material/deep_learning/anomaly_detection_fahsion_mnist.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 537,
      "metadata": {
        "id": "UCih9D443dmd"
      },
      "outputs": [],
      "source": [
        "import pandas as pd\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 데이터 준비"
      ],
      "metadata": {
        "id": "iU5AJt91PAGR"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import tensorflow as tf\n",
        "\n",
        "(train_x, train_y), (test_x, test_y) = tf.keras.datasets.fashion_mnist.load_data()"
      ],
      "metadata": {
        "id": "SlADaXCidbrr"
      },
      "execution_count": 538,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(train_x.shape)\n",
        "print(test_x.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LbsEx-6tdlH8",
        "outputId": "dba0bb4d-fba7-41d0-bc4a-2fd952b6e0de"
      },
      "execution_count": 539,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(60000, 28, 28)\n",
            "(10000, 28, 28)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.figure(figsize=(10,10))\n",
        "for i in range(25):\n",
        "    plt.subplot(5,5,i+1)\n",
        "    plt.imshow(test_x[i], cmap=\"binary\")\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 594
        },
        "id": "QP2HjoxadyZ5",
        "outputId": "e438c02b-3844-40d0-b8c5-d083ef129712"
      },
      "execution_count": 540,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x720 with 25 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Test 데이터 생성"
      ],
      "metadata": {
        "id": "mvmS1VlleRzX"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# load_data()로 받은 것은 read only이다 카피해서 write가능하게 한다.\n",
        "test_x = test_x.copy()\n",
        "test_y = np.zeros_like(test_y)"
      ],
      "metadata": {
        "id": "-zHSSLBvgFOY"
      },
      "execution_count": 541,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def draw_scar(image, size=2):\n",
        "    x = np.random.randint(0+size, 27-size)\n",
        "    y = np.random.randint(0+size, 27-size)\n",
        "    image[y-size:y+size+1,x] = 255\n",
        "    image[y,x-size:x+size+1] = 255\n",
        "\n",
        "# 반절을 낙서를 그린다.\n",
        "for i in range(int(len(test_x)/2)):    \n",
        "    draw_scar(test_x[i], size=1)\n",
        "    test_y[i] = 1"
      ],
      "metadata": {
        "id": "9VjZH9XxeTn4"
      },
      "execution_count": 542,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plt.figure(figsize=(10,10))\n",
        "for i in range(25):\n",
        "    plt.subplot(5,5,i+1)\n",
        "    plt.imshow(test_x[i], cmap=\"binary\")\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 594
        },
        "id": "iw3FWS3ufZ3F",
        "outputId": "bd73db03-9752-437b-b2b9-e54895de11f8"
      },
      "execution_count": 543,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x720 with 25 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Normalize"
      ],
      "metadata": {
        "id": "kW6KfHSEO68L"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "train_x = train_x / 255.\n",
        "test_x = test_x / 255."
      ],
      "metadata": {
        "id": "Posl_rHBhKD4"
      },
      "execution_count": 544,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_x = np.reshape(train_x, (-1, 28, 28, 1))\n",
        "test_x = np.reshape(test_x, (-1, 28, 28, 1))"
      ],
      "metadata": {
        "id": "yDhnfJNKtNJv"
      },
      "execution_count": 545,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 모델 학습"
      ],
      "metadata": {
        "id": "MwZaDQIzNPNs"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 546,
      "metadata": {
        "id": "Zo9OQ4A-5rog"
      },
      "outputs": [],
      "source": [
        "from tensorflow.keras.datasets import mnist\n",
        "from tensorflow.keras.layers import Dense\n",
        "from tensorflow.keras import Sequential"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 547,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bSOoocZb9tld",
        "outputId": "a1b995a4-b9b3-4b22-df77-ffd7cee693a0"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/1000\n",
            "53/53 [==============================] - 5s 48ms/step - loss: 0.4934 - mape: 99631792.0000 - val_loss: 0.3730 - val_mape: 18796304.0000\n",
            "Epoch 2/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.3499 - mape: 26425858.0000 - val_loss: 0.3344 - val_mape: 30996720.0000\n",
            "Epoch 3/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.3191 - mape: 19099880.0000 - val_loss: 0.3142 - val_mape: 9992366.0000\n",
            "Epoch 4/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.3062 - mape: 14960645.0000 - val_loss: 0.3347 - val_mape: 28473458.0000\n",
            "Epoch 5/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2971 - mape: 12772235.0000 - val_loss: 0.3021 - val_mape: 7622666.0000\n",
            "Epoch 6/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2910 - mape: 10852307.0000 - val_loss: 0.2876 - val_mape: 13329054.0000\n",
            "Epoch 7/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2862 - mape: 9688834.0000 - val_loss: 0.2905 - val_mape: 6542227.0000\n",
            "Epoch 8/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2829 - mape: 8622563.0000 - val_loss: 0.2809 - val_mape: 10798671.0000\n",
            "Epoch 9/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2800 - mape: 7885874.0000 - val_loss: 0.2777 - val_mape: 4905250.0000\n",
            "Epoch 10/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2776 - mape: 7136744.5000 - val_loss: 0.2756 - val_mape: 8597800.0000\n",
            "Epoch 11/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2754 - mape: 6602166.0000 - val_loss: 0.2826 - val_mape: 5240365.5000\n",
            "Epoch 12/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2736 - mape: 6123278.5000 - val_loss: 0.2768 - val_mape: 9221236.0000\n",
            "Epoch 13/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2730 - mape: 5884595.5000 - val_loss: 0.2718 - val_mape: 3933731.5000\n",
            "Epoch 14/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2713 - mape: 5550008.5000 - val_loss: 0.2715 - val_mape: 7104188.5000\n",
            "Epoch 15/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2710 - mape: 5365673.5000 - val_loss: 0.2712 - val_mape: 3950999.2500\n",
            "Epoch 16/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2694 - mape: 5118546.0000 - val_loss: 0.2715 - val_mape: 3680003.7500\n",
            "Epoch 17/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2690 - mape: 4881725.0000 - val_loss: 0.2694 - val_mape: 4072718.2500\n",
            "Epoch 18/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2680 - mape: 4739616.0000 - val_loss: 0.2685 - val_mape: 5091137.0000\n",
            "Epoch 19/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2677 - mape: 4752030.5000 - val_loss: 0.2704 - val_mape: 2478924.2500\n",
            "Epoch 20/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2669 - mape: 4507908.5000 - val_loss: 0.2729 - val_mape: 5048465.5000\n",
            "Epoch 21/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2668 - mape: 4370829.0000 - val_loss: 0.2691 - val_mape: 3942937.2500\n",
            "Epoch 22/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2659 - mape: 4271259.5000 - val_loss: 0.2683 - val_mape: 4732437.5000\n",
            "Epoch 23/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2656 - mape: 4226256.5000 - val_loss: 0.2687 - val_mape: 2847746.5000\n",
            "Epoch 24/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2646 - mape: 4079737.0000 - val_loss: 0.2664 - val_mape: 4338795.5000\n",
            "Epoch 25/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2655 - mape: 4009527.2500 - val_loss: 0.2659 - val_mape: 4702955.5000\n",
            "Epoch 26/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2640 - mape: 3848811.2500 - val_loss: 0.2647 - val_mape: 4133396.0000\n",
            "Epoch 27/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2645 - mape: 3899416.0000 - val_loss: 0.2651 - val_mape: 4451408.5000\n",
            "Epoch 28/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2639 - mape: 3824084.2500 - val_loss: 0.2651 - val_mape: 3105885.2500\n",
            "Epoch 29/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2636 - mape: 3764695.0000 - val_loss: 0.2682 - val_mape: 2879373.2500\n",
            "Epoch 30/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2631 - mape: 3638519.5000 - val_loss: 0.2690 - val_mape: 5293554.5000\n",
            "Epoch 31/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2630 - mape: 3626543.0000 - val_loss: 0.2640 - val_mape: 2199829.7500\n",
            "Epoch 32/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2627 - mape: 3494787.2500 - val_loss: 0.2627 - val_mape: 3722514.5000\n",
            "Epoch 33/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2625 - mape: 3537660.5000 - val_loss: 0.2631 - val_mape: 4439441.5000\n",
            "Epoch 34/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2623 - mape: 3447841.2500 - val_loss: 0.2630 - val_mape: 2368429.0000\n",
            "Epoch 35/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2622 - mape: 3431121.7500 - val_loss: 0.2640 - val_mape: 3722963.7500\n",
            "Epoch 36/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2619 - mape: 3394856.7500 - val_loss: 0.2627 - val_mape: 2735243.0000\n",
            "Epoch 37/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2612 - mape: 3263579.7500 - val_loss: 0.2661 - val_mape: 5260735.0000\n",
            "Epoch 38/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2609 - mape: 3217402.0000 - val_loss: 0.2730 - val_mape: 4824767.0000\n",
            "Epoch 39/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2619 - mape: 3357807.0000 - val_loss: 0.2649 - val_mape: 2383548.0000\n",
            "Epoch 40/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2610 - mape: 3188377.5000 - val_loss: 0.2628 - val_mape: 4442387.0000\n",
            "Epoch 41/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2610 - mape: 3167637.5000 - val_loss: 0.2615 - val_mape: 2803381.2500\n",
            "Epoch 42/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2607 - mape: 3150560.0000 - val_loss: 0.2620 - val_mape: 2196859.2500\n",
            "Epoch 43/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2605 - mape: 3056938.0000 - val_loss: 0.2617 - val_mape: 3423968.5000\n",
            "Epoch 44/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2602 - mape: 3081443.5000 - val_loss: 0.2616 - val_mape: 3457768.2500\n",
            "Epoch 45/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2603 - mape: 3018135.7500 - val_loss: 0.2615 - val_mape: 2114869.2500\n",
            "Epoch 46/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2597 - mape: 2928199.0000 - val_loss: 0.2602 - val_mape: 2504996.2500\n",
            "Epoch 47/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2604 - mape: 3016266.5000 - val_loss: 0.2615 - val_mape: 2127762.7500\n",
            "Epoch 48/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2597 - mape: 2953733.2500 - val_loss: 0.2630 - val_mape: 3187708.5000\n",
            "Epoch 49/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2598 - mape: 2932234.5000 - val_loss: 0.2632 - val_mape: 4327023.5000\n",
            "Epoch 50/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2594 - mape: 2884066.7500 - val_loss: 0.2616 - val_mape: 1915852.0000\n",
            "Epoch 51/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2592 - mape: 2848362.5000 - val_loss: 0.2646 - val_mape: 2019440.6250\n",
            "Epoch 52/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2597 - mape: 2863305.0000 - val_loss: 0.2615 - val_mape: 3755479.5000\n",
            "Epoch 53/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2591 - mape: 2836618.2500 - val_loss: 0.2627 - val_mape: 2876494.0000\n",
            "Epoch 54/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2591 - mape: 2795172.0000 - val_loss: 0.2623 - val_mape: 2166859.5000\n",
            "Epoch 55/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2592 - mape: 2800907.7500 - val_loss: 0.2607 - val_mape: 1913902.6250\n",
            "Epoch 56/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2587 - mape: 2719496.0000 - val_loss: 0.2610 - val_mape: 3283543.5000\n",
            "Epoch 57/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2584 - mape: 2693523.7500 - val_loss: 0.2624 - val_mape: 3734424.7500\n",
            "Epoch 58/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2589 - mape: 2733548.0000 - val_loss: 0.2600 - val_mape: 2064445.7500\n",
            "Epoch 59/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2586 - mape: 2702145.5000 - val_loss: 0.2597 - val_mape: 1922557.0000\n",
            "Epoch 60/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2584 - mape: 2669969.0000 - val_loss: 0.2615 - val_mape: 2910745.0000\n",
            "Epoch 61/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2583 - mape: 2636153.2500 - val_loss: 0.2595 - val_mape: 2145618.0000\n",
            "Epoch 62/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2582 - mape: 2625029.2500 - val_loss: 0.2598 - val_mape: 1955367.3750\n",
            "Epoch 63/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2582 - mape: 2630102.0000 - val_loss: 0.2601 - val_mape: 3106882.2500\n",
            "Epoch 64/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2581 - mape: 2594235.0000 - val_loss: 0.2604 - val_mape: 2297863.7500\n",
            "Epoch 65/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2580 - mape: 2576253.2500 - val_loss: 0.2594 - val_mape: 1802801.7500\n",
            "Epoch 66/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2575 - mape: 2512953.0000 - val_loss: 0.2595 - val_mape: 3138798.5000\n",
            "Epoch 67/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2582 - mape: 2561059.7500 - val_loss: 0.2602 - val_mape: 1880497.1250\n",
            "Epoch 68/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2576 - mape: 2504197.2500 - val_loss: 0.2593 - val_mape: 1766956.6250\n",
            "Epoch 69/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2579 - mape: 2539745.7500 - val_loss: 0.2598 - val_mape: 2753007.5000\n",
            "Epoch 70/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2575 - mape: 2483065.5000 - val_loss: 0.2592 - val_mape: 1995247.1250\n",
            "Epoch 71/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2575 - mape: 2473428.0000 - val_loss: 0.2592 - val_mape: 1797531.5000\n",
            "Epoch 72/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2575 - mape: 2481226.7500 - val_loss: 0.2592 - val_mape: 1938422.8750\n",
            "Epoch 73/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2574 - mape: 2432460.7500 - val_loss: 0.2598 - val_mape: 2887735.2500\n",
            "Epoch 74/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2570 - mape: 2406576.5000 - val_loss: 0.2589 - val_mape: 3165843.7500\n",
            "Epoch 75/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2573 - mape: 2439069.5000 - val_loss: 0.2593 - val_mape: 2058786.5000\n",
            "Epoch 76/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2570 - mape: 2392568.0000 - val_loss: 0.2591 - val_mape: 3278395.5000\n",
            "Epoch 77/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2571 - mape: 2420950.0000 - val_loss: 0.2591 - val_mape: 1874197.0000\n",
            "Epoch 78/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2569 - mape: 2355749.7500 - val_loss: 0.2609 - val_mape: 2973648.5000\n",
            "Epoch 79/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2570 - mape: 2351712.7500 - val_loss: 0.2582 - val_mape: 2331405.5000\n",
            "Epoch 80/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2569 - mape: 2379753.2500 - val_loss: 0.2584 - val_mape: 1917450.0000\n",
            "Epoch 81/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2567 - mape: 2324171.7500 - val_loss: 0.2591 - val_mape: 2925363.0000\n",
            "Epoch 82/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2563 - mape: 2279756.0000 - val_loss: 0.2578 - val_mape: 2633958.0000\n",
            "Epoch 83/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2571 - mape: 2350734.7500 - val_loss: 0.2579 - val_mape: 2008256.3750\n",
            "Epoch 84/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2567 - mape: 2316739.0000 - val_loss: 0.2582 - val_mape: 2621781.5000\n",
