47157034]      0.55         366 - 20X80_25.bmp\n",
      "\n",
      "76 : [0.8211596]      0.85         103 - 10X70_103.bmp\n",
      "\n",
      "77 : [0.9968215]      0.95         608 - test7_132.bmp\n",
      "\n",
      "Min value:0.08280453 index: [40]\n",
      "Max value:1.0 index: [71]\n"
     ]
    }
   ],
   "source": [
    "\n",
    "#%% loaded_model predict\n",
    "\n",
    "loaded_model_predict = loaded_model.predict(x_test)\n",
    "\n",
    "print(\" ----Predict----      -------Real-------    -----Images Tags - Paths----- \")\n",
    "for i in range(len(loaded_model_predict)):\n",
    "     print( str(i) + \" : \"+ str(loaded_model_predict[i]) + \"      \" + str(y_test.T[0][i]) + \"         \" + str(int(y_test.T[1][i])) + \" - \" + imagePaths[int(y_test.T[1][i])])\n",
    "\n",
    "\n",
    "#%% print min and max values\n",
    "\n",
    "print(\"Min value:\" + str(np.amin(loaded_model_predict)) + \" index: \" + str(np.argmin(loaded_model_predict, axis=0)))\n",
    "print(\"Max value:\" + str(np.amax(loaded_model_predict)) + \" index: \" + str(np.argmax(loaded_model_predict, axis=0)))"
   ]
  }
 ],
 "metadata": {
  "accelerator": "TPU",
  "colab": {
   "authorship_tag": "ABX9TyNYNx0lUIvX73O5+BUZoDhe",
   "collapsed_sections": [],
   "include_colab_link": true,
   "name": "FFT_Variance_ANN_Tasarim.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": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}