            "Epoch 85/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2565 - mape: 2274778.5000 - val_loss: 0.2587 - val_mape: 2728435.0000\n",
            "Epoch 86/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2564 - mape: 2287513.2500 - val_loss: 0.2582 - val_mape: 1634385.7500\n",
            "Epoch 87/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2565 - mape: 2273303.7500 - val_loss: 0.2587 - val_mape: 2892419.0000\n",
            "Epoch 88/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2564 - mape: 2248341.2500 - val_loss: 0.2589 - val_mape: 1840999.7500\n",
            "Epoch 89/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2563 - mape: 2242284.7500 - val_loss: 0.2581 - val_mape: 2831096.2500\n",
            "Epoch 90/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2563 - mape: 2237066.7500 - val_loss: 0.2588 - val_mape: 2744501.2500\n",
            "Epoch 91/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2562 - mape: 2206112.0000 - val_loss: 0.2581 - val_mape: 1628171.3750\n",
            "Epoch 92/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2559 - mape: 2184631.0000 - val_loss: 0.2574 - val_mape: 1716145.6250\n",
            "Epoch 93/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2564 - mape: 2205968.7500 - val_loss: 0.2582 - val_mape: 2239615.7500\n",
            "Epoch 94/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2560 - mape: 2190155.0000 - val_loss: 0.2580 - val_mape: 1611779.6250\n",
            "Epoch 95/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2557 - mape: 2135041.7500 - val_loss: 0.2574 - val_mape: 1623385.2500\n",
            "Epoch 96/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2562 - mape: 2195563.5000 - val_loss: 0.2581 - val_mape: 2323835.7500\n",
            "Epoch 97/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2559 - mape: 2179006.0000 - val_loss: 0.2572 - val_mape: 1627624.2500\n",
            "Epoch 98/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2557 - mape: 2117688.0000 - val_loss: 0.2574 - val_mape: 2002943.5000\n",
            "Epoch 99/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2561 - mape: 2169341.5000 - val_loss: 0.2574 - val_mape: 1701355.2500\n",
            "Epoch 100/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2555 - mape: 2103478.2500 - val_loss: 0.2579 - val_mape: 2309596.2500\n",
            "Epoch 101/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2558 - mape: 2111609.0000 - val_loss: 0.2575 - val_mape: 2122956.2500\n",
            "Epoch 102/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2558 - mape: 2130229.5000 - val_loss: 0.2575 - val_mape: 1657255.5000\n",
            "Epoch 103/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2554 - mape: 2039931.1250 - val_loss: 0.2599 - val_mape: 3288432.7500\n",
            "Epoch 104/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2557 - mape: 2122410.5000 - val_loss: 0.2570 - val_mape: 1743495.5000\n",
            "Epoch 105/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2556 - mape: 2096329.8750 - val_loss: 0.2574 - val_mape: 2610358.5000\n",
            "Epoch 106/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2553 - mape: 2054496.0000 - val_loss: 0.2576 - val_mape: 1472763.7500\n",
            "Epoch 107/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2555 - mape: 2058904.8750 - val_loss: 0.2574 - val_mape: 2393698.2500\n",
            "Epoch 108/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2554 - mape: 2058806.8750 - val_loss: 0.2578 - val_mape: 2691536.7500\n",
            "Epoch 109/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2553 - mape: 2021306.6250 - val_loss: 0.2569 - val_mape: 1695471.1250\n",
            "Epoch 110/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2553 - mape: 2046451.0000 - val_loss: 0.2573 - val_mape: 2504513.0000\n",
            "Epoch 111/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2552 - mape: 2013701.2500 - val_loss: 0.2569 - val_mape: 2414941.5000\n",
            "Epoch 112/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2554 - mape: 2036211.1250 - val_loss: 0.2569 - val_mape: 1653453.2500\n",
            "Epoch 113/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2551 - mape: 2003263.3750 - val_loss: 0.2577 - val_mape: 2554308.5000\n",
            "Epoch 114/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2551 - mape: 2021470.5000 - val_loss: 0.2578 - val_mape: 1849972.0000\n",
            "Epoch 115/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2551 - mape: 1970914.5000 - val_loss: 0.2568 - val_mape: 1779403.3750\n",
            "Epoch 116/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2550 - mape: 1976687.6250 - val_loss: 0.2580 - val_mape: 1478272.0000\n",
            "Epoch 117/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2550 - mape: 1980373.2500 - val_loss: 0.2572 - val_mape: 2525522.2500\n",
            "Epoch 118/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2550 - mape: 1959450.1250 - val_loss: 0.2576 - val_mape: 2529983.7500\n",
            "Epoch 119/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2549 - mape: 1956704.6250 - val_loss: 0.2565 - val_mape: 1595725.7500\n",
            "Epoch 120/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2549 - mape: 1942327.1250 - val_loss: 0.2562 - val_mape: 1961095.7500\n",
            "Epoch 121/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2550 - mape: 1981749.0000 - val_loss: 0.2567 - val_mape: 1624730.3750\n",
            "Epoch 122/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2545 - mape: 1902054.6250 - val_loss: 0.2583 - val_mape: 1481359.5000\n",
            "Epoch 123/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2550 - mape: 1946763.7500 - val_loss: 0.2567 - val_mape: 1925993.3750\n",
            "Epoch 124/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2543 - mape: 1857518.8750 - val_loss: 0.2586 - val_mape: 1335827.7500\n",
            "Epoch 125/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2550 - mape: 1955240.5000 - val_loss: 0.2562 - val_mape: 1579760.6250\n",
            "Epoch 126/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2548 - mape: 1939692.6250 - val_loss: 0.2562 - val_mape: 2188269.0000\n",
            "Epoch 127/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2546 - mape: 1898429.0000 - val_loss: 0.2566 - val_mape: 2188767.2500\n",
            "Epoch 128/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2544 - mape: 1889948.2500 - val_loss: 0.2568 - val_mape: 1531998.2500\n",
            "Epoch 129/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2546 - mape: 1855326.5000 - val_loss: 0.2564 - val_mape: 2036031.0000\n",
            "Epoch 130/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2546 - mape: 1891758.7500 - val_loss: 0.2560 - val_mape: 1891619.2500\n",
            "Epoch 131/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2547 - mape: 1901557.7500 - val_loss: 0.2561 - val_mape: 1502784.6250\n",
            "Epoch 132/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2544 - mape: 1865151.7500 - val_loss: 0.2562 - val_mape: 2240603.7500\n",
            "Epoch 133/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2545 - mape: 1862785.3750 - val_loss: 0.2563 - val_mape: 1990548.1250\n",
            "Epoch 134/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2543 - mape: 1832508.3750 - val_loss: 0.2560 - val_mape: 1823260.5000\n",
            "Epoch 135/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2545 - mape: 1871515.0000 - val_loss: 0.2559 - val_mape: 1498034.0000\n",
            "Epoch 136/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2542 - mape: 1821263.7500 - val_loss: 0.2571 - val_mape: 2512269.2500\n",
            "Epoch 137/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2541 - mape: 1820464.6250 - val_loss: 0.2559 - val_mape: 2128762.0000\n",
            "Epoch 138/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2543 - mape: 1846892.2500 - val_loss: 0.2565 - val_mape: 1462657.2500\n",
            "Epoch 139/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2544 - mape: 1822713.7500 - val_loss: 0.2565 - val_mape: 1404641.3750\n",
            "Epoch 140/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2541 - mape: 1801945.1250 - val_loss: 0.2564 - val_mape: 2225613.5000\n",
            "Epoch 141/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2540 - mape: 1786308.2500 - val_loss: 0.2564 - val_mape: 2212867.5000\n",
            "Epoch 142/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2543 - mape: 1829144.0000 - val_loss: 0.2565 - val_mape: 1583712.8750\n",
            "Epoch 143/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2543 - mape: 1809870.1250 - val_loss: 0.2570 - val_mape: 2230670.0000\n",
            "Epoch 144/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2539 - mape: 1771749.5000 - val_loss: 0.2554 - val_mape: 1790305.6250\n",
            "Epoch 145/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2541 - mape: 1810857.1250 - val_loss: 0.2557 - val_mape: 1476950.0000\n",
            "Epoch 146/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2540 - mape: 1776526.8750 - val_loss: 0.2562 - val_mape: 2109144.2500\n",
            "Epoch 147/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2538 - mape: 1746471.5000 - val_loss: 0.2557 - val_mape: 2029278.1250\n",
            "Epoch 148/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2541 - mape: 1786979.6250 - val_loss: 0.2560 - val_mape: 1444646.0000\n",
            "Epoch 149/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2540 - mape: 1753047.6250 - val_loss: 0.2557 - val_mape: 2065675.3750\n",
            "Epoch 150/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2539 - mape: 1768373.2500 - val_loss: 0.2558 - val_mape: 1378906.2500\n",
            "Epoch 151/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2539 - mape: 1750170.3750 - val_loss: 0.2559 - val_mape: 2252969.7500\n",
            "Epoch 152/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2540 - mape: 1761719.8750 - val_loss: 0.2561 - val_mape: 1342964.6250\n",
            "Epoch 153/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2537 - mape: 1713549.8750 - val_loss: 0.2552 - val_mape: 1661810.3750\n",
            "Epoch 154/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2540 - mape: 1771397.5000 - val_loss: 0.2555 - val_mape: 2052231.2500\n",
            "Epoch 155/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2536 - mape: 1710170.7500 - val_loss: 0.2555 - val_mape: 1412388.6250\n",
            "Epoch 156/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2538 - mape: 1738833.5000 - val_loss: 0.2558 - val_mape: 2054672.6250\n",
            "Epoch 157/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2538 - mape: 1726423.3750 - val_loss: 0.2556 - val_mape: 1636109.2500\n",
            "Epoch 158/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2537 - mape: 1718759.7500 - val_loss: 0.2561 - val_mape: 1513115.5000\n",
            "Epoch 159/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2536 - mape: 1706760.3750 - val_loss: 0.2564 - val_mape: 1278762.7500\n",
            "Epoch 160/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2537 - mape: 1702678.1250 - val_loss: 0.2561 - val_mape: 2376285.7500\n",
            "Epoch 161/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2537 - mape: 1717520.7500 - val_loss: 0.2555 - val_mape: 1363878.2500\n",
            "Epoch 162/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2535 - mape: 1699257.8750 - val_loss: 0.2551 - val_mape: 1723538.2500\n",
            "Epoch 163/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2537 - mape: 1702276.0000 - val_loss: 0.2555 - val_mape: 1832846.2500\n",
            "Epoch 164/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2535 - mape: 1687764.7500 - val_loss: 0.2563 - val_mape: 1424897.2500\n",
            "Epoch 165/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2535 - mape: 1674338.1250 - val_loss: 0.2559 - val_mape: 1692092.0000\n",
            "Epoch 166/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2534 - mape: 1656881.5000 - val_loss: 0.2566 - val_mape: 2477762.0000\n",
            "Epoch 167/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2536 - mape: 1713608.0000 - val_loss: 0.2554 - val_mape: 1349877.6250\n",
            "Epoch 168/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2535 - mape: 1672912.5000 - val_loss: 0.2556 - val_mape: 1793706.0000\n",
            "Epoch 169/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2535 - mape: 1665893.2500 - val_loss: 0.2566 - val_mape: 2117015.0000\n",
            "Epoch 170/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2532 - mape: 1655433.6250 - val_loss: 0.2561 - val_mape: 1348513.5000\n",
            "Epoch 171/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2536 - mape: 1667568.7500 - val_loss: 0.2557 - val_mape: 1773448.7500\n",
            "Epoch 172/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2535 - mape: 1665331.5000 - val_loss: 0.2557 - val_mape: 1315824.6250\n",
            "Epoch 173/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2532 - mape: 1633546.3750 - val_loss: 0.2553 - val_mape: 1901971.1250\n",
            "Epoch 174/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2533 - mape: 1662411.0000 - val_loss: 0.2551 - val_mape: 1396364.2500\n",
            "Epoch 175/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2534 - mape: 1641055.0000 - val_loss: 0.2563 - val_mape: 1278339.5000\n",
            "Epoch 176/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2533 - mape: 1637225.2500 - val_loss: 0.2554 - val_mape: 1920194.3750\n",
            "Epoch 177/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2532 - mape: 1634104.0000 - val_loss: 0.2558 - val_mape: 1477918.5000\n",
            "Epoch 178/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2532 - mape: 1635864.8750 - val_loss: 0.2554 - val_mape: 1731658.7500\n",
            "Epoch 179/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2531 - mape: 1608838.6250 - val_loss: 0.2552 - val_mape: 1925521.5000\n",
            "Epoch 180/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2532 - mape: 1638219.2500 - val_loss: 0.2552 - val_mape: 1461896.7500\n",
            "Epoch 181/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2533 - mape: 1617683.2500 - val_loss: 0.2553 - val_mape: 1731377.1250\n",
            "Epoch 182/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2531 - mape: 1611184.1250 - val_loss: 0.2549 - val_mape: 1550095.8750\n",
            "Epoch 183/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2533 - mape: 1623779.8750 - val_loss: 0.2558 - val_mape: 2108748.5000\n",
            "Epoch 184/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2530 - mape: 1600328.1250 - val_loss: 0.2558 - val_mape: 1958369.8750\n",
            "Epoch 185/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2531 - mape: 1605712.0000 - val_loss: 0.2550 - val_mape: 1625401.1250\n",
            "Epoch 186/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2532 - mape: 1625007.1250 - val_loss: 0.2549 - val_mape: 1760830.1250\n",
            "Epoch 187/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2531 - mape: 1595596.7500 - val_loss: 0.2553 - val_mape: 1455339.6250\n",
            "Epoch 188/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2531 - mape: 1592247.1250 - val_loss: 0.2558 - val_mape: 1295839.1250\n",
            "Epoch 189/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2529 - mape: 1593128.8750 - val_loss: 0.2550 - val_mape: 1681939.2500\n",
            "Epoch 190/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2530 - mape: 1588248.6250 - val_loss: 0.2547 - val_mape: 1510794.6250\n",
            "Epoch 191/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2532 - mape: 1590499.1250 - val_loss: 0.2550 - val_mape: 1798020.7500\n",
            "Epoch 192/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2529 - mape: 1576672.7500 - val_loss: 0.2550 - val_mape: 1236199.6250\n",
            "Epoch 193/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2530 - mape: 1572022.2500 - val_loss: 0.2552 - val_mape: 1982022.3750\n",
            "Epoch 194/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2529 - mape: 1567192.3750 - val_loss: 0.2556 - val_mape: 2041599.5000\n",
            "Epoch 195/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2529 - mape: 1583796.1250 - val_loss: 0.2551 - val_mape: 1459251.3750\n",
            "Epoch 196/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2530 - mape: 1577504.6250 - val_loss: 0.2549 - val_mape: 1887921.8750\n",
            "Epoch 197/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2528 - mape: 1546838.6250 - val_loss: 0.2556 - val_mape: 2128721.7500\n",
            "Epoch 198/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2529 - mape: 1579205.0000 - val_loss: 0.2550 - val_mape: 1205930.8750\n",
            "Epoch 199/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2529 - mape: 1559445.0000 - val_loss: 0.2549 - val_mape: 1803293.8750\n",
            "Epoch 200/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2527 - mape: 1534449.2500 - val_loss: 0.2546 - val_mape: 1726304.6250\n",
            "Epoch 201/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2529 - mape: 1579753.0000 - val_loss: 0.2545 - val_mape: 1306364.6250\n",
            "Epoch 202/1000\n",
            "53/53 [==============================] - 2s 43ms/step - loss: 0.2529 - mape: 1545283.1250 - val_loss: 0.2550 - val_mape: 2016765.1250\n",
            "Epoch 203/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2527 - mape: 1546968.1250 - val_loss: 0.2548 - val_mape: 1196139.2500\n",
            "Epoch 204/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2528 - mape: 1536410.8750 - val_loss: 0.2553 - val_mape: 2130093.0000\n",
            "Epoch 205/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2527 - mape: 1542675.0000 - val_loss: 0.2550 - val_mape: 1193253.1250\n",
            "Epoch 206/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2526 - mape: 1526210.7500 - val_loss: 0.2544 - val_mape: 1320763.0000\n",
            "Epoch 207/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2528 - mape: 1545889.6250 - val_loss: 0.2545 - val_mape: 1696322.6250\n",
            "Epoch 208/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2526 - mape: 1520931.5000 - val_loss: 0.2568 - val_mape: 1989086.0000\n",
            "Epoch 209/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2528 - mape: 1534828.0000 - val_loss: 0.2544 - val_mape: 1638973.3750\n",
            "Epoch 210/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2528 - mape: 1548623.6250 - val_loss: 0.2543 - val_mape: 1301248.2500\n",
            "Epoch 211/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2526 - mape: 1499133.1250 - val_loss: 0.2551 - val_mape: 2029870.3750\n",
            "Epoch 212/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2527 - mape: 1528061.7500 - val_loss: 0.2548 - val_mape: 1385723.1250\n",
            "Epoch 213/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2526 - mape: 1509891.5000 - val_loss: 0.2560 - val_mape: 1234433.2500\n",
            "Epoch 214/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2526 - mape: 1507409.0000 - val_loss: 0.2549 - val_mape: 1776958.0000\n",
            "Epoch 215/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2526 - mape: 1509831.1250 - val_loss: 0.2543 - val_mape: 1572756.1250\n",
            "Epoch 216/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2527 - mape: 1535007.2500 - val_loss: 0.2544 - val_mape: 1240407.5000\n",
            "Epoch 217/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2526 - mape: 1499728.8750 - val_loss: 0.2546 - val_mape: 1785127.2500\n",
            "Epoch 218/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2525 - mape: 1491330.6250 - val_loss: 0.2554 - val_mape: 1883183.3750\n",
            "Epoch 219/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2526 - mape: 1516345.1250 - val_loss: 0.2548 - val_mape: 1327875.3750\n",
            "Epoch 220/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2524 - mape: 1485956.3750 - val_loss: 0.2554 - val_mape: 1248532.7500\n",
            "Epoch 221/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2526 - mape: 1501500.5000 - val_loss: 0.2545 - val_mape: 1516622.7500\n",
            "Epoch 222/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2525 - mape: 1501069.3750 - val_loss: 0.2547 - val_mape: 1384836.7500\n",
            "Epoch 223/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2523 - mape: 1470619.2500 - val_loss: 0.2555 - val_mape: 1182380.0000\n",
            "Epoch 224/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2526 - mape: 1496084.8750 - val_loss: 0.2547 - val_mape: 1623147.8750\n",
            "Epoch 225/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2525 - mape: 1492072.5000 - val_loss: 0.2551 - val_mape: 1245692.0000\n",
            "Epoch 226/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2524 - mape: 1477993.2500 - val_loss: 0.2554 - val_mape: 1593195.3750\n",
            "Epoch 227/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2525 - mape: 1488055.3750 - val_loss: 0.2545 - val_mape: 1205815.3750\n",
            "Epoch 228/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2524 - mape: 1468499.2500 - val_loss: 0.2541 - val_mape: 1423051.1250\n",
            "Epoch 229/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2524 - mape: 1478357.6250 - val_loss: 0.2548 - val_mape: 1782556.2500\n",
            "Epoch 230/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2522 - mape: 1468920.6250 - val_loss: 0.2552 - val_mape: 1217674.1250\n",
            "Epoch 231/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2523 - mape: 1461443.6250 - val_loss: 0.2543 - val_mape: 1394140.3750\n",
            "Epoch 232/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2525 - mape: 1479270.1250 - val_loss: 0.2546 - val_mape: 1435104.7500\n",
            "Epoch 233/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2524 - mape: 1472393.6250 - val_loss: 0.2544 - val_mape: 1240999.1250\n",
            "Epoch 234/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2524 - mape: 1452639.5000 - val_loss: 0.2544 - val_mape: 1665929.1250\n",
            "Epoch 235/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2523 - mape: 1461697.5000 - val_loss: 0.2546 - val_mape: 1252536.1250\n",
            "Epoch 236/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2523 - mape: 1468219.1250 - val_loss: 0.2543 - val_mape: 1582031.2500\n",
            "Epoch 237/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2522 - mape: 1442516.3750 - val_loss: 0.2545 - val_mape: 1761379.0000\n",
            "Epoch 238/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2524 - mape: 1474474.8750 - val_loss: 0.2542 - val_mape: 1226538.5000\n",
            "Epoch 239/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2522 - mape: 1435988.7500 - val_loss: 0.2545 - val_mape: 1812201.2500\n",
            "Epoch 240/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2524 - mape: 1467482.0000 - val_loss: 0.2545 - val_mape: 1154404.8750\n",
            "Epoch 241/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2522 - mape: 1441305.6250 - val_loss: 0.2544 - val_mape: 1635677.0000\n",
            "Epoch 242/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2522 - mape: 1448469.1250 - val_loss: 0.2544 - val_mape: 1211015.0000\n",
            "Epoch 243/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2522 - mape: 1436504.8750 - val_loss: 0.2540 - val_mape: 1498857.8750\n",
            "Epoch 244/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2523 - mape: 1441846.3750 - val_loss: 0.2542 - val_mape: 1641516.6250\n",
            "Epoch 245/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2522 - mape: 1437988.8750 - val_loss: 0.2541 - val_mape: 1284411.2500\n",
            "Epoch 246/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2523 - mape: 1443140.1250 - val_loss: 0.2543 - val_mape: 1708458.2500\n",
            "Epoch 247/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2521 - mape: 1429148.2500 - val_loss: 0.2541 - val_mape: 1308818.1250\n",
            "Epoch 248/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2522 - mape: 1444554.6250 - val_loss: 0.2542 - val_mape: 1637794.6250\n",
            "Epoch 249/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2520 - mape: 1408821.3750 - val_loss: 0.2543 - val_mape: 1166865.0000\n",
            "Epoch 250/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2523 - mape: 1448280.5000 - val_loss: 0.2540 - val_mape: 1609116.3750\n",
            "Epoch 251/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2522 - mape: 1431645.0000 - val_loss: 0.2541 - val_mape: 1248618.3750\n",
            "Epoch 252/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2522 - mape: 1426452.7500 - val_loss: 0.2541 - val_mape: 1550026.2500\n",
            "Epoch 253/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2521 - mape: 1423407.0000 - val_loss: 0.2540 - val_mape: 1226731.7500\n",
            "Epoch 254/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2521 - mape: 1412941.2500 - val_loss: 0.2540 - val_mape: 1548647.5000\n",
            "Epoch 255/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2522 - mape: 1433308.2500 - val_loss: 0.2540 - val_mape: 1208927.3750\n",
            "Epoch 256/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2521 - mape: 1406403.1250 - val_loss: 0.2543 - val_mape: 1687628.0000\n",
            "Epoch 257/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2520 - mape: 1411127.7500 - val_loss: 0.2540 - val_mape: 1252718.7500\n",
            "Epoch 258/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2521 - mape: 1407394.6250 - val_loss: 0.2542 - val_mape: 1460911.6250\n",
            "Epoch 259/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2521 - mape: 1417792.7500 - val_loss: 0.2542 - val_mape: 1228465.8750\n",
            "Epoch 260/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2521 - mape: 1408405.3750 - val_loss: 0.2543 - val_mape: 1649586.6250\n",
            "Epoch 261/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2520 - mape: 1408531.2500 - val_loss: 0.2541 - val_mape: 1274096.8750\n",
            "Epoch 262/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2519 - mape: 1407316.3750 - val_loss: 0.2546 - val_mape: 1561130.1250\n",
            "Epoch 263/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2520 - mape: 1406134.8750 - val_loss: 0.2540 - val_mape: 1226142.3750\n",
            "Epoch 264/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2521 - mape: 1407879.3750 - val_loss: 0.2539 - val_mape: 1607935.2500\n",
            "Epoch 265/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2520 - mape: 1403713.0000 - val_loss: 0.2541 - val_mape: 1231243.0000\n",
            "Epoch 266/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2520 - mape: 1397297.6250 - val_loss: 0.2541 - val_mape: 1465111.5000\n",
            "Epoch 267/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2519 - mape: 1391449.2500 - val_loss: 0.2550 - val_mape: 1658383.6250\n",
            "Epoch 268/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2518 - mape: 1392919.3750 - val_loss: 0.2540 - val_mape: 1137032.0000\n",
            "Epoch 269/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2521 - mape: 1398184.0000 - val_loss: 0.2538 - val_mape: 1494994.1250\n",
            "Epoch 270/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2519 - mape: 1394290.8750 - val_loss: 0.2540 - val_mape: 1201708.7500\n",
            "Epoch 271/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2520 - mape: 1385198.1250 - val_loss: 0.2544 - val_mape: 1654652.7500\n",
            "Epoch 272/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2519 - mape: 1389997.1250 - val_loss: 0.2540 - val_mape: 1240140.2500\n",
            "Epoch 273/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2520 - mape: 1393138.5000 - val_loss: 0.2538 - val_mape: 1432310.6250\n",
            "Epoch 274/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2520 - mape: 1384608.1250 - val_loss: 0.2539 - val_mape: 1529012.0000\n",
            "Epoch 275/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2518 - mape: 1363464.0000 - val_loss: 0.2538 - val_mape: 1450406.2500\n",
            "Epoch 276/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2520 - mape: 1401043.3750 - val_loss: 0.2539 - val_mape: 1519745.1250\n",
            "Epoch 277/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2518 - mape: 1365296.2500 - val_loss: 0.2538 - val_mape: 1494469.0000\n",
            "Epoch 278/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2520 - mape: 1401112.3750 - val_loss: 0.2538 - val_mape: 1176652.7500\n",
            "Epoch 279/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2518 - mape: 1358290.3750 - val_loss: 0.2541 - val_mape: 1402995.3750\n",
            "Epoch 280/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2519 - mape: 1387411.7500 - val_loss: 0.2538 - val_mape: 1213912.1250\n",
            "Epoch 281/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2519 - mape: 1376192.0000 - val_loss: 0.2540 - val_mape: 1508569.2500\n",
            "Epoch 282/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2517 - mape: 1358469.8750 - val_loss: 0.2545 - val_mape: 1762199.0000\n",
            "Epoch 283/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2519 - mape: 1388904.8750 - val_loss: 0.2538 - val_mape: 1189168.0000\n",
            "Epoch 284/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2518 - mape: 1349573.6250 - val_loss: 0.2536 - val_mape: 1344482.6250\n",
            "Epoch 285/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2519 - mape: 1387562.1250 - val_loss: 0.2537 - val_mape: 1196642.6250\n",
            "Epoch 286/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2518 - mape: 1349747.7500 - val_loss: 0.2538 - val_mape: 1506903.2500\n",
            "Epoch 287/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2518 - mape: 1374715.1250 - val_loss: 0.2539 - val_mape: 1159141.1250\n",
            "Epoch 288/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2518 - mape: 1360296.7500 - val_loss: 0.2542 - val_mape: 1087628.3750\n",
            "Epoch 289/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2518 - mape: 1355251.7500 - val_loss: 0.2536 - val_mape: 1437900.1250\n",
            "Epoch 290/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2518 - mape: 1373521.8750 - val_loss: 0.2538 - val_mape: 1141428.6250\n",
            "Epoch 291/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2518 - mape: 1357250.0000 - val_loss: 0.2539 - val_mape: 1209578.3750\n",
            "Epoch 292/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2518 - mape: 1356461.1250 - val_loss: 0.2538 - val_mape: 1372303.6250\n",
            "Epoch 293/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2515 - mape: 1337572.3750 - val_loss: 0.2539 - val_mape: 1635730.0000\n",
            "Epoch 294/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2519 - mape: 1366505.6250 - val_loss: 0.2539 - val_mape: 1300726.3750\n",
            "Epoch 295/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2518 - mape: 1354496.7500 - val_loss: 0.2540 - val_mape: 1603420.5000\n",
            "Epoch 296/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2517 - mape: 1344911.3750 - val_loss: 0.2535 - val_mape: 1407268.6250\n",
            "Epoch 297/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2519 - mape: 1371261.6250 - val_loss: 0.2537 - val_mape: 1124671.0000\n",
            "Epoch 298/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2516 - mape: 1329302.7500 - val_loss: 0.2539 - val_mape: 1472068.1250\n",
            "Epoch 299/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2518 - mape: 1356446.5000 - val_loss: 0.2539 - val_mape: 1187746.8750\n",
            "Epoch 300/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2515 - mape: 1335610.8750 - val_loss: 0.2536 - val_mape: 1375016.1250\n",
            "Epoch 301/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2518 - mape: 1352477.3750 - val_loss: 0.2537 - val_mape: 1218357.0000\n",
            "Epoch 302/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2517 - mape: 1344755.2500 - val_loss: 0.2538 - val_mape: 1559419.5000\n",
            "Epoch 303/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2517 - mape: 1337598.8750 - val_loss: 0.2536 - val_mape: 1258672.5000\n",
            "Epoch 304/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2517 - mape: 1345497.8750 - val_loss: 0.2542 - val_mape: 1098282.6250\n",
            "Epoch 305/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2517 - mape: 1342357.6250 - val_loss: 0.2544 - val_mape: 1601946.6250\n",
            "Epoch 306/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2515 - mape: 1331868.8750 - val_loss: 0.2537 - val_mape: 1162295.0000\n",
            "Epoch 307/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2517 - mape: 1327941.1250 - val_loss: 0.2543 - val_mape: 1891389.8750\n",
            "Epoch 308/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2517 - mape: 1348441.2500 - val_loss: 0.2537 - val_mape: 1115081.7500\n",
            "Epoch 309/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2516 - mape: 1317743.3750 - val_loss: 0.2536 - val_mape: 1385989.0000\n",
            "Epoch 310/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2517 - mape: 1351574.0000 - val_loss: 0.2536 - val_mape: 1133966.3750\n",
            "Epoch 311/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2517 - mape: 1326835.7500 - val_loss: 0.2541 - val_mape: 1504819.1250\n",
            "Epoch 312/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2516 - mape: 1336894.8750 - val_loss: 0.2537 - val_mape: 1222683.5000\n",
            "Epoch 313/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2516 - mape: 1328406.0000 - val_loss: 0.2538 - val_mape: 1434564.0000\n",
            "Epoch 314/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2516 - mape: 1332679.0000 - val_loss: 0.2536 - val_mape: 1242721.2500\n",
            "Epoch 315/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2516 - mape: 1320533.5000 - val_loss: 0.2537 - val_mape: 1515829.1250\n",
            "Epoch 316/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2516 - mape: 1329277.7500 - val_loss: 0.2536 - val_mape: 1125172.1250\n",
            "Epoch 317/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2517 - mape: 1327898.5000 - val_loss: 0.2536 - val_mape: 1204750.1250\n",
            "Epoch 318/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2515 - mape: 1315133.6250 - val_loss: 0.2537 - val_mape: 1393342.3750\n",
            "Epoch 319/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2517 - mape: 1326131.3750 - val_loss: 0.2536 - val_mape: 1245151.1250\n",
            "Epoch 320/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2516 - mape: 1328768.7500 - val_loss: 0.2540 - val_mape: 1108726.8750\n",
            "Epoch 321/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2514 - mape: 1298960.1250 - val_loss: 0.2534 - val_mape: 1320331.8750\n",
            "Epoch 322/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2517 - mape: 1338070.5000 - val_loss: 0.2534 - val_mape: 1160943.8750\n",
            "Epoch 323/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2516 - mape: 1309694.6250 - val_loss: 0.2544 - val_mape: 1741530.6250\n",
            "Epoch 324/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2515 - mape: 1322706.8750 - val_loss: 0.2537 - val_mape: 1161599.7500\n",
            "Epoch 325/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2516 - mape: 1312705.0000 - val_loss: 0.2539 - val_mape: 1565550.0000\n",
            "Epoch 326/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2515 - mape: 1319106.8750 - val_loss: 0.2537 - val_mape: 1059247.8750\n",
            "Epoch 327/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2514 - mape: 1300822.3750 - val_loss: 0.2534 - val_mape: 1196408.0000\n",
            "Epoch 328/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2513 - mape: 1294098.3750 - val_loss: 0.2535 - val_mape: 1132666.7500\n",
            "Epoch 329/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2517 - mape: 1330631.0000 - val_loss: 0.2534 - val_mape: 1425897.5000\n",
            "Epoch 330/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2516 - mape: 1319391.2500 - val_loss: 0.2536 - val_mape: 1172078.7500\n",
            "Epoch 331/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2514 - mape: 1295491.7500 - val_loss: 0.2536 - val_mape: 1293993.3750\n",
            "Epoch 332/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2516 - mape: 1313156.6250 - val_loss: 0.2538 - val_mape: 1459068.3750\n",
            "Epoch 333/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2514 - mape: 1306113.0000 - val_loss: 0.2538 - val_mape: 1042986.3750\n",
            "Epoch 334/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2515 - mape: 1310589.6250 - val_loss: 0.2537 - val_mape: 1488614.5000\n",
            "Epoch 335/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2515 - mape: 1297675.6250 - val_loss: 0.2535 - val_mape: 1108706.1250\n",
            "Epoch 336/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2515 - mape: 1307575.6250 - val_loss: 0.2538 - val_mape: 1573265.5000\n",
            "Epoch 337/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2513 - mape: 1285037.1250 - val_loss: 0.2534 - val_mape: 1401648.2500\n",
            "Epoch 338/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2516 - mape: 1324679.7500 - val_loss: 0.2535 - val_mape: 1159558.7500\n",
            "Epoch 339/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2514 - mape: 1294749.3750 - val_loss: 0.2539 - val_mape: 1453094.8750\n",
            "Epoch 340/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2514 - mape: 1301826.7500 - val_loss: 0.2533 - val_mape: 1179399.6250\n",
            "Epoch 341/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2515 - mape: 1295799.5000 - val_loss: 0.2534 - val_mape: 1134882.5000\n",
            "Epoch 342/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2515 - mape: 1296631.2500 - val_loss: 0.2534 - val_mape: 1381153.0000\n",
            "Epoch 343/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2514 - mape: 1299106.5000 - val_loss: 0.2535 - val_mape: 1140980.2500\n",
            "Epoch 344/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2513 - mape: 1292155.2500 - val_loss: 0.2535 - val_mape: 1399607.3750\n",
            "Epoch 345/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2515 - mape: 1293794.5000 - val_loss: 0.2535 - val_mape: 1225983.3750\n",
            "Epoch 346/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2514 - mape: 1295000.2500 - val_loss: 0.2533 - val_mape: 1389151.6250\n",
            "Epoch 347/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2514 - mape: 1288800.3750 - val_loss: 0.2536 - val_mape: 1066278.3750\n",
            "Epoch 348/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2514 - mape: 1281670.7500 - val_loss: 0.2539 - val_mape: 1366268.2500\n",
            "Epoch 349/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2515 - mape: 1300607.7500 - val_loss: 0.2534 - val_mape: 1146990.3750\n",
            "Epoch 350/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2513 - mape: 1278495.6250 - val_loss: 0.2538 - val_mape: 1440717.1250\n",
            "Epoch 351/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2514 - mape: 1291792.8750 - val_loss: 0.2536 - val_mape: 1169248.0000\n",
            "Epoch 352/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2515 - mape: 1293549.6250 - val_loss: 0.2536 - val_mape: 1555439.8750\n",
            "Epoch 353/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2513 - mape: 1279869.3750 - val_loss: 0.2535 - val_mape: 1188486.6250\n",
            "Epoch 354/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2514 - mape: 1286700.6250 - val_loss: 0.2534 - val_mape: 1297923.3750\n",
            "Epoch 355/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2513 - mape: 1286692.1250 - val_loss: 0.2538 - val_mape: 1128020.3750\n",
            "Epoch 356/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2513 - mape: 1278009.7500 - val_loss: 0.2533 - val_mape: 1349471.5000\n",
            "Epoch 357/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2515 - mape: 1295573.2500 - val_loss: 0.2533 - val_mape: 1174498.1250\n",
            "Epoch 358/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2513 - mape: 1271113.8750 - val_loss: 0.2535 - val_mape: 1481019.7500\n",
            "Epoch 359/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2513 - mape: 1284045.2500 - val_loss: 0.2534 - val_mape: 1144599.7500\n",
            "Epoch 360/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2513 - mape: 1269372.7500 - val_loss: 0.2536 - val_mape: 1051285.5000\n",
            "Epoch 361/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2514 - mape: 1286780.7500 - val_loss: 0.2533 - val_mape: 1408069.0000\n",
            "Epoch 362/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2514 - mape: 1282001.2500 - val_loss: 0.2534 - val_mape: 1145451.7500\n",
            "Epoch 363/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2513 - mape: 1272488.1250 - val_loss: 0.2536 - val_mape: 1582321.7500\n",
            "Epoch 364/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2513 - mape: 1267895.3750 - val_loss: 0.2535 - val_mape: 1416834.6250\n",
            "Epoch 365/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2514 - mape: 1287733.7500 - val_loss: 0.2533 - val_mape: 1114889.7500\n",
            "Epoch 366/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2511 - mape: 1254727.5000 - val_loss: 0.2542 - val_mape: 1134198.1250\n",
            "Epoch 367/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2514 - mape: 1285826.8750 - val_loss: 0.2532 - val_mape: 1344083.0000\n",
            "Epoch 368/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2514 - mape: 1275739.2500 - val_loss: 0.2533 - val_mape: 1396542.5000\n",
            "Epoch 369/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2512 - mape: 1268154.2500 - val_loss: 0.2542 - val_mape: 1125794.8750\n",
            "Epoch 370/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2513 - mape: 1272084.8750 - val_loss: 0.2535 - val_mape: 1375549.5000\n",
            "Epoch 371/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2513 - mape: 1275030.2500 - val_loss: 0.2532 - val_mape: 1228839.2500\n",
            "Epoch 372/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2513 - mape: 1268082.7500 - val_loss: 0.2533 - val_mape: 1354811.2500\n",
            "Epoch 373/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2513 - mape: 1262466.2500 - val_loss: 0.2531 - val_mape: 1250624.1250\n",
            "Epoch 374/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2513 - mape: 1273825.3750 - val_loss: 0.2539 - val_mape: 1097533.5000\n",
            "Epoch 375/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2512 - mape: 1259789.0000 - val_loss: 0.2533 - val_mape: 1344058.3750\n",
            "Epoch 376/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2513 - mape: 1272556.6250 - val_loss: 0.2535 - val_mape: 1209915.7500\n",
            "Epoch 377/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2513 - mape: 1264560.2500 - val_loss: 0.2536 - val_mape: 1435959.2500\n",
            "Epoch 378/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2512 - mape: 1255971.3750 - val_loss: 0.2534 - val_mape: 1229986.8750\n",
            "Epoch 379/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2513 - mape: 1268291.8750 - val_loss: 0.2539 - val_mape: 1532597.6250\n",
            "Epoch 380/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2512 - mape: 1263354.7500 - val_loss: 0.2535 - val_mape: 1117578.5000\n",
            "Epoch 381/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2513 - mape: 1253805.7500 - val_loss: 0.2535 - val_mape: 1507463.8750\n",
            "Epoch 382/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2513 - mape: 1261707.2500 - val_loss: 0.2531 - val_mape: 1200844.2500\n",
            "Epoch 383/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2513 - mape: 1265080.7500 - val_loss: 0.2534 - val_mape: 1390796.2500\n",
            "Epoch 384/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2511 - mape: 1255574.0000 - val_loss: 0.2532 - val_mape: 1184252.2500\n",
            "Epoch 385/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2513 - mape: 1265626.2500 - val_loss: 0.2532 - val_mape: 1116670.7500\n",
            "Epoch 386/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2511 - mape: 1244039.7500 - val_loss: 0.2538 - val_mape: 1352625.1250\n",
            "Epoch 387/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2512 - mape: 1264952.3750 - val_loss: 0.2535 - val_mape: 1070618.0000\n",
            "Epoch 388/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2512 - mape: 1250311.2500 - val_loss: 0.2531 - val_mape: 1141736.2500\n",
            "Epoch 389/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2513 - mape: 1263060.0000 - val_loss: 0.2533 - val_mape: 1386700.3750\n",
            "Epoch 390/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2512 - mape: 1253609.6250 - val_loss: 0.2535 - val_mape: 1011575.8125\n",
            "Epoch 391/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2512 - mape: 1252876.1250 - val_loss: 0.2535 - val_mape: 1538844.2500\n",
            "Epoch 392/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2512 - mape: 1262094.0000 - val_loss: 0.2533 - val_mape: 1077263.6250\n",
            "Epoch 393/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2511 - mape: 1238973.6250 - val_loss: 0.2532 - val_mape: 1191744.2500\n",
            "Epoch 394/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2513 - mape: 1265963.7500 - val_loss: 0.2534 - val_mape: 1097812.1250\n",
            "Epoch 395/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2511 - mape: 1242732.1250 - val_loss: 0.2537 - val_mape: 1407170.8750\n",
            "Epoch 396/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2511 - mape: 1250482.3750 - val_loss: 0.2531 - val_mape: 1158768.3750\n",
            "Epoch 397/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2513 - mape: 1255380.5000 - val_loss: 0.2532 - val_mape: 1065737.1250\n",
            "Epoch 398/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2512 - mape: 1243471.3750 - val_loss: 0.2534 - val_mape: 1392885.3750\n",
            "Epoch 399/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2511 - mape: 1249496.8750 - val_loss: 0.2533 - val_mape: 1117949.0000\n",
            "Epoch 400/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2511 - mape: 1235189.5000 - val_loss: 0.2533 - val_mape: 1048631.0000\n",
            "Epoch 401/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2512 - mape: 1249519.6250 - val_loss: 0.2532 - val_mape: 1388125.0000\n",
            "Epoch 402/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2512 - mape: 1249076.3750 - val_loss: 0.2533 - val_mape: 1033249.3750\n",
            "Epoch 403/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2511 - mape: 1237048.8750 - val_loss: 0.2531 - val_mape: 1165574.2500\n",
            "Epoch 404/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2512 - mape: 1248002.7500 - val_loss: 0.2535 - val_mape: 1411894.7500\n",
            "Epoch 405/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2511 - mape: 1243554.1250 - val_loss: 0.2535 - val_mape: 1120816.2500\n",
            "Epoch 406/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2511 - mape: 1244257.0000 - val_loss: 0.2534 - val_mape: 1379875.0000\n",
            "Epoch 407/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2511 - mape: 1225067.5000 - val_loss: 0.2533 - val_mape: 1301484.8750\n",
            "Epoch 408/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2513 - mape: 1265229.0000 - val_loss: 0.2533 - val_mape: 1047877.1875\n",
            "Epoch 409/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2510 - mape: 1217732.7500 - val_loss: 0.2534 - val_mape: 1245293.5000\n",
            "Epoch 410/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2512 - mape: 1252169.3750 - val_loss: 0.2533 - val_mape: 1102348.3750\n",
            "Epoch 411/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2511 - mape: 1237419.5000 - val_loss: 0.2533 - val_mape: 1383192.1250\n",
            "Epoch 412/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2510 - mape: 1230523.6250 - val_loss: 0.2537 - val_mape: 1309272.0000\n",
            "Epoch 413/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2512 - mape: 1248389.6250 - val_loss: 0.2534 - val_mape: 1036928.4375\n",
            "Epoch 414/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2511 - mape: 1233748.8750 - val_loss: 0.2535 - val_mape: 1292465.1250\n",
            "Epoch 415/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2511 - mape: 1237440.7500 - val_loss: 0.2533 - val_mape: 1121490.2500\n",
            "Epoch 416/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2510 - mape: 1235983.8750 - val_loss: 0.2537 - val_mape: 1339960.2500\n",
            "Epoch 417/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2510 - mape: 1225996.1250 - val_loss: 0.2533 - val_mape: 1454268.3750\n",
            "Epoch 418/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2512 - mape: 1249487.1250 - val_loss: 0.2532 - val_mape: 1081195.8750\n",
            "Epoch 419/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2511 - mape: 1230295.2500 - val_loss: 0.2531 - val_mape: 1369540.7500\n",
            "Epoch 420/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2511 - mape: 1229306.0000 - val_loss: 0.2533 - val_mape: 1047991.5000\n",
            "Epoch 421/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2510 - mape: 1227154.3750 - val_loss: 0.2535 - val_mape: 1454767.5000\n",
            "Epoch 422/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2511 - mape: 1233481.5000 - val_loss: 0.2532 - val_mape: 1092914.5000\n",
            "Epoch 423/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1228091.7500 - val_loss: 0.2531 - val_mape: 1328308.3750\n",
            "Epoch 424/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2512 - mape: 1238520.2500 - val_loss: 0.2531 - val_mape: 1143280.3750\n",
            "Epoch 425/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2511 - mape: 1223532.3750 - val_loss: 0.2534 - val_mape: 1477116.8750\n",
            "Epoch 426/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2510 - mape: 1221168.2500 - val_loss: 0.2531 - val_mape: 1347516.1250\n",
            "Epoch 427/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2511 - mape: 1243633.6250 - val_loss: 0.2531 - val_mape: 1066673.5000\n",
            "Epoch 428/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2511 - mape: 1227658.7500 - val_loss: 0.2532 - val_mape: 1329606.2500\n",
            "Epoch 429/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2510 - mape: 1225759.7500 - val_loss: 0.2531 - val_mape: 1160330.8750\n",
            "Epoch 430/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2510 - mape: 1217487.2500 - val_loss: 0.2537 - val_mape: 1451781.6250\n",
            "Epoch 431/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2510 - mape: 1233313.3750 - val_loss: 0.2531 - val_mape: 1074421.7500\n",
            "Epoch 432/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2510 - mape: 1222838.1250 - val_loss: 0.2531 - val_mape: 1060317.1250\n",
            "Epoch 433/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2511 - mape: 1224028.1250 - val_loss: 0.2530 - val_mape: 1320201.3750\n",
            "Epoch 434/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2511 - mape: 1230440.0000 - val_loss: 0.2532 - val_mape: 1055310.2500\n",
            "Epoch 435/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1210968.2500 - val_loss: 0.2532 - val_mape: 1116401.7500\n",
            "Epoch 436/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2511 - mape: 1225416.6250 - val_loss: 0.2533 - val_mape: 1507588.8750\n",
            "Epoch 437/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2511 - mape: 1232614.0000 - val_loss: 0.2531 - val_mape: 1064089.2500\n",
            "Epoch 438/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2509 - mape: 1206445.8750 - val_loss: 0.2535 - val_mape: 1582323.7500\n",
            "Epoch 439/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2510 - mape: 1224352.6250 - val_loss: 0.2533 - val_mape: 1132168.3750\n",
            "Epoch 440/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2510 - mape: 1221639.0000 - val_loss: 0.2531 - val_mape: 1284202.5000\n",
            "Epoch 441/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2510 - mape: 1218141.2500 - val_loss: 0.2531 - val_mape: 1125234.0000\n",
            "Epoch 442/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2510 - mape: 1224375.2500 - val_loss: 0.2533 - val_mape: 1273027.0000\n",
            "Epoch 443/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2510 - mape: 1214872.2500 - val_loss: 0.2535 - val_mape: 1380256.3750\n",
            "Epoch 444/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2510 - mape: 1218250.1250 - val_loss: 0.2531 - val_mape: 1179108.0000\n",
            "Epoch 445/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2510 - mape: 1221218.8750 - val_loss: 0.2531 - val_mape: 1330328.8750\n",
            "Epoch 446/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2509 - mape: 1210493.1250 - val_loss: 0.2539 - val_mape: 1437302.0000\n",
            "Epoch 447/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2510 - mape: 1218423.3750 - val_loss: 0.2530 - val_mape: 1202109.2500\n",
            "Epoch 448/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2511 - mape: 1224936.5000 - val_loss: 0.2530 - val_mape: 1327076.7500\n",
            "Epoch 449/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2510 - mape: 1213046.7500 - val_loss: 0.2533 - val_mape: 1011297.3750\n",
            "Epoch 450/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1212071.3750 - val_loss: 0.2533 - val_mape: 1271418.2500\n",
            "Epoch 451/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2510 - mape: 1219078.2500 - val_loss: 0.2532 - val_mape: 1020617.7500\n",
            "Epoch 452/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2510 - mape: 1207268.7500 - val_loss: 0.2533 - val_mape: 1250600.7500\n",
            "Epoch 453/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2510 - mape: 1219574.1250 - val_loss: 0.2530 - val_mape: 1127423.0000\n",
            "Epoch 454/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2510 - mape: 1207818.3750 - val_loss: 0.2534 - val_mape: 1045295.4375\n",
            "Epoch 455/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1204671.7500 - val_loss: 0.2534 - val_mape: 1297238.5000\n",
            "Epoch 456/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2510 - mape: 1219854.1250 - val_loss: 0.2534 - val_mape: 1014815.0625\n",
            "Epoch 457/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1198316.1250 - val_loss: 0.2531 - val_mape: 1257066.1250\n",
            "Epoch 458/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2510 - mape: 1224305.0000 - val_loss: 0.2530 - val_mape: 1098679.7500\n",
            "Epoch 459/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2508 - mape: 1191489.0000 - val_loss: 0.2530 - val_mape: 1061759.6250\n",
            "Epoch 460/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2511 - mape: 1226005.7500 - val_loss: 0.2529 - val_mape: 1333123.0000\n",
            "Epoch 461/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2509 - mape: 1209264.7500 - val_loss: 0.2529 - val_mape: 1062418.7500\n",
            "Epoch 462/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1202144.7500 - val_loss: 0.2535 - val_mape: 991157.1250\n",
            "Epoch 463/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1207983.6250 - val_loss: 0.2532 - val_mape: 1286079.8750\n",
            "Epoch 464/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1205857.8750 - val_loss: 0.2534 - val_mape: 985531.6875\n",
            "Epoch 465/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2510 - mape: 1206376.7500 - val_loss: 0.2533 - val_mape: 1502755.2500\n",
            "Epoch 466/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2510 - mape: 1212629.7500 - val_loss: 0.2530 - val_mape: 1098961.3750\n",
            "Epoch 467/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2509 - mape: 1195175.8750 - val_loss: 0.2529 - val_mape: 1200215.1250\n",
            "Epoch 468/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1206395.8750 - val_loss: 0.2532 - val_mape: 1315053.8750\n",
            "Epoch 469/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2508 - mape: 1203894.3750 - val_loss: 0.2529 - val_mape: 1095590.1250\n",
            "Epoch 470/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2510 - mape: 1206239.8750 - val_loss: 0.2530 - val_mape: 1166012.0000\n",
            "Epoch 471/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1200484.5000 - val_loss: 0.2542 - val_mape: 1017204.1250\n",
            "Epoch 472/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1199567.7500 - val_loss: 0.2530 - val_mape: 1305013.2500\n",
            "Epoch 473/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2510 - mape: 1208979.7500 - val_loss: 0.2529 - val_mape: 1075227.3750\n",
            "Epoch 474/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1199443.0000 - val_loss: 0.2534 - val_mape: 1045527.4375\n",
            "Epoch 475/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2509 - mape: 1198812.5000 - val_loss: 0.2533 - val_mape: 1393185.5000\n",
            "Epoch 476/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2509 - mape: 1198763.3750 - val_loss: 0.2533 - val_mape: 1099171.2500\n",
            "Epoch 477/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2509 - mape: 1203700.5000 - val_loss: 0.2531 - val_mape: 1321026.1250\n",
            "Epoch 478/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2509 - mape: 1198560.5000 - val_loss: 0.2529 - val_mape: 1097064.1250\n",
            "Epoch 479/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1199218.0000 - val_loss: 0.2532 - val_mape: 1325558.5000\n",
            "Epoch 480/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1200984.5000 - val_loss: 0.2533 - val_mape: 1026286.6250\n",
            "Epoch 481/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2508 - mape: 1193979.3750 - val_loss: 0.2529 - val_mape: 1289029.8750\n",
            "Epoch 482/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1202959.1250 - val_loss: 0.2531 - val_mape: 1039297.0625\n",
            "Epoch 483/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2509 - mape: 1189012.2500 - val_loss: 0.2531 - val_mape: 1398385.2500\n",
            "Epoch 484/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1203732.3750 - val_loss: 0.2530 - val_mape: 1068340.6250\n",
            "Epoch 485/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2508 - mape: 1179202.5000 - val_loss: 0.2532 - val_mape: 1007283.0000\n",
            "Epoch 486/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1209004.8750 - val_loss: 0.2529 - val_mape: 1316713.2500\n",
            "Epoch 487/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1197567.6250 - val_loss: 0.2529 - val_mape: 1072587.1250\n",
            "Epoch 488/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1192918.2500 - val_loss: 0.2531 - val_mape: 1257264.7500\n",
            "Epoch 489/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2507 - mape: 1189286.2500 - val_loss: 0.2535 - val_mape: 1040696.6250\n",
            "Epoch 490/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1194844.5000 - val_loss: 0.2530 - val_mape: 1326495.2500\n",
            "Epoch 491/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1192835.0000 - val_loss: 0.2530 - val_mape: 1079503.3750\n",
            "Epoch 492/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1188041.1250 - val_loss: 0.2530 - val_mape: 1312713.7500\n",
            "Epoch 493/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2508 - mape: 1192915.5000 - val_loss: 0.2531 - val_mape: 1202624.2500\n",
            "Epoch 494/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2509 - mape: 1193944.8750 - val_loss: 0.2529 - val_mape: 1052498.3750\n",
            "Epoch 495/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2508 - mape: 1189659.7500 - val_loss: 0.2533 - val_mape: 1321485.1250\n",
            "Epoch 496/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2508 - mape: 1189549.0000 - val_loss: 0.2530 - val_mape: 1038789.8750\n",
            "Epoch 497/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1190766.0000 - val_loss: 0.2528 - val_mape: 1260964.5000\n",
            "Epoch 498/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2508 - mape: 1179133.8750 - val_loss: 0.2529 - val_mape: 1049181.0000\n",
            "Epoch 499/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2509 - mape: 1198644.3750 - val_loss: 0.2531 - val_mape: 1417801.2500\n",
            "Epoch 500/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2508 - mape: 1190924.0000 - val_loss: 0.2532 - val_mape: 1092874.5000\n",
            "Epoch 501/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2508 - mape: 1185467.1250 - val_loss: 0.2529 - val_mape: 1274993.2500\n",
            "Epoch 502/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2508 - mape: 1191557.5000 - val_loss: 0.2529 - val_mape: 1054592.8750\n",
            "Epoch 503/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2509 - mape: 1186910.2500 - val_loss: 0.2529 - val_mape: 1297100.0000\n",
            "Epoch 504/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2506 - mape: 1165788.7500 - val_loss: 0.2531 - val_mape: 1116839.6250\n",
            "Epoch 505/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2510 - mape: 1198814.0000 - val_loss: 0.2531 - val_mape: 1127012.7500\n",
            "Epoch 506/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1193444.0000 - val_loss: 0.2529 - val_mape: 1335904.5000\n",
            "Epoch 507/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2508 - mape: 1186535.8750 - val_loss: 0.2531 - val_mape: 1087463.7500\n",
            "Epoch 508/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2508 - mape: 1189657.3750 - val_loss: 0.2531 - val_mape: 1307184.2500\n",
            "Epoch 509/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2508 - mape: 1175356.8750 - val_loss: 0.2532 - val_mape: 1458192.6250\n",
            "Epoch 510/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2508 - mape: 1196892.7500 - val_loss: 0.2529 - val_mape: 1012324.8125\n",
            "Epoch 511/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2508 - mape: 1178295.7500 - val_loss: 0.2530 - val_mape: 1409416.5000\n",
            "Epoch 512/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2508 - mape: 1184493.2500 - val_loss: 0.2531 - val_mape: 966823.7500\n",
            "Epoch 513/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2508 - mape: 1183697.6250 - val_loss: 0.2529 - val_mape: 1176575.7500\n",
            "Epoch 514/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2507 - mape: 1175972.8750 - val_loss: 0.2533 - val_mape: 1378298.3750\n",
            "Epoch 515/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2508 - mape: 1181669.1250 - val_loss: 0.2530 - val_mape: 1105292.0000\n",
            "Epoch 516/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2508 - mape: 1181445.3750 - val_loss: 0.2529 - val_mape: 1267772.2500\n",
            "Epoch 517/1000\n",
            "53/53 [==============================] - 3s 50ms/step - loss: 0.2508 - mape: 1183753.8750 - val_loss: 0.2528 - val_mape: 1078391.8750\n",
            "Epoch 518/1000\n",
            "53/53 [==============================] - 3s 48ms/step - loss: 0.2508 - mape: 1180046.3750 - val_loss: 0.2531 - val_mape: 1394519.6250\n",
            "Epoch 519/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2508 - mape: 1177072.6250 - val_loss: 0.2531 - val_mape: 1394851.8750\n",
            "Epoch 520/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2508 - mape: 1183940.1250 - val_loss: 0.2529 - val_mape: 1070006.2500\n",
            "Epoch 521/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2507 - mape: 1170854.2500 - val_loss: 0.2529 - val_mape: 1010795.6250\n",
            "Epoch 522/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2508 - mape: 1179537.8750 - val_loss: 0.2529 - val_mape: 1286436.0000\n",
            "Epoch 523/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2508 - mape: 1188019.0000 - val_loss: 0.2531 - val_mape: 1022327.2500\n",
            "Epoch 524/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2508 - mape: 1175482.8750 - val_loss: 0.2531 - val_mape: 1305703.3750\n",
            "Epoch 525/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2507 - mape: 1173603.8750 - val_loss: 0.2529 - val_mape: 1179257.6250\n",
            "Epoch 526/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2508 - mape: 1182082.2500 - val_loss: 0.2528 - val_mape: 1221536.3750\n",
            "Epoch 527/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2508 - mape: 1177426.0000 - val_loss: 0.2530 - val_mape: 1075540.8750\n",
            "Epoch 528/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2507 - mape: 1174955.8750 - val_loss: 0.2528 - val_mape: 1329066.5000\n",
            "Epoch 529/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2508 - mape: 1179078.3750 - val_loss: 0.2528 - val_mape: 1024812.0625\n",
            "Epoch 530/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2508 - mape: 1176243.7500 - val_loss: 0.2528 - val_mape: 1293848.0000\n",
            "Epoch 531/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2507 - mape: 1166411.5000 - val_loss: 0.2531 - val_mape: 987400.5000\n",
            "Epoch 532/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2507 - mape: 1175066.8750 - val_loss: 0.2529 - val_mape: 1310224.1250\n",
            "Epoch 533/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2508 - mape: 1175123.6250 - val_loss: 0.2528 - val_mape: 1130189.0000\n",
            "Epoch 534/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2508 - mape: 1177458.3750 - val_loss: 0.2531 - val_mape: 1267697.3750\n",
            "Epoch 535/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2506 - mape: 1160327.8750 - val_loss: 0.2527 - val_mape: 1240658.7500\n",
            "Epoch 536/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2509 - mape: 1189145.8750 - val_loss: 0.2527 - val_mape: 1072496.8750\n",
            "Epoch 537/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2507 - mape: 1167508.1250 - val_loss: 0.2532 - val_mape: 1345336.5000\n",
            "Epoch 538/1000\n",
            "53/53 [==============================] - 3s 47ms/step - loss: 0.2506 - mape: 1167113.8750 - val_loss: 0.2530 - val_mape: 1213562.8750\n",
            "Epoch 539/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2508 - mape: 1182833.0000 - val_loss: 0.2528 - val_mape: 1053585.7500\n",
            "Epoch 540/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2507 - mape: 1163697.8750 - val_loss: 0.2528 - val_mape: 1173181.5000\n",
            "Epoch 541/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2508 - mape: 1178323.1250 - val_loss: 0.2530 - val_mape: 1162010.1250\n",
            "Epoch 542/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2506 - mape: 1159910.2500 - val_loss: 0.2529 - val_mape: 1012485.2500\n",
            "Epoch 543/1000\n",
            "53/53 [==============================] - 3s 48ms/step - loss: 0.2508 - mape: 1177541.1250 - val_loss: 0.2527 - val_mape: 1221980.0000\n",
            "Epoch 544/1000\n",
            "53/53 [==============================] - 3s 48ms/step - loss: 0.2508 - mape: 1171580.3750 - val_loss: 0.2528 - val_mape: 1273210.5000\n",
            "Epoch 545/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2507 - mape: 1165987.2500 - val_loss: 0.2538 - val_mape: 1051161.8750\n",
            "Epoch 546/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2507 - mape: 1166777.5000 - val_loss: 0.2528 - val_mape: 1222582.8750\n",
            "Epoch 547/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2508 - mape: 1175783.2500 - val_loss: 0.2528 - val_mape: 1100698.5000\n",
            "Epoch 548/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2507 - mape: 1163408.5000 - val_loss: 0.2537 - val_mape: 1379782.2500\n",
            "Epoch 549/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2508 - mape: 1173587.2500 - val_loss: 0.2532 - val_mape: 942332.6875\n",
            "Epoch 550/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2507 - mape: 1168313.2500 - val_loss: 0.2530 - val_mape: 1197911.3750\n",
            "Epoch 551/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2507 - mape: 1169877.7500 - val_loss: 0.2531 - val_mape: 1006232.7500\n",
            "Epoch 552/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2507 - mape: 1164594.8750 - val_loss: 0.2530 - val_mape: 1389104.8750\n",
            "Epoch 553/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2506 - mape: 1151986.8750 - val_loss: 0.2528 - val_mape: 1070933.8750\n",
            "Epoch 554/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2508 - mape: 1172671.2500 - val_loss: 0.2530 - val_mape: 1370893.0000\n",
            "Epoch 555/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2507 - mape: 1173597.1250 - val_loss: 0.2529 - val_mape: 1031116.2500\n",
            "Epoch 556/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2508 - mape: 1166656.3750 - val_loss: 0.2529 - val_mape: 1352534.7500\n",
            "Epoch 557/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2507 - mape: 1165001.2500 - val_loss: 0.2529 - val_mape: 1023835.2500\n",
            "Epoch 558/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1153511.0000 - val_loss: 0.2531 - val_mape: 1058711.7500\n",
            "Epoch 559/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2508 - mape: 1173753.7500 - val_loss: 0.2529 - val_mape: 1351468.8750\n",
            "Epoch 560/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2507 - mape: 1165859.7500 - val_loss: 0.2529 - val_mape: 979352.6250\n",
            "Epoch 561/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2507 - mape: 1158509.8750 - val_loss: 0.2536 - val_mape: 1347490.2500\n",
            "Epoch 562/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2507 - mape: 1167943.2500 - val_loss: 0.2528 - val_mape: 996189.6875\n",
            "Epoch 563/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2508 - mape: 1164686.1250 - val_loss: 0.2528 - val_mape: 1027371.0000\n",
            "Epoch 564/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2507 - mape: 1161680.3750 - val_loss: 0.2529 - val_mape: 1342722.5000\n",
            "Epoch 565/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2506 - mape: 1152361.1250 - val_loss: 0.2528 - val_mape: 1053911.7500\n",
            "Epoch 566/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2507 - mape: 1167700.0000 - val_loss: 0.2529 - val_mape: 1333822.0000\n",
            "Epoch 567/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1162275.0000 - val_loss: 0.2531 - val_mape: 1016623.3125\n",
            "Epoch 568/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2507 - mape: 1164079.0000 - val_loss: 0.2527 - val_mape: 1255645.2500\n",
            "Epoch 569/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2507 - mape: 1162035.7500 - val_loss: 0.2530 - val_mape: 1323369.2500\n",
            "Epoch 570/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1154249.5000 - val_loss: 0.2527 - val_mape: 1203739.7500\n",
            "Epoch 571/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2508 - mape: 1168332.7500 - val_loss: 0.2528 - val_mape: 1016803.4375\n",
            "Epoch 572/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2506 - mape: 1156042.0000 - val_loss: 0.2530 - val_mape: 1239456.1250\n",
            "Epoch 573/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2507 - mape: 1161067.1250 - val_loss: 0.2529 - val_mape: 1056965.2500\n",
            "Epoch 574/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2507 - mape: 1161607.6250 - val_loss: 0.2527 - val_mape: 1251274.8750\n",
            "Epoch 575/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2507 - mape: 1152859.3750 - val_loss: 0.2532 - val_mape: 1464237.2500\n",
            "Epoch 576/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2506 - mape: 1161109.3750 - val_loss: 0.2529 - val_mape: 1096232.5000\n",
            "Epoch 577/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2507 - mape: 1163509.7500 - val_loss: 0.2528 - val_mape: 1298096.5000\n",
            "Epoch 578/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1149798.6250 - val_loss: 0.2529 - val_mape: 1137644.2500\n",
            "Epoch 579/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2508 - mape: 1167796.7500 - val_loss: 0.2528 - val_mape: 1024121.3750\n",
            "Epoch 580/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1151861.5000 - val_loss: 0.2529 - val_mape: 1291912.0000\n",
            "Epoch 581/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2506 - mape: 1158518.3750 - val_loss: 0.2528 - val_mape: 1023369.2500\n",
            "Epoch 582/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2507 - mape: 1159030.1250 - val_loss: 0.2527 - val_mape: 1236517.6250\n",
            "Epoch 583/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2506 - mape: 1155995.2500 - val_loss: 0.2530 - val_mape: 988872.5000\n",
            "Epoch 584/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2506 - mape: 1151190.3750 - val_loss: 0.2528 - val_mape: 1215490.0000\n",
            "Epoch 585/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2507 - mape: 1157319.7500 - val_loss: 0.2528 - val_mape: 1024863.3125\n",
            "Epoch 586/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1153863.1250 - val_loss: 0.2530 - val_mape: 1197358.2500\n",
            "Epoch 587/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2507 - mape: 1157256.6250 - val_loss: 0.2528 - val_mape: 1023028.7500\n",
            "Epoch 588/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1141754.7500 - val_loss: 0.2528 - val_mape: 986203.5625\n",
            "Epoch 589/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2507 - mape: 1164046.7500 - val_loss: 0.2527 - val_mape: 1190894.1250\n",
            "Epoch 590/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1157330.2500 - val_loss: 0.2527 - val_mape: 1045347.1250\n",
            "Epoch 591/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2506 - mape: 1139436.0000 - val_loss: 0.2529 - val_mape: 979817.9375\n",
            "Epoch 592/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2507 - mape: 1162438.7500 - val_loss: 0.2527 - val_mape: 1230818.7500\n",
            "Epoch 593/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2506 - mape: 1153041.3750 - val_loss: 0.2531 - val_mape: 1000032.7500\n",
            "Epoch 594/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1147072.0000 - val_loss: 0.2528 - val_mape: 1193743.0000\n",
            "Epoch 595/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2507 - mape: 1157319.1250 - val_loss: 0.2526 - val_mape: 1241362.3750\n",
            "Epoch 596/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2505 - mape: 1139022.0000 - val_loss: 0.2534 - val_mape: 1353946.6250\n",
            "Epoch 597/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2507 - mape: 1164900.7500 - val_loss: 0.2529 - val_mape: 1097109.5000\n",
            "Epoch 598/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1148329.1250 - val_loss: 0.2527 - val_mape: 1230762.6250\n",
            "Epoch 599/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2506 - mape: 1148923.5000 - val_loss: 0.2528 - val_mape: 966921.9375\n",
            "Epoch 600/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1145752.6250 - val_loss: 0.2530 - val_mape: 1408176.2500\n",
            "Epoch 601/1000\n",
            "53/53 [==============================] - 2s 44ms/step - loss: 0.2507 - mape: 1161538.6250 - val_loss: 0.2526 - val_mape: 1074408.8750\n",
            "Epoch 602/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2506 - mape: 1146379.2500 - val_loss: 0.2534 - val_mape: 1007226.5625\n",
            "Epoch 603/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1146355.8750 - val_loss: 0.2526 - val_mape: 1181036.2500\n",
            "Epoch 604/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2507 - mape: 1154118.2500 - val_loss: 0.2528 - val_mape: 1388079.2500\n",
            "Epoch 605/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1142989.0000 - val_loss: 0.2533 - val_mape: 907676.7500\n",
            "Epoch 606/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1152206.1250 - val_loss: 0.2530 - val_mape: 1288959.1250\n",
            "Epoch 607/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1145319.8750 - val_loss: 0.2529 - val_mape: 1193211.8750\n",
            "Epoch 608/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2507 - mape: 1155332.7500 - val_loss: 0.2528 - val_mape: 1049678.7500\n",
            "Epoch 609/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1145105.5000 - val_loss: 0.2528 - val_mape: 1247879.5000\n",
            "Epoch 610/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1152338.0000 - val_loss: 0.2527 - val_mape: 1073494.1250\n",
            "Epoch 611/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1139184.0000 - val_loss: 0.2530 - val_mape: 1447117.1250\n",
            "Epoch 612/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2506 - mape: 1155231.5000 - val_loss: 0.2527 - val_mape: 1003796.1875\n",
            "Epoch 613/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1132484.8750 - val_loss: 0.2529 - val_mape: 1001268.5625\n",
            "Epoch 614/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2507 - mape: 1159253.5000 - val_loss: 0.2528 - val_mape: 1314437.0000\n",
            "Epoch 615/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1139358.3750 - val_loss: 0.2526 - val_mape: 1130792.3750\n",
            "Epoch 616/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2505 - mape: 1146671.5000 - val_loss: 0.2533 - val_mape: 1011110.1250\n",
            "Epoch 617/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2506 - mape: 1148023.0000 - val_loss: 0.2528 - val_mape: 1284092.8750\n",
            "Epoch 618/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2507 - mape: 1149441.0000 - val_loss: 0.2528 - val_mape: 1009451.6875\n",
            "Epoch 619/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1145495.6250 - val_loss: 0.2527 - val_mape: 1251180.3750\n",
            "Epoch 620/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2506 - mape: 1138380.5000 - val_loss: 0.2527 - val_mape: 1206257.3750\n",
            "Epoch 621/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2506 - mape: 1153333.1250 - val_loss: 0.2530 - val_mape: 941443.8125\n",
            "Epoch 622/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2505 - mape: 1130836.5000 - val_loss: 0.2529 - val_mape: 977957.7500\n",
            "Epoch 623/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2507 - mape: 1154721.6250 - val_loss: 0.2528 - val_mape: 1203844.1250\n",
            "Epoch 624/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1139274.0000 - val_loss: 0.2530 - val_mape: 1001384.5000\n",
            "Epoch 625/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2506 - mape: 1145902.5000 - val_loss: 0.2526 - val_mape: 1189020.0000\n",
            "Epoch 626/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1141469.8750 - val_loss: 0.2527 - val_mape: 1014299.6875\n",
            "Epoch 627/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1144688.3750 - val_loss: 0.2528 - val_mape: 1330531.8750\n",
            "Epoch 628/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1134713.2500 - val_loss: 0.2532 - val_mape: 1227191.6250\n",
            "Epoch 629/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1151902.5000 - val_loss: 0.2529 - val_mape: 1034379.8750\n",
            "Epoch 630/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2506 - mape: 1138399.1250 - val_loss: 0.2527 - val_mape: 1327511.0000\n",
            "Epoch 631/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2506 - mape: 1146325.1250 - val_loss: 0.2530 - val_mape: 1226901.8750\n",
            "Epoch 632/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1147105.8750 - val_loss: 0.2527 - val_mape: 1024640.7500\n",
            "Epoch 633/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2504 - mape: 1118753.8750 - val_loss: 0.2526 - val_mape: 1033177.3750\n",
            "Epoch 634/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2507 - mape: 1155843.8750 - val_loss: 0.2526 - val_mape: 1198778.8750\n",
            "Epoch 635/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1141497.8750 - val_loss: 0.2528 - val_mape: 1015774.2500\n",
            "Epoch 636/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2505 - mape: 1132963.5000 - val_loss: 0.2531 - val_mape: 1305681.5000\n",
            "Epoch 637/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1144497.8750 - val_loss: 0.2527 - val_mape: 1011092.3125\n",
            "Epoch 638/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1134532.7500 - val_loss: 0.2528 - val_mape: 1351064.7500\n",
            "Epoch 639/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1143446.6250 - val_loss: 0.2527 - val_mape: 1060420.8750\n",
            "Epoch 640/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2506 - mape: 1141975.3750 - val_loss: 0.2530 - val_mape: 993911.1875\n",
            "Epoch 641/1000\n",
            "53/53 [==============================] - 3s 47ms/step - loss: 0.2504 - mape: 1133435.6250 - val_loss: 0.2527 - val_mape: 1218677.3750\n",
            "Epoch 642/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2507 - mape: 1145426.0000 - val_loss: 0.2527 - val_mape: 1226724.0000\n",
            "Epoch 643/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2505 - mape: 1136233.0000 - val_loss: 0.2529 - val_mape: 975178.5625\n",
            "Epoch 644/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1133112.5000 - val_loss: 0.2529 - val_mape: 1410253.7500\n",
            "Epoch 645/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1143400.6250 - val_loss: 0.2528 - val_mape: 1048733.3750\n",
            "Epoch 646/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2505 - mape: 1134962.1250 - val_loss: 0.2529 - val_mape: 1336563.7500\n",
            "Epoch 647/1000\n",
            "53/53 [==============================] - 3s 47ms/step - loss: 0.2505 - mape: 1135781.6250 - val_loss: 0.2527 - val_mape: 996393.7500\n",
            "Epoch 648/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1135807.8750 - val_loss: 0.2526 - val_mape: 1029537.8125\n",
            "Epoch 649/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2506 - mape: 1141251.3750 - val_loss: 0.2528 - val_mape: 1358289.7500\n",
            "Epoch 650/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2505 - mape: 1134411.8750 - val_loss: 0.2528 - val_mape: 1011074.8125\n",
            "Epoch 651/1000\n",
            "53/53 [==============================] - 3s 48ms/step - loss: 0.2505 - mape: 1134004.0000 - val_loss: 0.2527 - val_mape: 1096783.0000\n",
            "Epoch 652/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1143681.3750 - val_loss: 0.2525 - val_mape: 1197209.0000\n",
            "Epoch 653/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1133897.7500 - val_loss: 0.2528 - val_mape: 947136.4375\n",
            "Epoch 654/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2504 - mape: 1127294.2500 - val_loss: 0.2531 - val_mape: 1249002.3750\n",
            "Epoch 655/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2506 - mape: 1145992.7500 - val_loss: 0.2527 - val_mape: 970591.6250\n",
            "Epoch 656/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2505 - mape: 1126956.6250 - val_loss: 0.2527 - val_mape: 972441.2500\n",
            "Epoch 657/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2506 - mape: 1140369.5000 - val_loss: 0.2525 - val_mape: 1210165.6250\n",
            "Epoch 658/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2505 - mape: 1131778.7500 - val_loss: 0.2527 - val_mape: 958399.1250\n",
            "Epoch 659/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1132526.1250 - val_loss: 0.2527 - val_mape: 1107017.7500\n",
            "Epoch 660/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1138832.3750 - val_loss: 0.2528 - val_mape: 1005932.1250\n",
            "Epoch 661/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2504 - mape: 1126358.8750 - val_loss: 0.2526 - val_mape: 1097772.8750\n",
            "Epoch 662/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2506 - mape: 1140902.7500 - val_loss: 0.2526 - val_mape: 1032588.8125\n",
            "Epoch 663/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1129766.3750 - val_loss: 0.2526 - val_mape: 1140795.6250\n",
            "Epoch 664/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2505 - mape: 1130265.1250 - val_loss: 0.2530 - val_mape: 1298611.7500\n",
            "Epoch 665/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1136199.1250 - val_loss: 0.2528 - val_mape: 1037645.1875\n",
            "Epoch 666/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1134125.1250 - val_loss: 0.2526 - val_mape: 1136893.5000\n",
            "Epoch 667/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2505 - mape: 1131239.8750 - val_loss: 0.2527 - val_mape: 979169.3750\n",
            "Epoch 668/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1124494.1250 - val_loss: 0.2527 - val_mape: 973864.0625\n",
            "Epoch 669/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1137323.6250 - val_loss: 0.2526 - val_mape: 1190311.8750\n",
            "Epoch 670/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1131421.7500 - val_loss: 0.2527 - val_mape: 1014336.3125\n",
            "Epoch 671/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2505 - mape: 1125440.3750 - val_loss: 0.2527 - val_mape: 1227254.7500\n",
            "Epoch 672/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2505 - mape: 1137243.7500 - val_loss: 0.2527 - val_mape: 1070348.5000\n",
            "Epoch 673/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2505 - mape: 1130276.5000 - val_loss: 0.2527 - val_mape: 1270218.5000\n",
            "Epoch 674/1000\n",
            "53/53 [==============================] - 3s 48ms/step - loss: 0.2505 - mape: 1129697.7500 - val_loss: 0.2526 - val_mape: 993480.3750\n",
            "Epoch 675/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1121173.5000 - val_loss: 0.2529 - val_mape: 978748.7500\n",
            "Epoch 676/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1137604.2500 - val_loss: 0.2527 - val_mape: 1261063.7500\n",
            "Epoch 677/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1125506.0000 - val_loss: 0.2525 - val_mape: 1096122.5000\n",
            "Epoch 678/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1133222.3750 - val_loss: 0.2528 - val_mape: 1000508.6875\n",
            "Epoch 679/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1129952.8750 - val_loss: 0.2527 - val_mape: 1258868.2500\n",
            "Epoch 680/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2505 - mape: 1133231.1250 - val_loss: 0.2526 - val_mape: 1035687.9375\n",
            "Epoch 681/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1117583.5000 - val_loss: 0.2527 - val_mape: 1310245.8750\n",
            "Epoch 682/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2505 - mape: 1135845.6250 - val_loss: 0.2526 - val_mape: 1039926.8125\n",
            "Epoch 683/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1123039.8750 - val_loss: 0.2530 - val_mape: 1426116.5000\n",
            "Epoch 684/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1128140.6250 - val_loss: 0.2528 - val_mape: 1351165.7500\n",
            "Epoch 685/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1137267.6250 - val_loss: 0.2527 - val_mape: 986435.7500\n",
            "Epoch 686/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2505 - mape: 1119323.5000 - val_loss: 0.2526 - val_mape: 1216206.0000\n",
            "Epoch 687/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1127012.3750 - val_loss: 0.2528 - val_mape: 1058167.5000\n",
            "Epoch 688/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1111094.7500 - val_loss: 0.2526 - val_mape: 1055366.5000\n",
            "Epoch 689/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1144501.6250 - val_loss: 0.2527 - val_mape: 1331616.1250\n",
            "Epoch 690/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1130054.3750 - val_loss: 0.2527 - val_mape: 955924.3750\n",
            "Epoch 691/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2505 - mape: 1124076.3750 - val_loss: 0.2526 - val_mape: 1112920.2500\n",
            "Epoch 692/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1114457.5000 - val_loss: 0.2525 - val_mape: 1136465.7500\n",
            "Epoch 693/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1146083.8750 - val_loss: 0.2525 - val_mape: 1016279.2500\n",
            "Epoch 694/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1115681.7500 - val_loss: 0.2529 - val_mape: 1352341.5000\n",
            "Epoch 695/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1126636.0000 - val_loss: 0.2531 - val_mape: 1047743.6250\n",
            "Epoch 696/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2505 - mape: 1130858.6250 - val_loss: 0.2526 - val_mape: 1277593.5000\n",
            "Epoch 697/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2505 - mape: 1122035.3750 - val_loss: 0.2531 - val_mape: 1410215.0000\n",
            "Epoch 698/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1130419.0000 - val_loss: 0.2526 - val_mape: 1010813.3125\n",
            "Epoch 699/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1123340.0000 - val_loss: 0.2526 - val_mape: 1264322.7500\n",
            "Epoch 700/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1123398.6250 - val_loss: 0.2527 - val_mape: 964964.0000\n",
            "Epoch 701/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1122976.7500 - val_loss: 0.2527 - val_mape: 1246539.2500\n",
            "Epoch 702/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1126377.0000 - val_loss: 0.2525 - val_mape: 1053200.1250\n",
            "Epoch 703/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1115446.7500 - val_loss: 0.2526 - val_mape: 986247.3750\n",
            "Epoch 704/1000\n",
            "53/53 [==============================] - 3s 47ms/step - loss: 0.2505 - mape: 1127320.0000 - val_loss: 0.2525 - val_mape: 1197572.2500\n",
            "Epoch 705/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1122333.8750 - val_loss: 0.2527 - val_mape: 1267237.7500\n",
            "Epoch 706/1000\n",
            "53/53 [==============================] - 3s 49ms/step - loss: 0.2504 - mape: 1118329.8750 - val_loss: 0.2527 - val_mape: 1145138.5000\n",
            "Epoch 707/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2506 - mape: 1133793.7500 - val_loss: 0.2527 - val_mape: 998910.5000\n",
            "Epoch 708/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2504 - mape: 1118552.2500 - val_loss: 0.2528 - val_mape: 1193153.6250\n",
            "Epoch 709/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1124102.2500 - val_loss: 0.2526 - val_mape: 954127.2500\n",
            "Epoch 710/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2505 - mape: 1117432.2500 - val_loss: 0.2527 - val_mape: 1308867.2500\n",
            "Epoch 711/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1116128.7500 - val_loss: 0.2528 - val_mape: 957432.5625\n",
            "Epoch 712/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1126085.3750 - val_loss: 0.2526 - val_mape: 995348.6250\n",
            "Epoch 713/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2505 - mape: 1124985.8750 - val_loss: 0.2531 - val_mape: 1384765.3750\n",
            "Epoch 714/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2503 - mape: 1116704.5000 - val_loss: 0.2527 - val_mape: 1045340.4375\n",
            "Epoch 715/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1126822.2500 - val_loss: 0.2526 - val_mape: 1261544.0000\n",
            "Epoch 716/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1108558.3750 - val_loss: 0.2525 - val_mape: 1209013.7500\n",
            "Epoch 717/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1134865.1250 - val_loss: 0.2525 - val_mape: 1049085.3750\n",
            "Epoch 718/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1117611.7500 - val_loss: 0.2530 - val_mape: 1251307.1250\n",
            "Epoch 719/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1121826.0000 - val_loss: 0.2526 - val_mape: 1000653.9375\n",
            "Epoch 720/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1120681.6250 - val_loss: 0.2526 - val_mape: 1020689.0000\n",
            "Epoch 721/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2504 - mape: 1113773.0000 - val_loss: 0.2531 - val_mape: 1262230.0000\n",
            "Epoch 722/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2505 - mape: 1122695.2500 - val_loss: 0.2527 - val_mape: 1017339.0625\n",
            "Epoch 723/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1116429.0000 - val_loss: 0.2527 - val_mape: 968709.8125\n",
            "Epoch 724/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1121812.8750 - val_loss: 0.2527 - val_mape: 1285642.6250\n",
            "Epoch 725/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2504 - mape: 1112757.0000 - val_loss: 0.2526 - val_mape: 1228881.7500\n",
            "Epoch 726/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1126401.7500 - val_loss: 0.2526 - val_mape: 996781.8125\n",
            "Epoch 727/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1110689.0000 - val_loss: 0.2530 - val_mape: 1029751.1250\n",
            "Epoch 728/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1117886.8750 - val_loss: 0.2529 - val_mape: 1179912.7500\n",
            "Epoch 729/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2504 - mape: 1128331.3750 - val_loss: 0.2526 - val_mape: 985577.1875\n",
            "Epoch 730/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2505 - mape: 1119397.6250 - val_loss: 0.2526 - val_mape: 1270457.8750\n",
            "Epoch 731/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1112236.8750 - val_loss: 0.2526 - val_mape: 954588.8125\n",
            "Epoch 732/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1120559.7500 - val_loss: 0.2528 - val_mape: 1407025.7500\n",
            "Epoch 733/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1123139.6250 - val_loss: 0.2526 - val_mape: 997320.6250\n",
            "Epoch 734/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2504 - mape: 1109893.7500 - val_loss: 0.2527 - val_mape: 1026503.3125\n",
            "Epoch 735/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1125700.3750 - val_loss: 0.2525 - val_mape: 1214708.1250\n",
            "Epoch 736/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2503 - mape: 1102320.5000 - val_loss: 0.2526 - val_mape: 1220096.2500\n",
            "Epoch 737/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1128368.8750 - val_loss: 0.2524 - val_mape: 1023106.8125\n",
            "Epoch 738/1000\n",
            "53/53 [==============================] - 3s 48ms/step - loss: 0.2503 - mape: 1103638.3750 - val_loss: 0.2530 - val_mape: 1286848.8750\n",
            "Epoch 739/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1128271.6250 - val_loss: 0.2525 - val_mape: 1006084.1250\n",
            "Epoch 740/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2505 - mape: 1118570.3750 - val_loss: 0.2526 - val_mape: 1027621.6250\n",
            "Epoch 741/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1110385.0000 - val_loss: 0.2531 - val_mape: 1263293.2500\n",
            "Epoch 742/1000\n",
            "53/53 [==============================] - 3s 47ms/step - loss: 0.2504 - mape: 1116835.0000 - val_loss: 0.2525 - val_mape: 1004483.8750\n",
            "Epoch 743/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1117663.3750 - val_loss: 0.2524 - val_mape: 1201321.7500\n",
            "Epoch 744/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1104761.2500 - val_loss: 0.2528 - val_mape: 970750.4375\n",
            "Epoch 745/1000\n",
            "53/53 [==============================] - 3s 47ms/step - loss: 0.2504 - mape: 1115315.1250 - val_loss: 0.2525 - val_mape: 1119730.0000\n",
            "Epoch 746/1000\n",
            "53/53 [==============================] - 3s 47ms/step - loss: 0.2505 - mape: 1124071.7500 - val_loss: 0.2527 - val_mape: 976345.6250\n",
            "Epoch 747/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2503 - mape: 1109360.6250 - val_loss: 0.2526 - val_mape: 1166772.8750\n",
            "Epoch 748/1000\n",
            "53/53 [==============================] - 3s 48ms/step - loss: 0.2505 - mape: 1122601.5000 - val_loss: 0.2525 - val_mape: 1254908.6250\n",
            "Epoch 749/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2502 - mape: 1096381.2500 - val_loss: 0.2530 - val_mape: 1194100.1250\n",
            "Epoch 750/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2505 - mape: 1127294.8750 - val_loss: 0.2527 - val_mape: 986576.0625\n",
            "Epoch 751/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1114963.3750 - val_loss: 0.2526 - val_mape: 1050712.1250\n",
            "Epoch 752/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2504 - mape: 1115087.1250 - val_loss: 0.2528 - val_mape: 981259.3750\n",
            "Epoch 753/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1117457.8750 - val_loss: 0.2526 - val_mape: 1253330.8750\n",
            "Epoch 754/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2502 - mape: 1099152.6250 - val_loss: 0.2526 - val_mape: 1141116.6250\n",
            "Epoch 755/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2506 - mape: 1130441.3750 - val_loss: 0.2525 - val_mape: 1010324.6875\n",
            "Epoch 756/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2504 - mape: 1111087.0000 - val_loss: 0.2528 - val_mape: 982569.7500\n",
            "Epoch 757/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2503 - mape: 1100746.8750 - val_loss: 0.2529 - val_mape: 1379470.8750\n",
            "Epoch 758/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1125064.6250 - val_loss: 0.2527 - val_mape: 1082880.2500\n",
            "Epoch 759/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2503 - mape: 1102112.2500 - val_loss: 0.2527 - val_mape: 994769.8125\n",
            "Epoch 760/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1120757.7500 - val_loss: 0.2524 - val_mape: 1185330.6250\n",
            "Epoch 761/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1106236.7500 - val_loss: 0.2533 - val_mape: 1437672.3750\n",
            "Epoch 762/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2503 - mape: 1115476.3750 - val_loss: 0.2525 - val_mape: 1027728.6250\n",
            "Epoch 763/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1113849.1250 - val_loss: 0.2525 - val_mape: 1236987.7500\n",
            "Epoch 764/1000\n",
            "53/53 [==============================] - 3s 47ms/step - loss: 0.2504 - mape: 1110366.6250 - val_loss: 0.2525 - val_mape: 1218868.8750\n",
            "Epoch 765/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2504 - mape: 1116715.8750 - val_loss: 0.2526 - val_mape: 973116.8750\n",
            "Epoch 766/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2502 - mape: 1098942.0000 - val_loss: 0.2525 - val_mape: 1093926.8750\n",
            "Epoch 767/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2505 - mape: 1122976.6250 - val_loss: 0.2525 - val_mape: 1136638.1250\n",
            "Epoch 768/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2504 - mape: 1108980.6250 - val_loss: 0.2526 - val_mape: 1014274.1250\n",
            "Epoch 769/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1103289.8750 - val_loss: 0.2528 - val_mape: 1372779.2500\n",
            "Epoch 770/1000\n",
            "53/53 [==============================] - 3s 48ms/step - loss: 0.2504 - mape: 1110674.7500 - val_loss: 0.2527 - val_mape: 1072526.3750\n",
            "Epoch 771/1000\n",
            "53/53 [==============================] - 3s 48ms/step - loss: 0.2504 - mape: 1117092.7500 - val_loss: 0.2526 - val_mape: 1129606.3750\n",
            "Epoch 772/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1111022.0000 - val_loss: 0.2527 - val_mape: 1001522.5000\n",
            "Epoch 773/1000\n",
            "53/53 [==============================] - 3s 48ms/step - loss: 0.2504 - mape: 1104713.2500 - val_loss: 0.2528 - val_mape: 1231026.6250\n",
            "Epoch 774/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1115604.8750 - val_loss: 0.2526 - val_mape: 929575.5000\n",
            "Epoch 775/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1105095.3750 - val_loss: 0.2527 - val_mape: 1002408.0000\n",
            "Epoch 776/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2503 - mape: 1106045.8750 - val_loss: 0.2526 - val_mape: 1137620.8750\n",
            "Epoch 777/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1110536.2500 - val_loss: 0.2526 - val_mape: 1056525.3750\n",
            "Epoch 778/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2503 - mape: 1107147.0000 - val_loss: 0.2525 - val_mape: 1251125.6250\n",
            "Epoch 779/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1115015.5000 - val_loss: 0.2525 - val_mape: 984533.6250\n",
            "Epoch 780/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2503 - mape: 1098900.1250 - val_loss: 0.2531 - val_mape: 1380253.6250\n",
            "Epoch 781/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1108227.1250 - val_loss: 0.2525 - val_mape: 1169176.6250\n",
            "Epoch 782/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2505 - mape: 1119212.0000 - val_loss: 0.2526 - val_mape: 980137.0000\n",
            "Epoch 783/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2503 - mape: 1102133.6250 - val_loss: 0.2527 - val_mape: 1234983.1250\n",
            "Epoch 784/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2504 - mape: 1112982.7500 - val_loss: 0.2525 - val_mape: 1009992.4375\n",
            "Epoch 785/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2504 - mape: 1106075.6250 - val_loss: 0.2526 - val_mape: 1205219.2500\n",
            "Epoch 786/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1112147.5000 - val_loss: 0.2526 - val_mape: 985624.8750\n",
            "Epoch 787/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2503 - mape: 1096716.7500 - val_loss: 0.2528 - val_mape: 1012917.7500\n",
            "Epoch 788/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1107757.6250 - val_loss: 0.2526 - val_mape: 1260262.1250\n",
            "Epoch 789/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2504 - mape: 1115462.1250 - val_loss: 0.2525 - val_mape: 979618.5000\n",
            "Epoch 790/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1107200.6250 - val_loss: 0.2527 - val_mape: 1273930.1250\n",
            "Epoch 791/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2502 - mape: 1099152.0000 - val_loss: 0.2528 - val_mape: 1125355.7500\n",
            "Epoch 792/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1112093.3750 - val_loss: 0.2527 - val_mape: 931217.7500\n",
            "Epoch 793/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2503 - mape: 1099644.1250 - val_loss: 0.2527 - val_mape: 1387046.6250\n",
            "Epoch 794/1000\n",
            "53/53 [==============================] - 2s 45ms/step - loss: 0.2504 - mape: 1109577.8750 - val_loss: 0.2528 - val_mape: 911480.8125\n",
            "Epoch 795/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2503 - mape: 1106133.5000 - val_loss: 0.2525 - val_mape: 1200048.1250\n",
            "Epoch 796/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2503 - mape: 1107170.8750 - val_loss: 0.2524 - val_mape: 971725.3125\n",
            "Epoch 797/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1104438.2500 - val_loss: 0.2526 - val_mape: 984431.7500\n",
            "Epoch 798/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2503 - mape: 1097341.0000 - val_loss: 0.2526 - val_mape: 1008640.8125\n",
            "Epoch 799/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1114972.3750 - val_loss: 0.2526 - val_mape: 1007710.7500\n",
            "Epoch 800/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2502 - mape: 1090785.3750 - val_loss: 0.2529 - val_mape: 954722.0000\n",
            "Epoch 801/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2504 - mape: 1115522.0000 - val_loss: 0.2525 - val_mape: 1145207.7500\n",
            "Epoch 802/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1105993.3750 - val_loss: 0.2527 - val_mape: 995251.7500\n",
            "Epoch 803/1000\n",
            "53/53 [==============================] - 3s 48ms/step - loss: 0.2503 - mape: 1101380.1250 - val_loss: 0.2527 - val_mape: 1161535.1250\n",
            "Epoch 804/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2504 - mape: 1104336.2500 - val_loss: 0.2526 - val_mape: 1057272.5000\n",
            "Epoch 805/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2503 - mape: 1101255.3750 - val_loss: 0.2528 - val_mape: 1262705.2500\n",
            "Epoch 806/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2503 - mape: 1103999.5000 - val_loss: 0.2526 - val_mape: 988470.1250\n",
            "Epoch 807/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1108298.0000 - val_loss: 0.2525 - val_mape: 977233.1250\n",
            "Epoch 808/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2503 - mape: 1102773.8750 - val_loss: 0.2525 - val_mape: 1245023.5000\n",
            "Epoch 809/1000\n",
            "53/53 [==============================] - 2s 47ms/step - loss: 0.2503 - mape: 1093197.3750 - val_loss: 0.2527 - val_mape: 1262212.6250\n",
            "Epoch 810/1000\n",
            "53/53 [==============================] - 2s 46ms/step - loss: 0.2504 - mape: 1114018.0000 - val_loss: 0.2526 - val_mape: 1043874.4375\n",
            "Epoch 00810: early stopping\n"
          ]
        }
      ],
      "source": [
        "from tensorflow.keras.layers import Input, Conv2D, MaxPooling2D, Conv2DTranspose, BatchNormalization\n",
        "\n",
        "model = Sequential()\n",
        "\n",
        "model.add(Input(shape=(28, 28, 1)))\n",
        "model.add(Conv2D(32, (3, 3), activation=\"relu\", padding=\"same\"))\n",
        "model.add(Conv2D(32, (3, 3), activation=\"relu\", padding=\"same\"))\n",
        "model.add(Conv2D(32, (3, 3), activation=\"relu\", padding=\"same\"))\n",
        "model.add(MaxPooling2D((2, 2), padding=\"same\"))\n",
        "model.add(Conv2D(32, (3, 3), activation=\"relu\", padding=\"same\"))\n",
        "model.add(Conv2D(32, (3, 3), activation=\"relu\", padding=\"same\"))\n",
        "model.add(Conv2D(32, (3, 3), activation=\"relu\", padding=\"same\"))\n",
        "model.add(MaxPooling2D((2, 2), padding=\"same\"))\n",
        "model.add(Dense(2))\n",
        "model.add(Conv2DTranspose(32, (3, 3), strides=2, activation=\"relu\", padding=\"same\"))\n",
        "model.add(Conv2D(32, (3, 3), activation=\"relu\", padding=\"same\"))\n",
        "model.add(Conv2D(32, (3, 3), activation=\"relu\", padding=\"same\"))\n",
        "model.add(Conv2DTranspose(32, (3, 3), strides=2, activation=\"relu\", padding=\"same\"))\n",
        "model.add(Conv2D(32, (3, 3), activation=\"relu\", padding=\"same\"))\n",
        "model.add(Conv2D(32, (3, 3), activation=\"relu\", padding=\"same\"))\n",
        "model.add(Conv2D(1, (3, 3), activation=\"sigmoid\", padding=\"same\"))\n",
        "\n",
        "\n",
        "\n",
        "model.compile(optimizer='Nadam', loss='binary_crossentropy', metrics=[\"mape\"])\n",
        "\n",
        "from tensorflow.keras.callbacks import EarlyStopping\n",
        "early_stopping = EarlyStopping(\n",
        "    monitor='val_loss',\n",
        "    verbose=1,\n",
        "    patience=50)  \n",
        "hist = model.fit(train_x, train_x, epochs=1000, batch_size=1024, validation_split=0.1, callbacks=[early_stopping])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 548,
      "metadata": {
        "id": "1C3rJbWR-FNn",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "outputId": "3372f27a-312a-41fd-d656-d9b018842a64"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.plot(hist.history['loss'][10:])\n",
        "plt.plot(hist.history['val_loss'][10:])\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# threshold 구하기"
      ],
      "metadata": {
        "id": "SPcqhxyQYyWp"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "reconstructed = model.predict(train_x)"
      ],
      "metadata": {
        "id": "Hd9bXRL8lmkA"
      },
      "execution_count": 549,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plt.figure(figsize=(10,10))\n",
        "for i in range(25):\n",
        "    plt.subplot(5,5,i+1)\n",
        "    plt.imshow(reconstructed[i].reshape((28,28)), cmap=\"binary\")\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 594
        },
        "id": "Vf0Bbk-NlkhE",
        "outputId": "0b1d0a69-8296-425d-8c2e-d8dd16f1460e"
      },
      "execution_count": 550,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x720 with 25 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "diff = np.abs(reconstructed - train_x)\n",
        "plt.hist(diff.flatten(), bins=100)\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 276
        },
        "id": "v638F7RF70t8",
        "outputId": "1bb9eebb-9c4c-4c16-ee36-f6ddd1ec971a"
      },
      "execution_count": 551,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.figure(figsize=(10,15))\n",
        "for i in range(5):\n",
        "    plt.subplot(5,3,i*3+1)\n",
        "    plt.imshow(train_x[i].reshape((28,28)), cmap=\"binary\")\n",
        "    plt.subplot(5,3,i*3+2)\n",
        "    plt.imshow(reconstructed[i].reshape((28,28)), cmap=\"binary\")\n",
        "    plt.subplot(5,3,i*3+3)\n",
        "    plt.imshow(diff[i].reshape((28,28)), cmap=\"binary\")\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 865
        },
        "id": "mq0LLISVDfsm",
        "outputId": "0a75171b-79b2-4fc5-ee31-28bcc34190df"
      },
      "execution_count": 552,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x1080 with 15 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "sorted_diff = np.sort(diff.flatten())\n",
        "threshold_index = int(len(sorted_diff)*0.999)\n",
        "threshold = sorted_diff[threshold_index]\n",
        "\n",
        "print(threshold)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KzzVAg6E8B1k",
        "outputId": "a3dbdb43-78c3-4993-b9d7-47b5c94720c6"
      },
      "execution_count": 553,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0.4629009309936972\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "lage_diff_count = (diff>threshold).sum(axis=(1,2,3))\n",
        "plt.hist(lage_diff_count, bins=50)\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "Pq23LiIgCVlh",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "outputId": "c1a4ab42-83bb-4ece-9fb9-fc3ddb2f83c8"
      },
      "execution_count": 554,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 탐지 실행"
      ],
      "metadata": {
        "id": "BV924dHvYvP-"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "reconstructed = model.predict(test_x)"
      ],
      "metadata": {
        "id": "WgwbVUElyuN1"
      },
      "execution_count": 555,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plt.figure(figsize=(10,10))\n",
        "for i in range(25):\n",
        "    plt.subplot(5,5,i+1)\n",
        "    plt.imshow(reconstructed[i].reshape((28,28)), cmap=\"binary\")\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 594
        },
        "id": "yRf5AvTtys9i",
        "outputId": "28d63f61-92b1-4667-df32-991e29420c2e"
      },
      "execution_count": 556,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x720 with 25 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "diff = np.abs(reconstructed - test_x)\n",
        "plt.hist(diff.flatten(), bins=100)\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "wmIaZWVe8W0Q",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 276
        },
        "outputId": "ec1b57a6-6e49-416c-a65c-fb9a3c57321e"
      },
      "execution_count": 557,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.figure(figsize=(10,15))\n",
        "for i in range(5):\n",
        "    plt.subplot(5,3,i*3+1)\n",
        "    plt.imshow(test_x[i].reshape((28,28)), cmap=\"binary\")\n",
        "    plt.subplot(5,3,i*3+2)\n",
        "    plt.imshow(reconstructed[i].reshape((28,28)), cmap=\"binary\")\n",
        "    plt.subplot(5,3,i*3+3)\n",
        "    plt.imshow(diff[i].reshape((28,28)), cmap=\"binary\")\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "GrUrXrY4DqYc",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 865
        },
        "outputId": "1d807be7-5556-426a-ed54-bbf3dfff32e6"
      },
      "execution_count": 558,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x1080 with 15 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "lage_diff_count = (diff>threshold).sum(axis=(1,2,3))\n",
        "\n",
        "half_index = int(len(test_x)*0.5)\n",
        "plt.hist(lage_diff_count[:half_index], bins=40, alpha=0.5)\n",
        "plt.hist(lage_diff_count[half_index:], bins=40, alpha=0.5)\n",
        "plt.show()\n",
        "\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "KJEaeVZN8YIe",
        "outputId": "8397f6dd-1168-4e46-d049-2a98ae898f4c"
      },
      "execution_count": 559,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "predicted = (lage_diff_count>0).astype(np.int16)\n",
        "\n",
        "n = len(test_x)\n",
        "accuracy = (predicted == test_y).sum() / n\n",
        "\n",
        "print(\"accuracy=\", accuracy)"
      ],
      "metadata": {
        "id": "8Z7RVlTCB_P3",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4a958f6d-a06d-4a03-f843-a13312a835dc"
      },
      "execution_count": 560,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "accuracy= 0.7196\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "_9ZoKZYJFm3B"
      },
      "execution_count": 560,
      "outputs": []
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "name": "anomaly_detection_fahsion_mnist.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}